I. Background
1. On May 18, 1998, plaintiffs the United States and twenty States and the
District of Columbia filed actions against defendant Microsoft Corporation, alleging
violations of the Sherman Act, 15 U.S.C. §§ 1 & 2, and the antitrust and consumer
protection laws of the respective plaintiff States. The actions were consolidated, and
expedited discovery ensued. Trial began on October 18, 1998, and concluded on June
26, 1999.
2. Defendant Microsoft Corporation ("Microsoft") is a corporation organized
under the laws of the State of Washington with its headquarters in Redmond,
Washington.
-
Answer ¶ 41.
3. Microsoft's principal business is the licensing of computer software, which it conducts
on a world-wide basis. Microsoft licenses computer software throughout the United
States and elsewhere and delivers operating systems to computer manufacturers and
others across states lines and international borders, and its business has had a
substantial effect on interstate commerce.
- Answer ¶ 5.
4. Microsoft, among other things, licenses operating system and application software for
personal computers. The personal computer industry, which has seen tremendous
growth over the last decade, is an important, robust sector of the United States
economy. Microsoft software dominates critical sectors of that industry.
- See infra Part II (Microsoft possess monopoly power in operating
systems).
- Tevanian Dir. ¶¶ 6, 14, 22, 35 (Microsoft is also dominant in a number of
applications, including office productivity suites).
5. A Personal Computer ("PC") is a computer designed for use by one person at a time.
- Microsoft Press Computer Dictionary, at 361 (3d ed. 1997) (GX 1050).
5.1 PCs (which include both desktop and laptop models), can be distinguished from more
powerful, more expensive computers known as Servers, which are designed to provide
services and functionality to multiple users, either in local area network or over the
Internet.
- Warren-Boulton Dir. ¶ 20.
- Microsoft Press Computer Dictionary, at 430 (3d ed. 1997) (GX
1050).
5.2 A typical PC system consists of a number of components, including a
microprocessor, dynamic memory, a hard disk, a keyboard, a monitor, and an operating
system.
- Warren-Boulton Dir. ¶ 20.
6. PCs are built primarily by firms known as Original Equipment Manufacturers
("OEMs"). OEMs typically purchase from different third-party vendors and preinstall
various hardware and software components for their systems, including the operating
system and application software.
- Warren-Boulton Dir. ¶ 23.
7. OEMs develop and sell their PCs to consumers in a competitive market and
design their PCs and their hardware and software features to respond to consumer
demand.
- See infra Part II.A; ¶ 15.1.1.
- Warren-Boulton Dir. ¶ 24.
8. An Operating System is the "central nervous system" of the PC.
- Barksdale Dir. ¶ 69.
8.1. An operating system performs two basic functions. First, the
operating system allows the various components of the PC to communicate and
function with each other; it provides "the software that controls the allocation and usage
of hardware resources such as memory, central processing unit time, disk space, and
peripheral devices."
- Microsoft Press Computer Dictionary, at 341 (3d ed. 1997).
- Farber Dir. ¶ 11 (the operating system "controls the execution of programs
on computer systems and may provide low-level services such as
resource allocation, scheduling and input-output control in a form which is
sufficiently simple and general so that these services are broadly useful to
software developers").
8.2. Second, an operating system provides a "platform" by exposing
Applications Programming Interfaces ("APIs") that applications use to "call upon" the
operating system's underlying software routines in order to perform various functions,
such as displaying a character on a monitor.
-
Schmalensee Dir. ¶¶ 93-94.
9. An Application is a software program "used to perform specific user-oriented
tasks".
- Farber Dir. ¶ 11.
9.1. Applications typically "run on top" of the operating system and draw
upon the services that the operating system's "platform" provides.
- Warren-Boulton Dir. ¶ 22
9.2. The term Platform is used in the software industry to describe
software that "provides features or services that can be used by software applications."
- Schmalensee Dir. ¶ 93.
10. Microsoft produces a number of PC operating systems, including MS-DOS
and successive versions of its Windows operating system, the most recent version of
which is Windows 98. Since at least the mid-1990s, Microsoft has dominated the
market for PC operating systems. As will be explained, Microsoft's market share has
remained well in excess of 90% during that period.
- See infra Part II.B.2; ¶ 21.
11. Applications are produced by numerous firms, including firms like Microsoft
that also produce operating systems and others, known as Independent Software
Vendors ("ISVs"). Microsoft's application software is dominant in several key
categories, most notably in office productivity suites.
- See infra Part V.F.1.b.(1).; ¶ 287.2.1.
12. All the components of a PC system -- the microprocessor and other
hardware, operating system, and applications software -- must be compatible with each
other. For instance, software, including the operating system and applications, must be
designed to be compatible with the PC's microprocessor, and application software must
be compatible with the operating system.
12.1. There are different types of PC systems.
12.1.1. An Intel-compatible PC is one designed to function with
Intel's x86/Pentium families of microprocessors or compatible microprocessors
manufactured by Intel or other firms. Microsoft's Windows operating system, and
different types of UNIX operating systems, are examples of operating systems that run
on Intel-compatible PCs.
- Fisher Dir. ¶ 62.
12.1.2. There are other types of PCs that use microprocessors that
are not Intel-compatible, such as the Apple Macintosh computer system. Operating
systems designed to run on Intel-compatible PCs, known as Intel-compatible PC
operating systems, will not function on an incompatible PC like the Macintosh; and
operating systems designed for an incompatible PC like the Macintosh will not function
on an Intel-compatible PC.
- Gosling Dir. ¶ 7.
12.2. Applications programs are typically written to run on a particular
operating system and cannot run on other operating systems unless the developer
goes to the time and expense to "port" the program to the other operating system. For
example, the version of Microsoft's popular Office productivity suite designed to run on
Microsoft's Windows operating system cannot run on the Apple Macintosh or even on
other Intel-compatible operating systems.
- See infra II.B.3.b.(1); ¶ 26.1.2.
13. One of the most important applications today is an Internet Web browser
("browser").
13.1. A browser is a "client application that enables a user to view HTML
documents on the World Wide Web, another network, or the users's computer; follow
the hyperlinks among them; and transfer files." A browser enables "the user to
examine, display, scan, and navigate via the Internet" information located on the "Web."
- Microsoft Press Computer Dictionary, at 505 (3d ed. 1997) (GX
1050).
- Farber Dir. ¶ 11.
13.1.1. The Internet is a global network that links many millions of
PCs and a smaller number of servers together. Begun in the early 1960s, the Internet
exploded in popularity with the emergence of the World Wide Web ("Web") in the mid-1990s.
- Maritz Dir. ¶ 50.
13.1.2. "The Internet is a global network of computers constructed
by patching together many local area networks that use widely varying communication
media such as telephone lines, dedicated data cables, and wireless links." The Internet
links PCs by means of servers, which run specialized operating systems and
applications designed for servicing a network environment.
- Felten Dir. ¶ 11.
13.1.3. In simplest terms, servers host and provide access to the
Internet's content. In the case of the Web, this content consists principally of Web
Pages, which are created by Internet Content Providers ("ICPs"). There are millions of
web pages located on the thousands of servers that comprise the Internet.
- See infra Part V.E.1.a; ¶ 255.
13.1.4. Web pages can be accessed over those thousands of
servers from millions of PCs because the Internet uses a number of widely-accepted
standards. For instance, web pages are typically written in Hypertext Markup Language
(HTML) and are transferred between servers and PCs using a common protocol known
as Hypertext Transfer Protocol (HTTP).
- Felten Dir. ¶ 13. (The Web is "characterized by a set of
standard data formats, including HyperText Markup
Language ('HTML'), and a set of standard communication
protocols, such as HyperText Transfer Protocol ('HTTP'),
that together allow computers to share multimedia
documents that may contain links to other such
documents.").
13.1.5. Consumers typically access the Internet through the
services of an Internet Access Provider, which can be an Internet Service Provider
("ISP"), such as Earthlink or AT&T Worldnet, or an On-Line Service ("OLS"), such as
America Online or Prodigy. Internet access providers are commercial firms that
connect users to the network of servers that comprise the Internet.
- See infra V.D.1; ¶ 213.
13.2. Although graphical web browsers have existed since 1993, the first
widely-popular commercial graphical browser was developed and brought to market by
Netscape Communications in late 1994. Microsoft introduced its browser, Internet
Explorer, in 1995.
- See infra Part III.B.1; ¶ 53.1.1; Part V.B.2.c; ¶ 126.
II. Microsoft Possesses Monopoly Power Over Operating Systems
14. Microsoft possesses monopoly power over operating systems for Intel-compatible personal computers.
14.1. Microsoft's monopoly power in Intel-compatible personal computers
is demonstrated by its customers' lack of any commercially viable alternative to
Windows and certain Microsoft conduct that makes sense only if there is a monopoly to
protect. See infra Part II.A; ¶¶ 15-16.
14.2. A traditional structural analysis, which shows that Microsoft
possesses a dominant market share protected by immense barriers to entry, confirms
that Microsoft has monopoly power. See infra Part II.B; ¶¶ 17-32.
14.3. Microsoft's monopoly power is also evidenced by its ability to
control price. See infra Part II.C; ¶¶ 33-38.
14.4. Dean Schmalensee's analysis that Microsoft lacks monopoly power
is contrary to the evidence, inconsistent with his prior testimony and writings, and
otherwise unreliable. See infra Part II.D; ¶¶ 39-50.
A. Microsoft's monopoly power is established by direct evidence of its
existence and exercise
15. That Microsoft has monopoly power in operating systems is directly
evidenced by the "sustained absence of realistic commercial alternatives" to Microsoft's
operating system product.
- Fisher, 6/1/99am, at 11:17-18.
15.1. Microsoft's principal customers, computer manufacturers (OEMs),
lack any commercially viable alternative to Windows.
15.1.1. OEMs are the most important direct customers of operating
systems. Because competition among OEMs is intense, they respond to consumer
demand. OEMs thus not only are important customers in their own right, but also are
surrogates for determining the commercial alternatives reasonably available to
consumers.
- Dr. Warren-Boulton testified that the "great majority of
operating systems installed on PCs are installed on new
machines by OEMs." Warren-Boulton Dir. ¶ 23; id. at ¶ 23
n.7 (noting that in 1997, 87.6% of all copies of Windows 95
were installed by OEMs).
- Professor Fisher testified: "OEMs's are, in some sense, the
representative of the consumer for certain purposes. They
are in competition with each other. They gain if they deliver
what end users actually want. They wouldn't care about the
restrictions on them if they don't think that it mattered in their
dealings with consumers." Fisher, 6/2/99am, at 22:1-6.
- Dean Schmalensee conceded that "OEMs respond to
consumer demand." Schmalensee, 1/25/99am, at 15:16
(sealed session).
- See also Rose Dir. ¶ 17 ("If there were sufficient customer
demand for a different operating systems for personal
computers, Compaq would consider licensing that operating
system."); Von Holle Dep., 1/13/99, at 299:15 - 300:1 ("if
viable alternative emerged" to Windows, Gateway "would
evaluate" them because Gateway likes "to make sure that"
its "customers are offered a . . . choice of products that
become popular in the market place"); Ransom Dep. (played
12/16/98pm), at 71:20 - 72:4 ("If there's a product with a
competitive advantage or a price advantage, frankly, we
would consider it. But it has not been presented to us.").
15.1.2. OEMs uniformly testify that they lack any commercially
viable alternative to Windows:
- The testimony of Garry Norris, former Director of Strategy
and Software at IBM Personal Computer Company, vividly
illustrates the absence of commercially viable alternatives to
Windows. Norris testified that, "without Windows 95, you
couldn't be in the P.C. business." Norris, 6/7/99am, at
66:18-20. Indeed, Norris explained, IBM concluded in the
summer of 1995 that, if it did not obtain a Windows 95
license, it would "lose . . . anywhere from 30 to 90 percent"
of its sales volume, and "the IBM P.C. company would be
out of business" in "three to twelve months." Norris,
6/7/99am, at 65:16 - 67:18.
- The testimony of Microsoft's own OEM witness, Compaq's
John Rose, illustrates OEMs' dependence on Windows.
Compaq preinstalls Microsoft operating systems on over
90% of its PCs, including 100% of its popular Presario line,
Rose, 2/17/99pm, at 12:25 - 15:3; Rose Dir. ¶ 17 (since
1993, Compaq has "not consistently loaded any alternatives
to Windows on personal computers it markets to
consumers."), because Compaq has no commercially viable
alternative to Windows. Rose, 2/17/99pm, at 8:16-20.
- Gateway's Penny Nash testified that for Gateway to stop
licensing Microsoft operating systems would "be suicide."
Fisher Dir. ¶ 63 (quoting Nash Dep. 11/18/97, at 5-6); see
also Von Holle Dep., 1/13/99, at 298:2-23, GX 357 (sealed);
Fisher Dir. ¶ 63 (quoting Brown Dep., 3/5/98, at 10-11).
- Other OEMs gave similar testimony: Mal Ransom of
Packard Bell, a leading OEM, testified that Packard Bell pre-installs Windows on 100% of its PCs and has done so for
several years. Ransom Dep. (played 12/16/98pm), at 68:14
- 69:23. Packard Bell loads Windows because it is "the only
viable choice." Ransom Dep. (played 12/16/98pm), at 69:5.
Frank Santos testified that Hewlett-Packard has not
considered any other operating system for its consumer line
of PCs "because there isn't any out there." Fisher Dir. ¶ 63
(quoting Santos Dep., 4/13/98, at 7-8).
15.1.3. All three economic experts in this case agreed that there is
no commercially viable alternative to Windows to which a significant OEM can switch in
response to a substantial price increase or its equivalent by Microsoft.
- Professor Fisher testified that Microsoft's power is shown by
evidence that "Microsoft's customers do not believe that they
have serious commercial alternatives to Windows." Fisher,
6/1/99am, at 11:9-19; see also Fisher Dir. ¶ 63.
- Dean Schmalensee conceded that there are no reasonable
substitutes for Windows to which a major OEM can switch
and that Microsoft can raise the short-term price of
Windows. Schmalensee, 1/20/99am, at 33:3-8; see also
1/13/99pm, at 68:17 - 69:2.
- Dr. Warren-Boulton testified that OEMs consider Windows
"commercially necessary" and that "if confronted with a 10%
increase in their Windows license, they would not switch to
operating system products for other hardware platforms."
Warren-Boulton Dir. ¶ 39 (summarizing OEM testimony);
Warren-Boulton, 11/23/98pm, at 70:9-12 (testifying that it is
"commercially necessary to be able to offer Microsoft
operating system . . . to end users").
15.1.4. Microsoft knows that OEMs have no choice but to load
Windows.
15.1.4.1. Microsoft told OEMs that they lack any alternative
to Windows and, indeed, that Microsoft was "the only game in town."
- Norris of IBM testified that Microsoft executives
repeatedly sought to use the fact that IBM had no
"commercially viable alternative" to Windows (Norris,
6/7/99am, at 66:18-20), and feared losing access to
Windows, to pressure IBM into dropping products that
competed with Microsoft. See infra Part V.C.2.b.(3);
¶¶ 209-212. Indeed, Norris testified, the Microsoft
executive in charge of its relationship with the IBM PC
company bluntly told IBM during negotiations, "'where
else are you going to go? This is the only game in
town.'" Norris, 6/7/99am, at 66:21 - 67:6.
15.1.4.2. OEMs told Microsoft that they lack any viable
alternative to Windows.
- John Romano of Hewlett Packard wrote to Microsoft,
when it imposed costly screen restrictions upon
Hewlett Packard, that "if we had another supplier, I
guarantee you would not be our supplier of choice."
GX 309.
- Gateway urged Microsoft
- redacted -
GX 357 (sealed).
15.1.4.3. Other operating system vendors recognize that
they do not provide a viable alternative to Windows.
- John Soyring of IBM testified: "As a result of the
applications and device support for Windows, in my
view, suppliers of PCs have no commercially viable
choice but to license Windows and to offering on the
vast majority of PCs they ship." Soyring Dir. ¶ 11.
- Avadis Tevanian of Apple computer testified: "For the
foreseeable future, Microsoft will maintain a market
share in excess of 90 percent of the desktop
operating system market, a dominance that will
enable it to continue to effectively control both price
and technologies." Tevanian Dir. ¶ 14.
- The CEO of Red Hat Linux also insists that Red Hat is not
a viable competitor to Microsoft. In a Washington Post
article he said: "'It just tells you how desperate Microsoft is
for a competitor that they're holding up a software box
produced by 100 guys in the hills of North Carolina.'" He
also said: "'We are absolutely not a viable competitor at
this time. We have every intention of being one. But how
long will it take? Realistically, it will be twenty years.'"
GX 1568.
15.1.5. Microsoft set the Windows royalty recognizing that OEMs
have no viable alternative to Windows.
- Joachim Kempin, Microsoft's Vice President for OEM sales,
testified that the prices set by other operating system
vendors were not a consideration in setting the Windows 98
royalty. Kempin, 2/25/99pm, at 97:24 - 98:23. To the
contrary, Microsoft set the royalty for Windows 98 by
"'compar[ing] it with Windows 95.'" Id. at 98:6 (quoting
Kempin's deposition, 21:20 - 22:6); see also Kempin,
2/25/99pm, at 98:15-23 (quoting Kempin's deposition, 22:10-22:6) (Kempin also did not consider "'competition more
generally'").
- Kempin testified that he did not consider the prices set by
other operating system vendors because, "with Windows 95
or 98, when it comes to value propositions, it just doesn't
come close to anything else. Meaning I believe competitors
are basically selling inferior-type products." Competitors
products are "inferior," Kempin explained, because "the
number of applications, peripheral devices, support on that
platform, basically, is so huge that the benefits of buying into
that platform is huge." Kempin, 2/25/99pm, at 98:24 - 99:5
(quoting Kempin's deposition, 22:19-24).
- Kempin, in contemplating "OEM pricing thoughts," wrote that
although conceivably, "[o]ur high prices could get a single
OEM . . . or a coalition to fund a competing effort," he
considers it "doubtful." He concluded: "Could they convince
customer to change their computing platform is the real
questions. [sic]. The existing investments in training,
infrastructure and applications in windows computing are
huge and will create a lot of inertia." GX 365.
15.1.6. OEMs do not believe alternatives to Windows are likely to
emerge in the next several years such that Microsoft is constrained from being able to
raise price or reduce quality today.
- Garry Norris testified that without a Windows 95 license, "the
IBM P.C. company would be out of business" in "three to
twelve months." Norris, 6/7/99am, at 65:16 - 67:18.
- Professor Fisher testified that there is no reason to "believe
that OEMs would substitute other operating systems for
Microsoft's Windows operating system in favor of anything
that can now be seen on the horizon"; that is, in "the next
few years." Fisher, 1/6/99am, at 69:23 - 70:1.
15.2. Both OEMs and applications developers (ISVs) recognize that they
are dependent on Microsoft and fear that Microsoft will use its monopoly power to harm
them if they favor Microsoft's rivals.
- When Microsoft released a Java development kit that reflected
Microsoft's "breaking away from pure Java," Paul White of
Symantec, an ISV, wrote that "it's better to say nothing than risk the
blast from MS." GX 2078.
- Barry Schuler of AOL testified that, because its applications must
run on Windows, "there's an absolute dependency on what the
future direction of that operating system." DX 2810.
- William Harris testified: "Intuit's dependence on the Windows
operating system creates additional dependence on the supplier of
the operating system, Microsoft. We depend on Microsoft for the
information, specifications, training, development assistance and
development tools necessary to develop our products in an
effective and timely manner." Harris Dir. ¶ 28.
- Hewlett Packard's John Romano testified that
- redacted - DX 2582A (sealed).
- A Compaq presentation entitled "Microsoft Meeting Preparation --
Portable and Software Marketing PC Division" dated January 13,
1993, states:
- redacted - The presentation continued: "Judgment: How
retaliatory would they get?" and lists the possibilities as follows:
"Pricing advantage -- Revenue from updates -- Access to early
SDKs -- Field sales activities (Microsoft has ~900 field sales
people) -- Support and training -- Inclusion in advertising -- Tone
toward Compaq in press and with customers -- Selection and
elevation of other OEMs as leaders -- Make integration relations
even more strained than they are today -- Access to source code,
modification ownership -- Microsoft directional information and
plans -- Customers." GX 433 (sealed).
16. Microsoft repeatedly took actions that make sense only because it has
monopoly power to protect.
- Fisher, 6/1/99am, at 12:14-17.
16.1. Microsoft's expensive effort to gain browser usage share can be
explained only as an effort to protect Microsoft's position in operating system and thus
demonstrates substantial and durable market power.
- As detailed below, Microsoft engaged in a very costly course of
conduct designed to gain a substantial share of the market for
Internet browsers. See infra Part V.G.
- This conduct evidences monopoly power because, as will be
explained (see infra Part V.G.), Microsoft could not have expected
to recoup its hundreds of millions of dollars in browser-related costs
except by thwarting threats to its position in operating systems and
thereby increasing or prolonging its monopoly profits in operating
systems.
16.2. Microsoft's monopoly power is also evidenced by its ability, for
several years, to force other firms to cooperate in Microsoft's efforts to exclude threats
to its dominant position in operating systems.
16.2.1. This conduct includes, among other things:
- Forcing OEMs to accept Microsoft's Internet Explorer
browser as a condition of licensing Microsoft's Windows
operating system. See infra Part V.B.
- Forcing OEMs to agree to costly restrictions on their ability
to customize their PC systems; OEMs agreed to those
restrictions, in the words of one executive, because they
lack any "choice of another supplier." GX 309. See infra
Part V.C.1.
- Threatening to retaliate against OEMs that favored products
that threaten Microsoft's operating system monopoly. See
infra Part V.C.2.
- Threatening to retaliate against Intel if Intel developed
platform-level software or favored Netscape or Sun in
various ways. See infra Part VI.
16.2.2. This conduct is part of a predatory course of conduct that
makes no sense unless Microsoft expected it to lead to monopoly recoupment in the
operating system market. All these acts reduced the value of Windows to end users.
Microsoft would not rationally have reduced the value of Windows unless it anticipated
that doing so would create or increase monopoly power and thereby enable it to earn
greater monopoly profits.
- Professor Fisher testified: "Microsoft has, I think, plainly
taken actions which only make sense if they believe that
they have a monopoly to protect. Those are, of course, the
actions which are in large part the subject of this case."
Fisher, 6/1/99am, at 12:14-17.
- Dean Schmalensee conceded that, if a firm can impose a
tie-in "that implies the firm has some power over price."
Schmalensee, 1/19/99am, at 40:12-22. Dean Schmalensee
also previously wrote that: "Evidence that competitors have
conspired to fix prices or divide markets is treated as very
good evidence that these competitors have market power"
(GX 1514), and that such evidence "perhaps" could indicate
"monopoly power." Schmalensee, 1/14/99pm, at 46:14 -
47:6.
- Dr. Warren-Boulton testified that "to the extent there is
evidence . . . which shows that Microsoft has . . . used its
position in the operating system market to exclude
competitors from either that market or from markets that
might facilitate the entry of a firm into that market, then that's
direct evidence of the ability to exclude" and "that by itself is
direct evidence of the existence of monopoly power."
Warren-Boulton, 12/1/98am, at 32:3-20.
B. Microsoft's monopoly power is also demonstrated by a structural
analysis
17. Microsoft's monopoly power is confirmed by a traditional structural analysis,
which shows that Microsoft possesses a dominant share of a well-defined market
protected by immense barriers to entry.
- Professor Fisher testified that "Microsoft's high market share is an
indication that it possesses monopoly power. The analysis of barriers to
entry confirms that monopoly power exists." Fisher Dir. ¶ 65.
- Dr. Warren-Boulton likewise testified that Microsoft "possesses monopoly
power" because it "for several years has enjoyed, and is projected for
several years to retain, a market share in excess of 90%," and this share
"is protected by substantial barriers to entry." Warren-Boulton Dir. ¶ 7.
17.1. The standard way to determine monopoly power is (1) to ascertain
whether a firm possesses a very large share of a properly defined market and then (2)
to determine whether substantial barriers to entry protect that share by impeding the
ability of rivals to enter or to expand.
- Professor Fisher testified that "the ordinary way you proceed in an
antitrust case is to define a market and look at market shares" and
then determine whether there are substantial barriers to entry.
Fisher, 6/1/99am, at 12:2-13; see also Fisher, 6/1/99am, at 6:1-3
(explaining that this is the "standard way" to determine monopoly
power); Fisher Dir. ¶¶ 32-39 (testifying that "monopoly power is
conventionally addressed by defining 'the relevant market' and
assessing shares in the market share"); Warren-Boulton Dir. ¶¶ 18,
42-44.
- Dean Schmalensee conceded that: "'The traditional and most
common approach in an instance where one can define a relevant
market in the antitrust sense'" is "'to first look at shares of that
arket and then if shares are large, to move on to consider
conditions of entry.'" Schmalensee, 1/13/99pm, at 24:9-25 (quoting
GX 1526 (Schmalensee's testimony in Bristol)).
17.2. A large share of a well-defined market protected by substantial
entry barriers warrants an inference of monopoly power.
- Professor Fisher testified: "A large share of a properly defined
market" is indicative of the ability to exercise substantial market
power, and that where "there are significant barriers to entry,
monopoly power can be present." Fisher Dir. ¶¶ 32-36, 39.
- Dean Schmalensee conceded that, if Microsoft's Windows
operating system enjoys the protection of substantial barriers to
entry, then he could not conclude that Microsoft lacks monopoly
power. Schmalensee, 1/14/99am, at 8:22 - 9:9.
- Dr. Warren-Boulton testified "that market share is an indicator of
monopoly power. It is one of several indicators of monopoly
power." Warren-Boulton, 11/19/98am, at 56:22-23.
1. Operating systems for Intel-compatible PCs comprise a relevant
market
18. The purpose of defining markets is to determine whether substantial and
durable market power can be exercised; accordingly, a properly defined relevant market
should include the set of products over which a single firm, if it controlled production of
those products, could exercise substantial market power.
- Dean Schmalensee testified that a relevant market consists of the
"smallest aggregate that could be profitably monopolized." Schmalensee,
6/24/99pm, at 58:15-23.
- Dr. Warren-Boulton testified that a properly delineated antitrust market
includes the set of products over which a single firm, if it controlled
production of those products, could exercise substantial market power.
Warren-Boulton Dir. ¶¶ 26-32.
- Professor Fisher testified that the purpose of defining a market is to
determine the "set of things that could constrain the power of the alleged
monopolist." Fisher, 6/1/99am, at 9:17-24.
18.1. The relevant market thus should include only reasonable
substitutes that in a reasonable period of time could constrain -- and thus defeat -- an
attempt to exercise substantial market power.
- Professor Fisher testified that a relevant market "should include all
those products that reasonably serve to constrain the behavior of
the alleged monopolist." Fisher Dir. ¶ 32; Fisher, 6/1/99am, at
9:18-21 (stating that "in defining a market and then in examining
market power, you typically look at . . . things that could constrain
the power of the alleged monopolist.").
- Dr. Warren-Boulton also testified that a relevant market should
include substitute products that could prevent the exercise of
monopoly power. Warren-Boulton Dir. ¶¶ 27-28. He further
testified that it is "important not to define the market too broadly" by
including products that are not reasonable substitutes, "for that
might understate the power of the firm whose conduct is being
examined." Warren-Boulton Dir. ¶ 28.
18.2. These include:
18.2.1. Demand responses. The relevant market should include
products to which consumers could switch, without substantial difficulty, in response to
an attempt by firms in the candidate market to exercise substantial market power.
- Professor Fisher testified that, in defining a market, one
must look at "demand substitutability," which "concerns the
question of what are the products or the firms to which the
alleged monopolist's customers could readily turn in the
event of an increase in price." Fisher, 6/2/99am, at 69:22 -
70:1; Fisher, 6/1/99am, at 9:21-24 ("demand substitutability"
refers to "the set of products to which customers can turn in
the event of an attempt to earn supernormal profits" by the
alleged monopolist); Fisher Dir. ¶¶ 32-33 (same).
18.2.2. Supply responses. The relevant market should also
include firms that do not presently produce the product in question or a reasonable
substitute for it but which, without substantial difficulty, could do so in response to an
attempt by firms in the candidate market to exercise substantial market power.
- Professor Fisher testified that, in defining a market, one
must look at "supply substitutibility," which "refers to the
ability of firms who do not now produce demand
substitutable products, easily to produce demand
substitutable products." Fisher, 6/2/99am, at 70:9-11;
Fisher, 6/1/99am, at 10:8-13 (same); Fisher Dir. ¶¶ 32, 34
(same).
19. Operating systems for Intel-compatible PCs comprise a relevant market
because they lack good substitutes; that is, there are no substitutes that in a
reasonable period of time could defeat -- i.e., render unprofitable -- an attempt by a
monopolist over such operating systems from exercising substantial market power.
19.1. Other "platform" products, such as Internet browsers and Java, are
not good substitutes for operating systems because they cannot function without an
operating system.
- Jim Barksdale testified: "I am not suggesting that the browser is a
replacement for the operating system; Navigator still needs an
operating system, such as Windows 98, running underneath it, but
Navigator can and does serve as a platform for certain network-centric applications." Barksdale Dir. ¶ 82; Barksdale, 10/20/98pm,
at 72-74 (Barksdale testified that while Netscape could serve as a
substitute for certain platform chacteristics, he does not believe that
Netscape could seriously substitute for all platform characteristics).
- James Clark, founder and former Chairman of Netscape, testified that:
"Netscape is not an operating system. It's not even a networked operating
system. . . . Netscape was developing a platform. A platform is not the
same as an operating system. . . . The idea was to make it independent of
the Microsoft operating system, but no attempt whatsoever to displace the
Microsoft operating system." Clark Dep. (7/22/98) at 44:25 - 46:16 (DX
2562). Clark explained Netscape intended to provide a software layer that
would run on top of otherwise incompatible operating systems and enable
them to use network or web based applications, but that "layer still relied
on there being some kind of machine and some kind of operating system
underneath." Clark Dep. (7/22/98) at 48:5 - 49:4 (DX 2562). Clark
categorically denied that Netscape intended for the browser to replace the
operating systems that it relied upon. Clark Dep. (7/22/98) at 48:5 - 50:4
(DX 2562).
- Netscape's Richard Schell similarly testified that Netscape intended to be
"operating system agnostics," (i.e., work well with all operating systems),
but not to replace operating systems. When Microsoft counsel followed
up by asking whether he regarded "the notion of Navigator replacing
Windows [as] a slightly ridiculous assertion," Schell explained:
"There are 14 million lines of code in Windows 9X. They must do
something. For us to have thought that we would replace all of
those would have been a stretch of the imagination. We thought
we could provide functionality that enhanced not only Windows but
Unix and the Macintosh and . . . for some developers and some
users, that would become their primary environment, but we would
never think that that meant we were replacing Windows." Schell
Dep. (9/15/98), at 103:17 - 104:22 (DX 2587).
- Dean Schmalensee testified that he is not aware of any "software
that only browses and does not do anything else and requires no
other software to run." Schmalensee, 6/23/99am, at 53:2-10; id. at
57:14-17 (same for other "web-based applications").
- Professor Fisher testified: "In the present case, the growth of the
Netscape browser or the widespread use of original Java might
have perfectly well have broken down the applications barrier to
entry and allowed other operating systems to compete. But it
would be the other operating systems that were then in the market,
not . . . either Netscape, the browser market, or Sun because of
Java." Fisher, 6/1/99am, at 18:5-11.
19.2. Intel-compatible server operating systems are not good substitutes
for Intel-based PC operating systems because they lack the features and breadth of
applications users demand and are prohibitively more expensive.
- Sean Sanders of Novell testified that server operating systems do
not compete with Windows. Sanders Dep., 1/13/99pm, at 184:13 -
185:1. He further explained that to convert Novell's server
operating systems into desktop operating system would require
starting "all over again" and building the operating system "from the
ground up." "It is not easily transferable to" the desktop "role at all."
DX 2584.
- Sun's Brian Croll testified that Sun's Solaris operating system does
not compete with Windows. Croll Dep. (played 12/15/98pm), at
56:23 - 57:13.
- Ron Rassmussen, of Santa Cruz Operating Systems, testified:
"People are not purchasing our operating system as a desktop as
much as they did at one time" and that it is "more effective for our
strategy to move into a purely server role." DX 2581 (testifying that
using SCO's operating system for desktop use is prohibitively
expensive for users).
- Paul Maritz agreed "that the applications you find on a server are
different from those you find on an Intel PC acting as a desktop."
Maritz, 1/27/99pm, at 28:18 - 29:1.
- Dr. Warren-Boulton testified that "Intel-compatible operating system
products that are designed . . . to operate 'servers' are not viable
substitutes for a desktop operating systems" because they "are
generally more expensive yet do not provide the features
consumers demand when they purchase PC operating systems."
Warren-Boulton Dir. ¶ 40.
19.3. Nor do other devices, which run other (non-Intel-compatible)
operating systems, constrain the exercise of substantial market power over Intel-compatible PC operating systems.
19.3.1. A PC operating system accounts for only a very small
percentage of the cost of a PC system; therefore, even a substantial increase in the
price of a PC operating system above competitive levels will result in only a trivial
increase in the cost of a PC computer system to users.
- Maritz testified that the Windows royalty is "less than 5% of
the price of a typical new computer." Maritz Dir. ¶¶ 21, 132.
- Professor Fisher testified that a 10% increase in the price of
a PC operating system will result in only approximately 1 %
increase in the price of PC. Fisher, 6/1/99am, at 27:7-25.
- Dr. Warren-Boulton similarly testified that "even a 10%
increase in the price of the OS would result in at most a 1%
increase in the price of even inexpensive PCs." Warren-Boulton Dir. ¶ 37.
19.3.2. A common-sense economic analysis, therefore, shows that
users will not in significant numbers incur the substantial costs of switching away from
Intel-based PCs, and hence from Windows, in response to even a large increase in the
price of the operating system.
- Professor Fisher testified that the "[q]uestion at issue in
assessing Microsoft's power is not whether a change--an
increase in the price of the P.C. as a whole would cause
people to turn to other non-P.C. devices, or for that matter,
to Apple," but rather "whether an increase in the operating
system price will cause that to happen." Fisher, 6/1/99am,
at 27:1-6. He then concludes that it will not because even a
10% increase in the price of the operating system would
result in "less than a 1 percent increase in the P.C. price."
Fisher, 6/1/99am, at 27:14-16.
- Dr. Warren-Boulton similarly observed that "even a 10%
increase in the price of the OS would result in at most a 1%
increase in the price of even inexpensive PCs," and in light
of "the cost to users of switching to another platform, such a
small increase in the price of the PC platform would not be
expected to result in a large reduction in the demand for
PCs, and thus for PC operating systems." These facts led
him to conclude "that PC operating systems are a separate
market." Warren-Boulton Dir. ¶ 37; see also Warren-Boulton, 11/23/98pm, at 8:20-25, 9:17-25.
19.3.3. The evidence confirms that a substantial price increase for
PC operating systems (a trivial increase in the price of the PC) will not result in
switching away from PC systems, and hence PC operating systems, sufficient to make
the substantial price increase in the operating system unprofitable.
19.3.3.1. OEMs. As explained, OEMs will not switch away
from Windows (let alone start building other types of personal computers) in response
to a substantial exercise of market power (such as increased restrictions or prices) over
Intel-compatible PC operating systems.
- See supra ¶ 15.1.
19.3.3.2. Apple. The most obvious possible substitute for
users are other personal computers, such as Apple's Macintosh. But even Apple -- the
closest substitute to PCs -- does not constrain the exercise of power over operating
systems for Intel-based PCs.
- Dean Schmalensee conceded that Microsoft's
present operating system competitors, including
Apple, are not "the primary constraint on Microsoft's
pricing." Schmalensee, 1/14/99am, at 24:16-25.
- Although some users do switch from PCs to the
Macintosh, Apple's Avadis Tevanian testified that
Apple still cannot gain substantial share and,
therefore, cannot effectively compete with Microsoft.
Tevanian, 1/4/99pm, at 9:20 - 12:18.
- Plaintiffs' economists testified that consumers'
switching from PCs to the Macintosh is not the result
of the exercise of market power over PC operating
systems and, therefore, does not show an effective
constraint on Microsoft's ability to exercise substantial
market power. Warren-Boulton, 11/23/98pm, at 6:18
- 15:12; see also Fisher Dir. ¶ 137 ("Apple represents
the main potential alternative to desktop PCs running
Microsoft's Windows. (Although that alternative is not
sufficient to keep Microsoft from having monopoly
power.)"); Warren-Boulton, 11/23/98pm, at 8:20-25
(testifying that if the cost of the Windows operating
system increased "by a small but significant amount .
. . not enough people are going to decide . . .to switch
to the Mac platform" to include Mac in the market).
Switching to the Macintosh simply means the value of
Microsoft's monopoly is shifting, not that its monopoly
power is dissipating. Warren-Boulton, 11/23/98pm, at
13:3 - 15:12 (testifying that the question is "'what is
the constraint on the monopoly pricing of the
operating system'" and that the "fact that demand for
the product, as a whole, is increasing or decreasing is
not the relevant question'").
19.3.3.3. Other information appliances. There is similarly
no evidence that other information appliances constrain Microsoft's ability to exercise
substantial market power over operating systems for Intel-compatible personal
computers.
19.3.3.3.1 First, most such appliances are complements to,
rather than substitutes for, personal computers, so switching is not likely.
- Steve Case stated publicly and testified that:
"It's hard to imagine that PCs won't be the
dominant way people connect with the Internet
for many years to come, and Microsoft has a
pretty amazing lock on that business. . . . Other
devices will emerge, but I doubt any will
challenge Windows." Case Dep. (played
6/4/99am), at 44:17 - 45:4; Ct. Ex. 1.
- AOL's Barry Schuler testified:
- redacted - Schuler Dep., 5/5/99, at 183:18-21 (sealed).
- redacted - Schuler
Dep., 5/5/99, at 183:24 - 184:12 (sealed).
- Professor Fisher testified that other devices
are not presently good "substitutes for PC's.
And you can perfectly well have a monopoly in
operating systems for PC's, despite the fact
that there are or may be a number of operating
systems for hand-held devices, TV set-top
boxes and so on." Fisher, 1/12/99am, at 7:14-16; Fisher, 1/12/99am, at 7:19 - 8:7. Professor
Fisher further testified that other information
appliances do not presently constrain
Microsoft's behavior. Fisher, 6/2/99am, at
83:20-23.
- Bill Gates stated that for "most people at home
and at work, the PC will remain the primary
computing tool; you'll still want a big screen
and a keyboard" for many applications and
"you'll need plenty of local processing power
for graphics, games, and so on. But the PC
will also work in tandem with other cool
devices. You'll be able to share your data--files, schedule, calendar, email, address book,
etc.--across different machines; and you wont
have to think about it; it will be automatic." GX
2059 (Newsweek article dated 5/31/99). In a
similar vein, the IDC forecasts that for PCs and
other information appliances, there will be
"some competition between these two
categories of devices. However, it is more true
that the two devices will help lift each other.
As a rising tide raises all ships, the growth of
the Internet as an important tool for
communication, commerce, and entertainment
will provide ample justification for both form
factors." DX 2423, at 35.
- See also infra Part VII.D.C.3; ¶ 396.2.
19.3.3.3.2 Second, even if other information appliances
became better substitutes for a wider range of PC functions in the future, a small
increase in the price of PC systems caused by a large increase in the price of the
operating system will not result in substantial switching to other information appliances.
In other words, while other information appliances may affect relative ubiquity of PCs,
and thus the value of Microsoft's monopoly over operating systems for Intel-based PC
operating systems, those appliances do not undermine the fact that there is a market
for such operating systems that is capable of being monopolized.
- Dr. Warren-Boulton testified that a small
increase in the price of the overall computer
system will not induce large numbers of users
to incur the costs required to switch to other
devices. Warren-Boulton, 11/23/98pm, at
14:16-23; Warren-Boulton Dir. ¶¶ 37-39.
- Professor Fisher testified that, for this reason,
the existence of other information appliances
was "basically totally irrelevant" to the
monopoly power analysis. Fisher, 6/3/99pm,
at 65:1-7. "The fact that other devices are
going to be important, too, is interesting, but
we're not talking here, by the way, about a
monopoly of PCs themselves. We're talking
about a monopoly of operating systems for
PCs, and to believe that this has something to
do with eroding Microsoft's monopoly power in
operating systems, you would have to believe
that small changes in the price of the operating
system for PCs would cause people no longer
to buy PCs, but to ship" "these other devices."
Fisher, 6/3/99pm, at 65:23 - 66:6. See also
Fisher, 6/1/99am, at 27:14-22.
19.3.3.3.3 Third, because the issue for market definition is
whether a non-trivial increase in the price of the operating system will cause switching
away from PC operating systems (to other information appliances running other
operating systems or otherwise) to a sufficient extent to render that price increase
unprofitable, there is no need to reach the question of whether PCs themselves
comprise a relevant market (that is, whether a large price increase in the cost of a PC
would be rendered unprofitable by switching).
- Fisher, 6/2/99pm, at 30:2-13; 6/3/99pm, at
65:23 - 66:6.
20. Microsoft internal documents and the testimony at trial of its witnesses also
support delineating a market for Intel-based desktop operating systems.
- Joachim Kempin testified, Microsoft tracks the share of "[o]perating
systems for Intel PC[s]." Kempin, 2/25/99pm, at 94:24 - 95:7.
- Microsoft internal documents analyze as "competition" other "x86 Os[s]" --
that is, other Intel-based operating systems -- but do not characterize as
competition other types of operating systems. GX 401.
2. Microsoft possesses a dominant, persistent, and increasing
share of the market for operating systems for Intel-compatible
PCs
21. Microsoft possesses a dominant, persistent, and increasing share of the
relevant market.
21.1. Microsoft presently enjoys a market share in excess of 90%.
- Data sponsored by Professor Fisher and Dr. Warren-Boulton
shows that Microsoft's share of Intel-based PC operating systems
is well over 90%. GX 1.
- Professor Fisher testified: "Microsoft's share of personal computer
operating systems is very high and has remained stable over time.
Microsoft's worldwide share of shipments of Intel-based operating
systems has been approximately 90 percent or more in recent
years . . . . Even if operating systems for non-Intel-based
computers are included in the market definition, Microsoft's share is
still very high and stable." Fisher Dir. ¶ 64.
21.2. This share, which Microsoft has possessed since at least the early
1990s, has been stable through the many changes that have occurred in the computer
industry.
- Dr. Warren-Boulton testified: "This high market share has been
remarkably stable." Warren-Boulton Dir. ¶ 45.
- Data sponsored by Professor Fisher and Dr. Warren-Boulton
shows that Microsoft's share of Intel-based PC operating systems
is projected to rise to 96% by 2001. GX 1.
- Professor Fisher testified: "Here, Microsoft's share of the P.C.
operating systems business has been high and stable for some
years. Further, it's expected that it will remain high for some
years." Fisher, 6/1/99am, at 12:2-8.
- Microsoft North America FY96 Reviews, an internal financial report
compiled in June 1996, reported that the - redacted - GX 402, at
MS6 6001734 (sealed), GX 403, at MS6 6006356 (Microsoft North
America FY97 Reviews) (sealed).
21.3. Microsoft's share is projected to rise even further in the next
century.
- Rational Software "believes its continued success will become
increasingly dependent on its ability to support the Microsoft
platform, including Windows 95, Windows 98, and Windows NT
operating systems." GX 1663 (SEC 10-Q), at 5. Mike Devlin, a
Microsoft witness, testified that Rational's "increased dependence"
on Microsoft will indeed be the result of "the increasing market
share of the Microsoft platform." Devlin, 2/4/99am, at 25:22 - 26:1;
Devlin, 2/4/99am, at 14:8 - 15:9.
- IBM's John Soyring testified that Microsoft's 92% market share will
"stay that high, if not get higher" in the next two or three years.
Soyring, 11/18am, 71:24 - 72:4.
- Professor Frank Fisher testified: "Here, Microsoft's share of the
P.C. operating systems business has been high and stable for
some years. Further, it's expected that it's going to remain high for
some years." Fisher, 6/1/99am, at 12:2-8.
- Dr. Warren-Boulton testified that Microsoft's share of operating
systems "has been above 90% since at least the early 1990s and
this dominance is forecast through at least 2001." Warren-Boulton
Dir. ¶ 45; see also Warren-Boulton, 11/19/98am, at 57:24 - 58:5
(referring to GX 1, which contains the IDC's "projections of
continuous and sustained and increasing market shares").
- A report prepared for Microsoft in September 1997 states: "Win32
penetration by household primary machines is currently 70% and
projected to reach 90% by December 1998." GX 447, at MS7
001195.
22. Precise calculation of Microsoft's market share or of the contours of the
market is, in any event, unimportant.
22.1. Even if one included in the market other products -- such as
"middleware" and other operating systems -- Microsoft would still possess monopoly
power.
- Dr. Warren-Boulton testified that "even if the market were defined
more broadly to include operating system products for all personal
computers--such as those offered by Apple or some vendors of
UNIX based operating systems that do not use an Intel-compatible
microprocessor--my conclusion that Microsoft possesses monopoly
power in a relevant market would still stand." Warren-Boulton Dir.
¶ 41.
- Professor Fisher similarly testified that even "if operating systems
for non-Intel-based computers are included in the market definition,
Microsoft's share is still very high and stable." Fisher Dir. ¶ 64.
- Professor Fisher testified that Microsoft possesses monopoly
power even if threats to its monopoly power, such as Netscape and
Java, are included in the relevant market. Fisher, 6/2/99am, at
61:11 - 62:10; 6/1/99am, at 46:12 - 47:19.
22.2. Market definition and calculation of market shares are intended only
to aid in determining whether a firm has monopoly power, so precise calculation is not
necessary where refinement and precision will not change the ultimate determination of
monopoly power.
- As Professor Fisher testified, "there will often be no bright line
between defining products as in the market" and "leaving them out
while remembering that firms that do not produce them can enter
fairly readily. But the lack of such a clear line will not matter, so
long as one remembers that market definition need not be precise
and that its purpose is to assist in analyzing the constraints on the
behavior of the alleged monopolist." Fisher Dir. ¶ 36; see also
Fisher, 6/2/99am, at 57:19 - 59:1 (discussing Fisher,
"Microecomomics: Essays in Theory and Applications" (DX 2487)).
3. Microsoft's dominant market share reflects monopoly power
because its position in operating systems is protected by high
barriers to entry
23. Microsoft's dominant market share reflects monopoly power because that
share is both the source of, and protected by, immense entry barriers that prevent rivals
from entering or expanding.
a. Definition of barriers to entry
24. An entry barrier is any factor that permits firms already in the market to earn
returns above the competitive level without inducing entry or expansion that would
erode those returns.
- Professor Fisher testified that a barrier to entry "permits the incumbent
firms" to "earn supernormal profits without having their business bid away
by the expansion of competitors or the entry of new firms." Fisher,
1/6/99am, at 52:20-23; Fisher, 6/1/99am, at 47:20-24.
- Dean Schmalensee characterized as consistent with his definition of an
entry barrier "'any factor that permits firms already in the market to earn
returns above the competitive level while deterring outsiders from
entering.'" Schmalensee, 1/14/99am, at 6:17 - 7:19 (quoting Areeda &
Hovenkamp).
b. The applications barrier to entry protects Microsoft's
dominant position in operating systems
25. The principal barrier to entry into operating systems is what has been
termed in this case the applications barrier to entry.
- Professor Fisher testified that the "dominant position of Microsoft's
operating system is protected by the applications programming barrier to
entry." Fisher Dir. ¶ 82; Fisher, 6/1/99am, at 48:4-11.
- Dr. Warren-Boulton testified that "the applications barrier to entry sustains
Microsoft's dominance, critically contributes to its monopoly power, and
helps explain why other Intel-compatible operating systems, such as OS/2
and Linux, have persistently small market shares." Warren-Boulton Dir. ¶
56.
25.1. The applications barrier to entry results from a chicken-and-egg
problem: Users will not in large numbers use an operating system other than Windows
unless it supports a set of applications comparable to the set of applications available
for Windows, but ISVs will tend not to write comparable applications for other operating
systems in large numbers because those operating systems lack a large number of
users.
- Avadis Tevanian testified that Microsoft's dominant position rests in
part on "a commercial symbiosis that exists between application
programs and the computer operating systems on which those
programs run. An application program is condemned to
commercial failure if it will not operate reliably on the operating
system of a sufficiently large installed base of computer systems.
Similarly, the commercial viability of an operating system is critically
dependent on the availability of application programs . . . ."
Tevanian Dir. ¶ 15.
- Dr. Warren-Boulton testified that as "an operating system gains
popularity, the incentive to develop software for the operating
system increases because the larger number of users for the
operating system product implies a greater potential market for
software developers. The development of yet more applications for
that operating system, in turn, increases the value of the operating
system to end users who, as explained, purchase operating
systems in significant part based upon the quality and variety of
applications available for it." Warren-Boulton Dir. ¶ 53.
25.2. In other words, Microsoft's very large market share and installed
base of users -- which create incentives for ISVs to write first and foremost to Windows
rather than to other operating systems -- are themselves the source of an immense
entry barrier that keeps the share of operating system rivals low and protects
Microsoft's monopoly power.
- Professor Fisher testified that "Microsoft's high market share leads
to more applications being written for its operating system, which
reinforces and increases Microsoft's market share, which in turn
leads to still more applications being written for Windows than for
other operating systems, and so on." Because of this pattern,
Microsoft's "share is not likely to be eroded by new entry as long as
the applications programming barrier to entry remains strong."
Fisher Dir. ¶ 70.
- Dr. Warren-Boulton testified that "an operating system product can
rise to dominate the market, and once that dominance is achieved
maintain it, because of both the large number of complementary
software applications available for it and the flow of new
applications that are written to it." Warren-Boulton Dir. ¶ 54.
(1) Microsoft possesses a dominant market share
because software developers have powerful
incentives to write applications first and foremost
to Windows
26. The economic factors that create incentives to write applications first and
foremost to Windows, and reinforce Microsoft's dominant market share, have three
aspects.
26.1. First, Microsoft has a dominant share of PC operating systems
because a much greater breadth, depth, and number of applications run on Windows
than on other operating systems.
26.1.1. Users demand operating systems in order to run
applications; and the greater the number, variety, and quality of applications available
for a particular operating system, the greater the demand for that operating system.
- In a Microsoft marketing plan entitled "Winning @ Internet
Content" dated June 22, 1996, Andrew Wright wrote,
"Microsoft's success to date as a platform company has
primarily been driven by the availability of compelling
applications for Microsoft operating systems. Operating
systems, including Windows 95, Windows NT etc, are a
means to an end and not an end in themselves. End users
buy computers and operating systems to run applications."
GX 407.
- Microsoft's Chris Jones wrote in August 1995 that: "While
there are many factors which determine an OS purchase,
fundamentally consumers purchase the system that runs the
cool applications first and best." GX 523, at MS98 0103654.
- Avadis Tevanian testified that "the commercial viability of an
operating system is critically dependent on the availability of
application programs--including well-accepted, broadly-used
application programs--that are written for use on that
system." Tevanian Dir. ¶ 15.
- Microsoft admitted in its Answer that the "popularity of an
operating system is to some extent a function of the number,
variety, and quality of applications available to use with that
operating system . . . ." Answer ¶ 58.
- Microsoft's pricing decisions reflect the fact that Windows is
demanded precisely because of the number of applications
written for Windows. Kempin testified that "competitors are
producing, essentially . . . inferior-type products" because
"the number of applications written for [Windows] is so huge"
is an observation of the "result of the applications barrier to
entry, and it's a fairly clear statement." Kempin, 2/25/99pm,
98:15-99:5 (quoting Kempin's deposition, 21:20-22:6, 22:19-24). This, Professor Fisher explained, is exactly what one
would expect Kempin, a non-economist, to say rather than
saying "I am protected by the applications barrier to entry
and so, I have freedom as to pricing." Fisher, 6/1/99pm, at
5:15 - 6:5.
26.1.2. Applications written for one operating system generally do
not run on another because each operating system has its own, unique set of
application programming interfaces ("APIs") to which applications are written.
- Because operating systems have different APIs, "software
applications written for one operating system will not run well
on any other operating system." Barksdale Dir. ¶ 71.
- See also Soyring Dir. ¶¶ 6-7 ("For an application to operate
properly on an operating system, it must be designed to
work" with that operating systems's APIs.); Gosling Dir. ¶ 12
(testifying that applications are largely "platform-specific");
Tevanian Dir. ¶ 12 ("Application programs must be
developed so that they are compatible with the APIs of the
underlying operating system. For example, Microsoft's
popular word processing program, Word for Windows, will
run on the Windows operating system; it cannot run on the
Mac OS operating system.").
26.1.3. A vastly larger number of applications are written for
Windows than the number written for other operating systems.
- There are "tens of thousands" of applications that run on
Windows. Martiz, 1/25/99pm, at 22:10-13; Rose,
2/17/99pm, at 24:24 - 25:9 (testifying that there are over
70,000 applications available for Windows).
- According to Microsoft's own economic expert, the number
of applications available for other operating systems is at
least an order of magnitude lower. DX 2098, at E2
(reporting that approximately 12,000 applications are
available for the Macintosh, 900 for BeOs, and 250 for
Linux).
26.1.4. As a result, Microsoft has a dominant share of the installed
base of operating system users and of the operating system market.
- Microsoft's Brad Chase explained, "Content drives systems.
Windows won the desktop OS battle because it had more
applications earlier than any other platforms." GX 510, at
MS7 004130.
- Microsoft's Ben Slivka testified that "an advantage Windows
has today in the marketplace and why customers prefer
Windows today over Macintosh OS or some other operating
systems is that there are a large number of applications that
customers need . . . that are available primarily on Windows
or have their best expression on Windows." Slivka Dep.,
1/13/99, at 717:22 - 718:4.
- Microsoft's own witness, Compaq's John Rose, conceded
that the huge number of applications available for Windows
relative to other operating systems is "certainly the prime
reason" why Compaq lacks a commercially viable alternative
to Windows. Rose, 2/17/99pm, at 19:21 - 20:20. As Rose
elaborated (Rose, 2/17/99pm, at 24:24 - 25:9):
Q: Now, is it fair to say that the absence of any other
operating system that can run those 70,000
applications or any predominant chunk of them is a
prime reason why you believe there is not at present
commercially viable alternative to Windows?
A: Yes, that is part of it.
Q: Okay.
A: The fact that other operating environments do not
support that rich set of applications which are being
utilized by hundreds of millions of personal computer
users.
- Joachim Kempin testified that he didn't consider other
operating systems in setting the royalty for either Windows
95 or Windows 98 because "'the simple fact that the number
of applications, peripheral devices, support on that platform,
basically, is so huge that the benefits for buying into that
platform is huge'" Kempin, 2/25/99pm, at 98:18 - 99:5
(quoting Kempin's deposition). As Professor Fisher testified,
Kempin's testimony reflects Microsoft's perception that it is
"protected by the applications barrier to entry." Fisher,
6/1/99pm, at 5:13 - 6:5.
- Packard-Bell's Mal Ransom testified: "There are
appropriate applications, be they games or education or
reference that are - that work with the operating system.
That's a major factor for us in the consumer business that
consumers can go buy solutions that match with our
operating system. And Windows has really become a
worldwide standard in that regard." Ransom Dep. (played
12/16/98pm), at 69:24 - 70:10.
- For additional evidence, see Von Holle Dep., 1/13/99, at
298:2-23 (testifying that Gateway lacks a commercially
viable alternative to Windows because "there's not enough
support in the form of applications in the marketplace to-to
run on alternative operating environments"); Tevanian,
11/4/98pm, at 11:12 - 12:18 (testifying that "it's still the case
that the predominant number of applications in the market
do not run on the Macintosh, and because of that, most
people will just refuse to buy a macintosh, they'll want safety
in the applications that are on Windows").
26.2. Second, because of the economic incentives they confront, ISVs
tend to write first and foremost to the operating system with the dominant share, which
is Windows.
- Professor Fisher testified that the principal reason "that ISVs write
for Windows first," is that "there are economies of scale and it pays
to write for the system that has the most users." Fisher, 6/1/99am,
at 54:2-5.
26.2.1. Software development is characterized by substantial
economies of scale. The fixed costs of producing software, including applications, is
very high. By contrast, marginal costs are very low. Moreover, the costs of developing
software are sunk; once expended to develop software, resources so devoted cannot
be used for another purpose.
- Paul Maritz testified that "software products can be
produced and distributed in vast quantities very rapidly.
Once a software product is created, the cost to copy is near
zero, and the product can be quickly distributed around the
block or around the world via the Internet or other networks."
Maritz Dir. ¶ 115.
- Intuit's William Harris testified that "the economics of
software development make high volume sales critical to
profitability. The fixed costs of developing software --
including, among other things, research, development,
programming and testing -- are very large and can only be
offset by high volume sales. By contrast, the variable costs
of manufacturing software once it has been developed are
quite low. Thus, it is essential for profitability of most PC-based software products that the product be compatible with
Windows. At Intuit, compatibility with Windows is so critical
that the company will focus on such compatibility even if this
requires slowing or abandoning development of software for
use with other operating systems." Harris Dir. ¶ 25.
- Dr. Warren-Boulton testified that "operating systems in
particular, and software in general, are characterized by
economies of scale. The bulk of the costs are development
costs" whereas the costs "of producing and marketing
individual copies of the product ('the marginal costs') are, by
comparison, quite small." Warren-Boulton Dir. ¶ 47.
26.2.2. The result of economies of scale and sunk costs is that
applications developers seek to sell the highest number of copies; for it is only through
selling a large number of copies (for which the marginal cost is low) that the large, sunk
fixed costs necessary to develop software can be recovered
- Harris Dir. ¶ 25.
- Professor Fisher testified that because of the "upfront costs
of writing the software" and the fact that marginal costs of
distributing it are "essentially zero," ISVs will have "a big
incentive to write for the most popular operating system and
write for it first because you have the possibility of lots of
sales, and that means your costs per sale will be very low."
Fisher, 6/1/99am, at 59:10-16.
26.2.3. This creates overwhelming incentives to write first and
foremost for Windows because writing for Windows -- the operating system with the
dominant share -- gives applications developers by far the highest expected return for
the sunk costs incurred.
26.2.3.1. An application that is written for one operating
system, like Windows, will operate on another operating system only if it is "ported" to
that system. As numerous witnesses testified, porting applications is both time-consuming and expensive.
- John Soyring testified that it took IBM "about a year
and a half to port Netscape Navigator from
Netscape's Windows implementation to OS/2, and
that was having access to the Netscape source code
and having the Netscape engineers working side by
side with us in their laboratories in California."
Soyring, 1/18/98pm, at 65:15 - 66:18; Soyring Dir. ¶
7 (porting "can be both costly and time consuming.").
- Jim Barksdale testified that "it is time-consuming and
expensive, however, to take a piece of applications software
developed for the Windows platform and port it to the OS/2
or Macintosh platform or to some other platform."
Barksdale Dir. ¶ 75.
- James Gosling testified that the "tedious process,
which is known as 'porting' software to other
platforms, dramatically increases the cost of software
programs, and consumes scarce time and resources
that could otherwise be devoted to developing
innovative applications." Gosling Dir. ¶ 13.
26.2.3.2. As a result of these factors, ISVs tend to write
applications first and foremost for the highest volume platform, Windows.
- Microsoft's Steve Ballmer wrote in July 1997: "It's
important for us to keep developer focus. And market
share is an important part of that. If you don't have
good market share, you're going to lose developer
interest." GX 679, at 8.
- iIntuit's William Harris testified that "it is essential for
profitability of most PC-based software products that
the product be compatible with Windows. At Intuit,
compatibility with Windows is so critical that the
company will focus on such compatibility even if this
requires slowing or abandoning development of
software for use with other operating systems."
Harris Dir. ¶ 25.
- Jim Barksdale testified that, because of Microsoft's
large market share, "if anybody wants to build a
product, they build it there first. You don't start a
company building for some niche operating system.
You always start with . . . the current version of
Windows . . . if you're going to be out there selling
any product, you have to be on that year's product or
you can't succeed in any reasonable way."
Barksdale, 10/27/98am, at 70:18 - 71:9; Barksdale
Dir. ¶ 73 (Barksdale explains that "ISVs looking at this
world quite sensibly write most of the software for the
platform with the widest use. That means that most
applications are written for the Windows platform.).
- Dr. Warren-Boulton testified that "market share is, . . .
overwhelmingly, the critical issue in determining . . .
developers' decisions." Warren-Boulton, 11/19/98am,
at 86:14-16; Warren-Boulton Dir. ¶ 53 (testifying that
the development of more applications for a given
operating system "increases the value of the
operating system to end users" who "purchase
operating systems in significant part based upon the
quality and variety of applications available for it." If
the operating system's market share increases, "that,
in turn, is likely to cause software developers to
devote yet more resources to writing applications for
that operating system").
- Ron Rasmussen, Vice-President of the Santa Cruz
Operation, testified at his deposition that "all the
application vendors look at market share and the
cost/benefit analysis of providing that application on
any operating system. So if it costs them more than
they believe they're going to get in revenue or if they
believe their revenue is just a trade from one
operating system to another, there's no financial
benefit for producing that application on other
operating systems." Rasmussen Dep., (played
12/15/98am), at 58:3-9.
26.3. Third, the result of the above factors is that Windows exhibits very
strong network effects that reinforce demand for Windows.
26.3.1. A network-effect is a phenomenon in which the
attractiveness of a product increases with the use of that product by others.
- Fisher Dir. ¶ 42.
26.3.2. Windows exhibits strong network effects because each
user benefits from the fact that there are a multitude of other Windows users, that
Windows has a dominant market share, and that ISVs therefore write first and foremost
to Windows. The fact that ISVs write first and foremost to Windows, in turn, reinforces
demand for Windows and thereby augments Microsoft's dominant position and
perpetuates ISV incentives to write applications principally for Windows; and so on.
- James Gosling testified that, as a result of the incentives to
write "first" and often "only" for Windows (Gosling Dir.¶ 15),
"more software applications are available for Windows
users, which makes that platform even more attractive for
customers. This, in turn, reinforces the dominance of
Windows, and leads even more developers to develop
software for Windows." Gosling Dir. ¶ 18.
- Intuit's William Harris testified: "The development of software that is
compatible with the Windows operating system itself
reinforces the dominance of Windows, because consumers
seek to purchase the operating system that is compatible
with the greatest number of software applications. In turn,
software producers want their products to be compatible with
the operating system that is most widely used by
consumers. This creates a self-reinforcing cycle (sometimes
referred to as a 'network effect'), which tends to perpetuate
and enhance the dominance of the leading operating
system." Harris Dir. ¶ 27.
- James Barksdale testified: "Because so much software is
written for the Windows platform, consumers who want to take full
advantage of their computers and to have the maximum number of
choices of applications available continue to purchase machines
with a preinstalled Windows operating system. At the same time,
the more personal computers sold with Windows operating
systems, the more ISVs continue to write applications for the
Windows platform. In other words, the sale of computers with
Windows operating systems feeds the development of software for
the Windows platform, which in turn, generates additional sales of
computers with Windows operating systems." Barksdale Dir. ¶
74.
- Professor Fisher summarized: "Microsoft's high market
share leads to more applications being written for its
operating system, which reinforces and increases
Microsoft's market share, which in turn leads to still more
applications being written for Windows than for other
operating systems, and so on." Fisher Dir. ¶ 70.
- Dr. Warren-Boulton testified that the development of more
applications for a given operating system "increases the
value of the operating system to end users" who "purchase
operating systems in significant part based upon the quality
and variety of applications available for it." If the operating
system's market share increases, "that, in turn, is likely to
cause software developers to devote yet more resources to
writing applications for that operating system." Warren-Boulton Dir. ¶ 53.
26.3.3. This self-reinforcing cycle is confirmed by the observed
market facts: Windows' market share has been, and remains, much larger than rivals;
most ISVs develop new applications first and in the great numbers for Windows; and
the continued assurance of a large, up-to-date stock of applications for Windows
ensures that users demand Windows.
- See supra ¶ 26.1.3.
- Dr. Warren-Boulton testified that "the applications barrier to
entry sustains Microsoft's dominance" and because of it "no
rival has succeeded in mounting a sustained effective threat
to Microsoft's market dominance." Warren-Boulton Dir. ¶ 56.
- John Soyring testified that "OEMs have no commercially
viable choice but to license Windows." Even though other
operating systems exist, OEMs "cannot reasonably base
their businesses on these alternatives, due, in large
measure, to the lack of applications and device support."
Soyring Dir. ¶ 11.
(2) The same factors that reinforce Microsoft's large
market share inhibit other operating systems from
challenging Windows
27. Just as Microsoft's high market share creates incentives for ISVs to develop
applications first and foremost to Windows, the absence of a significant installed base
makes it much more expensive -- indeed, prohibitively so -- for other operating
systems to ensure the availability of a sufficient set of applications to enable those
operating systems to become good substitutes for Windows.
- Professor Fisher testified that when a firm gains a large market share due
to network effects, "it will prove increasingly difficult for other firms to
persuade customers to buy their products in the presence of a product
that is widely used. The firm with a large share may then be able to
charge high prices or slow down innovation without having its business
bid away." Fisher Dir. ¶ 43.
- Dean Schmalensee agreed with John Soyring's testimony that part of the
reason for OS/2's failure was that "IBM did not have a sufficient number of
applications to compete effectively with Microsoft." Schmalensee,
1/14/99am, at 34:15-25. Similarly, Dr. Warren-Boulton observed that "IBM
has found with OS/2 that it is simply impossible to effectively compete with
Microsoft in the home computer market because of the problem that it
doesn't have enough applications." Warren-Boulton, 11/24/98am, at 53:5-8. This competition between OS/2 and Windows illustrates the operation
of network effects, in which "the firm with the largest market share
becomes larger and the firm with the smaller market share becomes
smaller." Warren-Boulton, 11/24/98am, 52:20-21.
27.1. First, contrary to Microsoft's contention that all it takes to create a
rival to Windows is applications in a few key categories (Schmalensee, 6/22/99pm, at
60:12-20; Maritz, 1/27/99pm, at 10:2 - 11:2), to provide a viable substitute for Windows,
a rival operating system would need to offer both (1) a large, diverse, and frequently
updated set of applications and (2) assurances to users that such applications will be
available in the future.
- See supra ¶ 26.2.3.
- Although both Linux and Be OS, two relatively new Intel-based PC
operating systems, support several hundred applications --
including applications in the categories users tend to use most
(such as word processing, personal finance, and browsing)--
neither, as Dean Schmalensee conceded, can effectively substitute
for Windows. (Schmalensee Dir. ¶ 107, 108). The reason, as
Microsoft's own OEM witness, John Rose, explained, is that such
"operating environments do not support that rich set of applications
which are being utilized by hundreds of millions of personal
computer users." Rose, 2/17/99pm, at 24:24 - 25:9.
- Avadis Tevanian testified that Apple -- despite having thousands of
applications, including applications in all the "categories" users
frequently employ -- cannot gain users from Microsoft because "it's
still the case that the predominant number of applications in the
market do not run on the Macintosh, and because of that, most
people will just refuse to buy a Macintosh, they'll want safety in the
applications that are on Windows." Tevanian, 11/4/98pm, at 11:12
- 12:18.
- Paul Maritz conceded that other information devices, running other
operating systems, cannot "be a real competitor" unless they
support "a wide range of applications." Maritz, 1/27/99pm, at 11:3-24 (quoting Maritz's deposition).
- Professor Fisher testified that an "entrant would have to get written
for it -- and show that there was an assurance that this would
continue -- applications of the general number and breath for
Windows, and I would suppose that for the more popular
applications, the entrant would probably need the same ones."
Fisher, 1/13/99am, 5:9-14; Fisher, 6/1/99am, at 56:2-9 (similar).
27.2. Second, Microsoft's large installed base makes it prohibitively
expensive for rival operating systems to acquire the large set of applications necessary
to compete effectively with Windows.
27.2.1. The sunk costs required for an operating system vendor
itself to create the necessary applications itself are prohibitively large.
- Dean Schmalensee conceded that no operating system
vendor will develop the necessary applications on its own.
Schmalensee, 1/14/99am, at 15:23 - 16:9.
- Professor Fisher testified that an entrant faced with incurring
significant sunk costs for an uncertain return "isn't going to
go in" because "it's going to have to battle the incumbent
and because it will have to give up these hostages to
fortune." Fisher, 6/1/99am, at 50:18-25.
- Dr. Warren-Boulton testified that "competition between two
suppliers, each with very high fixed costs and very low marginal
costs, would likely result in a decrease in prices, further reducing
the profitability of entry to the would-be entrant. Entry into head-to-head operating system competition with Microsoft thus would
be time consuming, risky, and costly; profiting from such entry
would be at best very uncertain and long in coming." Warren-Boulton Dir. ¶ 48.
27.2.2. Accordingly, in order to ensure the availability of a set of
applications comparable to that available for Windows, a potential rival would need to
induce a large number of ISVs to write to its operating system.
- Dean Schmalensee testified that the question is whether
"the ISV community, can be convinced to provide
applications programming for an alternative operating
system." Schmalensee, 1/14/99am, at 15:23 - 16:9.
- Dr. Warren-Boulton testified that to "offer a product that a
significant number of consumers wish to have installed on
their PCs," vendors of alternative "operating systems would
have to create, or induce others to create, an extensive set
of compatible software applications. This would be not
merely expensive, but also very risky." Warren-Boulton Dir.
¶ 57.
27.2.3. The cost to an entrant of inducing ISVs to write
applications for their operating system exceeds the cost faced by Microsoft when it
induced ISVs to write applications for the DOS and/or Windows operating system
because Microsoft did not face a highly penetrated market dominated by a single
competitor.
- Professor Fisher testified: "After Microsoft's victory, the cost
of pursuading ISV's to build such a stock rather than write
for Windows has got to be much more substantial than it
was for Microsoft to persuade them in the first place."
Fisher, 6/1/99am, at 53:22 - 54:1.
27.2.3.1. In deciding whether to write for a particular
operating system, an ISV will consider the return it expects from incurring sunk costs,
and that depends on the number of users it expects the operating system will have.
- Dr. Warren-Boulton testified that it is not the return if
the firm succeeds that governs investment decisions,
but rather expected return, including the risk if the
venture fails. Warren-Boulton, 11/19/98pm, at 52:11 -
53:7, 70:2 - 71:10.
- Dean Schmalensee testified that "ISVs will not write
applications software for an operating system unless
they expect enough consumers to use that operating
system." Schmalensee Dir. ¶ 100; Schmalensee,
6/23/99pm, at 59:10-22 (same).
27.2.3.2. ISVs will not in large numbers expect that a niche
(or new) operating system will succeed in competing against Windows because ISVs
face a "collective action problem": a rival operating system cannot succeed without a
large number of applications, but no individual ISV can be assured that a sufficient
number of ISVs will write all the applications necessary for rival operating systems to
succeed. As a result, each individual ISV will continue to write first and foremost for
Windows because that is what it will expect its rivals to do; and other operating systems
will therefore be unable to gain appreciable share from Windows.
- Professor Fisher testified that for a new operating
system vendor to be successful, it "takes an awful lot
of people" writing applications. But in assembling this
critical mass, there "is a collective action problem.
That is, in deciding to write for a new system, each
ISV will not take into account the fact that his action"
will have "some influence on the success of the new
operating system." Fisher, 6/1/99am, at 58:10-18.
- Tevanian testified, regarding Apple's inability to
persuade developers to write for the proposed
Rhapsody operating system: "Developers, including
Microsoft, told Apple that they were concerned that
Apple would not be able to obtain a critical mass of
application programs written to work with the new
Rhapsody APIs and that customers, accordingly,
would not buy computers containing the new
operating system." Tevanian Dir. ¶ 19.
- Dr. Warren-Boulton summarized developer
incentives: "If you think of it as a trojan horse, any
individual applications writer looks at the market for
operating systems, and he says, 'I'm writing to the PC
platform. 90, 95 percent of the people who are likely
to use my application are using Windows; and
therefore, it's worth it for me individually to make a
decision to use J/Direct.' On the other hand, if you
look at the interests of applications writers as a whole,
if they all do that, nobody will write in cross-platform
applications. So, it is a quandary. What is in the
interests of individual application writers to do may
not be in the interests of applications writers as a
group." Warren-Boulton, 11/23/98pm, at 40:2-13.
- Microsoft's Steve Ballmer wrote in July 1997: "It's
important for us to keep developer focus. And market
share is an important part of that. If you don't have
good market share, you're going to lose developer
interest." GX 679, at 8.
- William Harris explained that, because of Microsoft's
dominant market share, Intuit had "abandoned
development of Macintosh-compatible versions of
QuickBooks and has dramatically reduced
development of Macintosh-compatible versions of
Quicken and TurboTax." Harris Dir. ¶¶ 25-26.
27.2.3.3. A rival operating system vendor cannot effectively
solve this problem by paying the necessary number of ISVs to write for its operating
system because the sunk costs of doing so are massive relative to the expected return.
- Professor Fisher testified that one "might pay ISVs to
write to your operating system. That in itself is part of
the barrier to entry, that you have to pay them to turn
away from Windows." Fisher, 6/1/99am, at 55:23 -
56:1. He further testified that doing so in order to
challenge Windows was infeasible because of the
very collective action problem that prevents ISVs from
doing so on their own. He explained: "There is a
collective action problem. That is, in deciding to write
for a new system, each ISV will not take into account
the fact that his action will have something to do with
the success of . . .some influence on the success of
the new operating system, because he won't reap all
the rewards from that. It takes an awful lot of people
doing this to make it a go." Fisher, 6/1/99am, at 59:2-18.
- John Soyring testified that "Microsoft's enormous
installed base, along with the wealth of applications
and hardware device support for Windows, noted
above, makes it difficult for IBM or any other company
to successfully offer a new operating system for
desktop and mobile PCs. Any company that
attempted to do so would have to spend an enormous
amount of money and time on development,
marketing, and support." He further observed that
this "task would be easier if there were some
reasonable way to ensure that all the applications
now on Windows would run on the new product.
Unfortunately, there is not." Soyring Dir. ¶ 13.
- MCI's David Limp testified that "it would be hard to
get into the PC space" because: "There's a lot of
home-grown application development, which has
been written directly to Windows and Win--not to the
languages of the Web but Windows languages, that
unseating that is--you know, I tried it for eight years of
my life at Apple. It's just a very hard problem, and it
takes a lot of resources, and nobody has been
successful, so, I mean, just putting on your business
hat, you kind of veer to the easier problem, right?
And that's a hard problem. IBM couldn't do it. Sun is
having a tough time. Apple basically couldn't do it, so
it's an uphill battle and, and we chose to fight our
competition in an area that was more wide open that
we could define ourselves, that was--that we could
redefine the playing field." Limp Dep., 7/30/98, at
143:6-25 (DX 2576).
- James Gosling testified that "it's very difficult for a
developer to financially justify developing software for
a platform like Solaris which has very low volume.
The differential between Solaris and Windows is
something like a hundred to one, which would mean
the financial return would be about a hundred to one
different, and yet the engineering effort is about the
same." Gosling, 12/10/98pm, at 26:16 - 27:3.
(3) The persistence of Microsoft's huge market share
is itself evidence of high entry barriers
28. That Microsoft's monopoly is protected by high entry barriers is reflected in
the fact that, for the last several years, Microsoft has possessed a dominant share of
the market and other operating systems have gained no more than a trivial share of the
market.
- Dr. Warren-Boulton testified that "the applications barrier to entry sustains
Microsoft's dominance, critically contributes to its monopoly power, and
helps explain why other Intel-compatible operating systems, such as OS/2
and Linux, have persistently small market shares." Warren-Boulton Dir. ¶
56.
(4) The testimony of Apple and IBM illustrates the
strength of the applications barrier to entry
29. The experience of Microsoft's most significant operating system rivals in the
middle and late 1990s, IBM and Apple, confirms the strength of the applications barrier
to entry.
30. IBM's inability to gain widespread developer support for its OS/2 Warp
operating system illustrates how the massive Windows installed base makes it
prohibitively costly for a rival operating systems to attract applications sufficient to
substitute for Windows.
30.1. IBM in 1994 introduced its Intel-based OS/2 Warp operating
system, targeted at the consumer market, and spent tens of millions of dollars in an
effort to attract ISVs and in an unsuccessful attempt to clone part of the Windows API
set.
- Soyring testified that IBM "spent tens of millions of dollars working
with ISV's around the world . . . to try to convince them to develop"
for OS/2. Soyring, 11/18/98pm, at 58:20 - 60:1, 66:19 - 67:8.
- Soyring further testified that IBM devoted substantial resources in
an ultimately unsuccessful attempt to clone part of the Windows
API set. Soyring, 11/18/98pm, at 61:15 - 62:1.
30.2. Despite these efforts, IBM could obtain neither significant market
share nor ISV support for OS/2 Warp.
- Soyring testified that, even when "it would have made economic
sense for an ISV to port their application to OS/2, many times they
felt those programmers could be better spent building new
functions or new applications for Windows because it provided a
potential for greater economic return for them" and because "of the
larger number of . . . Windows application users." Soyring,
11/18/98pm, at 67:11-24.
- As Soyring summarized, IBM found that it was caught "in a vicious
cycle. First, the limited number and type of OS/2 applications has
resulted in a limited demand for OS/2. That, in turn, has meant that
relatively few PCs are shipped with OS/2, and that the installed
base of OS/2 is relatively small. This relatively small installed base
of OS/2 installations has further reduced the incentive for
application developers to spend the resources necessary to port
their existing applications to OS/2 and to then offer and support
them on OS/2." Soyring Dir. ¶ 9.
- OEMs -- including IBM's PC business -- will not preinstall OS/2,
and the reason is the absence of applications. Romano Dep.
(played 12/16/98pm), at 33:4-19 (Hewlett Packard has "not
seriously" considered installing OS/2); Ransom Dep. (played
12/16/98pm), at 70:11 - 71:8 (OS/2 was "trying to make a push at
the consumer market. And the big problem with it is we needed
OS/2 plus Windows because OS/2 did not have the compatibility.
OS/2 was an operating system and worked fine on the systems,
but you needed Windows for the compatibility of all the applicants.
So it didn't make any sense resource-wise -- and by resource, I
don't mean just double charging, but the resources of the machine
to have two operating systems on it."); Romano Dep. (played
12/16/98pm), at 72:5-23 (because of the lack of applications
compatible with OS/2, it was not a viable choice for Packard Bell.).
30.3. Thus, although at its peak OS/2 ran approximately 2,500
applications and had 10% of the market, IBM determined that the applications barrier to
entry was too severe to compete against Windows in the consumer segment of the
market and, for that reason, in 1996 stopped trying to convince ISVs to write to OS/2.
- Soyring Dir. ¶ 5; Soyring, 11/18/98pm, at 61:2-4.
- Soyring testified that IBM determined that it "would not be able to
compete" against Windows because the "application barrier was
just too high for us to be able to compete" by promoting "OS/2
Warp 3 to consumer users." Soyring, 11/18/98pm, at 99:22 -
100:5. Thus, he explained, in 1996 IBM stopped trying to induce
developers to write for OS/2's APIs altogether because of it's
inability to compete against Windows. Soyring, 11/18/98pm, at
93:19-21.
- Dean Schmalensee agreed with Soyring's testimony that part of the
reason for OS/2's failure was that "IBM did not have a sufficient
number of applications to compete effectively with Microsoft."
Schmalensee, 1/14/99am, at 34:15-25.
- Dr. Warren-Boulton testified that "IBM has found with OS/2 that it is
simply impossible to effectively compete with Microsoft in the home
computer market because of the problem that it doesn't have
enough applications." Warren-Boulton, 11/24/98am, at 53:5-8.
This competition between OS/2 and Windows illustrates the
operation of network effects, in which "the firm with the largest
market share becomes larger and the firm with the smaller market
share becomes smaller." Warren-Boulton, 11/24/98am, at 52:20-21.
30.4. Microsoft's contention that OS/2's failure was a consequence of
IBM's own mistakes is misplaced because it confuses the reasons for the failure of
early versions of OS/2 with the reason -- the applications barrier to entry -- that OS/2
Warp cannot gain substantial market share today.
- As Soyring testified, IBM rectified many of OS/2's problems by the
time of OS/2 Warp's release. Soyring explained that "the
reductions in size that we made in the operating system program
were such that it made it very competitive in terms of the amount of
memory that was required, so it turned out to be quite suitable, and
we had a fair amount of success initially selling the products at
least to a particular subset of the home users." Soyring
11/18/98pm, at 58:25 - 59:7.
- Microsoft suggested that OS/2 Warp failed because IBM didn't
spend enough to attract developers. Soyring, 11/18/98pm, at
92:20 - 93:1. This, however, is entirely consistent with the
applications barrier to entry. As Soyring testified, because of
Microsoft's installed base, the cost to IBM of attracting significant
developer interest was prohibitive. Soyring Dir. ¶ 13.
31. The inability of Apple effectively to compete with Windows also evidences
the operation of the applications barrier to entry.
31.1. Although Apple's Macintosh operating system supports more than
12,000 applications, that stock of applications is not sufficient to enable Apple to
substitute for Windows for a large number of users.
- Avadis Tevanian testified that "the predominant number of
applications in the market do not run on the Macintosh, and
because of that, most people will just refuse to buy a Macintosh.
They'll want safety in the applications that are on Windows. Or in
some cases they'll be required to run Windows. For example, in
almost every corporation in the world, they have to run some
specific applications that are only on Windows." Accordingly,
despite the fact that the iMac is selling well, "in the grand scheme
of things, there is still the Windows monopoly, that it's a situation
where people need to run Windows applications, and they buy
Windows computers." Tevanian, 11/4/98pm, at 11:21 - 12:13.
- Dr. Warren-Boulton testified that there are approximately 12,000
applications available for users of the Macintosh operating system,
but that Apple cannot constrain Microsoft's ability to exercise
market power. Warren-Boulton, 11/23/99pm, at 16:7-13.
31.2. The absence of a large installed base, in turn, reinforces the
disparity between the applications available for the Macintosh operating system and
those available for Windows, further inhibiting Apple sales.
- Microsoft's Paul Maritz conceded that "fewer software developers
create products for the Apple Macintosh because there are fewer
Apple Macintosh customers to buy such products." Maritz Dir. ¶
179.
- Apple's Avadis Tevanian testified that an "application program is
condemned to commercial failure if it will not operate reliably on the
operating system of a sufficiently large installed base of computer
systems. Similarly, the commercial viability of an operating system
is critically dependent on the availability of application programs--including well-accepted, broadly-used application programs--that
are written for use on that system." Tevanian Dir. ¶ 15.
Consequently, "Apple has learned through experience" that "the
symbiosis between operating system[s] and application programs
creates significant barriers to the introduction and growth of
competing operating systems." Id. at ¶ 16.
31.3. Also illustrative is Apple's inability to gain developer support for its
Rhapsody operating system in 1997.
31.3.1. Rhapsody offered users new, attractive technologies; but
taking advantage of these technologies would have required ISVs substantially to
rewrite their applications, a process requiring a substantial investment and, therefore, a
significant volume of sales to recoup.
- Avie Tevanian testified that "the biggest reason" ISVs would
not write Rhapsody applications was that "they needed to
have an economic incentive, they needed to know that they
could sell a lot of copies of their applications; and to sell a lot
of copies of their applications, they needed to know that
there were going to be lots of copies of the operating
system, and they just didn't believe that Apple had any
chance of selling a lot of copies of this operating system."
Tevanian, 11/4/98pm, at 44:5-13.
31.3.2. Developers refused to make this investment because they
did not believe that Apple could gain significant volume against Windows to make the
additional sunk costs worthwhile.
- Tevanian testified that developers "didn't see that Apple
would ever get sufficient volume on Rhaspody so that they
thought they would have an economic return on their
investment." Tevanian, 11/4/98pm, at 83:20-23.
- Tevanian explained that the Windows installed base was the
reason why developers thought Apple "had no chance of
achieving any significant volume with a new operating
system." Tevanian, 11/4/98pm, at 85:19-23.
31.3.3. Other reasons may have contributed to Rhapsody's failure
-- Apple's financial difficulties and Microsoft's refusal to support its ability to work with
Windows NT -- do not detract from the illustration Rhapsody provides of the
applications barrier to entry.
- The very document Microsoft introduced in support of its
assertion that Apple's financial distress hurt Rhapsody
shows, in fact, developer concern as to whether Apple could
gain sufficient share to make their investment worthwhile.
DX 1769 ("For Developers, the ramp for Rhapsody is not
irrelevant."); see also Tevanian, 11/4/98pm, at 96:23 - 99:23.
- The force of the applications barrier to entry is demonstrated
by the steps Apple took following Rhapsody's initial failure.
Apple incorporated some of the Rhapsody technology into
its new Macintosh operating system in a way that did not
require ISVs significantly to rewrite their applications. As
Tevanian testified, this greatly reduced the costs to
developers of supporting Rhapsody because: "The
economic model for them is very simple. They just keep
their existing investment." Tevanian, 11/4/98pm, at 91:13-21. In short, ISVs are willing to develop for Apple when they
can recoup their past investments. But because of the
Windows installed base, they are generally unwilling to make
substantial investments required "to go into new areas."
Tevanian, 11/4/98pm, at 83:2-7.
c. Other entry barriers reinforce the applications barrier to
entry
32. Although the applications barrier to entry is an important factor that prevents
other operating systems from developing into reasonable substitutes for Windows, other
factors also inhibit the ability of other operating systems to enter or expand.
32.1. Switching costs. Switching to a new operating system requires
users of existing systems to scrap existing investments in applications, training, and
certain hardware.
- Dr. Warren-Boulton testified that computer users "are reluctant to
switch from Windows to another operating system, even another
PC operating system, because to do so requires them to replace
application software, to convert files, and to learn how to operate
the new software. Often, switching also means replacing or
modifying hardware. Businesses can face even greater switching
costs, as they must integrate PCs using the new operating systems
and application software within their PC networks and train their
employees to use the new software." Warren-Boulton Dir. ¶ 49; id.
¶ 36.
- James Gosling testified that a Windows user switching to the Apple
iMac would "have to buy every piece of software all over again."
Gosling, 12/10/98pm, at 19:15 - 20:1.
32.2. Other network effects. In addition to augmenting ISVs' incentives
to write for Windows, Microsoft's high market share increases the value of Windows in
other ways. These include, among other things, common file formats and low training
costs because of user familiarity.
- Professor Fisher testified that the ubiquity of Windows "may enable
firms to avoid training costs when personnel are moved within the
firm or new personnel are hired from outside. This gives firms an
incentive to have the same user interface throughout its own
computers and the same interface that is widely used by other
firms. Other network effects include the ease of exchanging files
and the opportunity to learn from others." Fisher Dir. ¶ 67.
- Dr. Warren-Boulton testified that switchers to another platform
would "need to expend time and money learning how to use a
computer designed for a different processor. And both switchers
and new users would have to bear costs resulting from any
incompatibility or impaired compatibility between their computer
and PCs used by colleagues or others with whom the users may
wish to communicate or share files." Warren-Boulton Dir. ¶ 17.
- Dr. Warren-Boulton also testified that the applications barrier to
entry "is supplemented by other barriers to entry that derive from
network effects. Books, publications, training, user groups, and
news groups for the incumbent operating system product provide a
large sense of community for its users. Users can exchange files,
and perhaps more readily use their computers to communicate,
with other members of the group. Finally, when the incumbent
operating system is installed at work, it leads users to select the
same operating system product for use at home." Warren-Boulton
Dir. ¶ 55.
- "It's important for them to be able to leverage one web browser
class -- for example, a training session -- among all the various
users of that browser, so that, to the extent it's possible, you want
the features of that browser to look and feel and act and work the
same, regardless of whether the employee is running a Unix work
station or an Intel-based PC." Weadock, 11/17/98am, 19:25 - 20:6
(discussing GX 217, at MS98 0109146) (corporations "want a
common platform for web apps, basic end user feature similarity,
simship, and it is the number one reason corps and ISPs wait or
don't go with IE as std. browser")
32.3. Sunk costs of developing an operating system. Like other software,
developing an operating system requires incurring significant sunk costs (although
actual production costs are low), and the significant sunk costs that must be incurred to
develop an operating system deter entry.
- Dr. Warren-Boulton testified: "If you build an operating system and
you fail, you can't take the OS and do much else with it. That
money is gone. And that makes it into a very risky business. And
economists generally recognize that the higher the share of costs
that are sunk, the greater the barrier to entry into that business,
which really makes good sense." Warren-Boulton, 12/1/98am, at
31:2 - 31:8.
C. Microsoft's ability to control the price of Windows evidences its
monopoly power
33. Microsoft's monopoly power is also evidenced by its ability to control the
price of its operating systems.
- Professor Fisher testified that a firm's "substantial ability to vary, and,
indeed, to raise" price "without fearing that its customers will turn
elsewhere" can be evidence monopoly power. Fisher, 6/1/99am, at 11:14
- 12:17.
1. Microsoft does not consider rival operating systems in pricing
Windows 95 or Windows 98
34. Microsoft does not consider competitors in setting the price for Windows 98,
and Microsoft does not fear that increasing the price of Windows will cause its
customers to turn elsewhere.
- See supra Part II.A; ¶ 15.1.5.
2. Microsoft raised the prices of obsolete versions of Windows
35. Microsoft's substantial pricing discretion is also demonstrated by its ability to
increase the royalty for older versions of Windows, versions that Microsoft
characterized as "obsolete," following the release of new versions.
a. Microsoft increased the Windows 95 price when it
released Windows 98
36. Following the release of Windows 98, Microsoft,
increased the price of Windows 95 to the same level as Windows 98.
- Professor Fisher testified that
- redacted -
Fisher, 1/12/99pm, at 47:2-9 (sealed session).
- Dean Schmalensee acknowledged that Microsoft
- redacted - Schmalensee, 1/25/99am, at 51:25 - 52:12 (sealed session);
Schmalensee, 1/25/99am, at 44:22 - 45:9
(sealed session).
- Current OEM licenses list royalties for
- redacted -
See e.g., GX 461 at MS98 0009500 (IBM license) (sealed); GX
1190 at MS98 0008922 (Compaq license) (sealed). See also
Schmalensee, 1/25/99am, 51:25 - 52:7
- redacted -
(sealed session).
- The average actual price of a Windows 95 standard license
- redacted -
GX 1404
(chart of prices sponsored by Professor Fisher) (sealed); DX 2330
(chart of license dates sponsored by Dean Schmalensee) (admitted
in sealed session).
36.1. Microsoft's increase of the Windows 95 royalty to the same level as
the Windows 98 royalty is not consistent with a competitive market.
- Microsoft witnesses repeatedly asserted that
- redacted -
Schmalensee, 1/25/99am, at 15:6-18
- redacted - (sealed session); Rose, 2/17/99pm, at 26:14
- redacted -
; id. at 30:9-31:11 (same) (sealed session).
-
- redacted -
Fisher, 1/12/99pm, at 45:16-22 (sealed
session).
- redacted - Fisher, 1/12/99pm, at 46:21-22 (sealed session). If
operating systems "were a competitive market, and Microsoft didn't have
some power over price, then when the better product came out,
you would expect to see the price of the older product at least stay
the same and, quite possibly, go down, but it didn't. It went up."
Fisher, 1/11/99pm, at 43:9-13.
36.1.1. Dean Schmalensee's testimony that
- redacted - (Schmalensee, 1/25/99am, at 27:9-11 (sealed session)) is
wrong, and he ultimately acknowledged that he did not investigate whether
- Professor Fisher presented a chart showing that
- redacted - GX 1404 (sealed); Fisher, 1/11/99am, at 19:18-22
(sealed session). These figures included
- redacted - Fisher, 1/12/99pm, at 46:16-25
(sealed session)
- Dean Schmalensee presented no evidence comparing
- redacted - Indeed he conceded
- redacted - Schmalensee, 1/25/99am, at 49:21 - 51:24 (sealed session).
- Instead, Dean Schmalensee presented a chart showing
- redacted - Schmalensee, 1/25/99am,
at 31:18 - 32:7 (sealed session); DX 2332 (admitted in sealed
session).
- Dean Schmalensee also asserted that "Microsoft did not in fact
increase prices for Windows 95/98 after December 1997."
Schmalensee Dir. ¶ 164. But he himself introduced a chart
showing
- redacted - DX 2330 (sealed).
- redacted - ; e.g., GX 1190, at
MS98 000892, MS98 0008930
(sealed); compare GX 449, at MSV 0002629 (1995
- redacted - (sealed). Dean Schmalensee appeared to ground his
erroneous assertion on the fact that
- redacted - (Schmalensee, 1/25/99am, at 50:3-9)
(sealed session);
- redacted - Dean
Schmalensee conceded he did not investigate whether
- redacted - Id. at
54:21 - 55:4.
b. Microsoft used the threat of withholding discounts on
Windows 95 to double the price charged IBM for
Windows 3.1 following the release of Windows 95
37. Similarly evidencing substantial and durable market power over operating
systems for Intel-compatible personal computers is Microsoft's threat to withhold
substantial discounts for Windows 95 in order to force IBM to accept a doubling of its
royalty for Windows 3.11.
37.1. Microsoft put IBM to the choice of abandoning its favorable royalty
for Windows 3.11 or sacrificing commercially crucial MDA discounts for Windows 95.
- Garry Norris testified that IBM, in part because of its assistance in
developing the product, enjoyed a $9 royalty for Windows 3.11.
Norris, 6/7/99pm, at 8:18-23, 12:8-18; 6/8/99am, at 81:23 - 82:19;
GX 2194, at 90353. IBM's contract with Microsoft guaranteed IBM
that rate until September 1997. Norris, 6/7/99pm, at 8:18-23.
- In April 1996, Norris testified, Microsoft proposed to IBM what
Microsoft termed its "Windows desktop family agreement." Norris,
6/7/99pm, at 14:13 - 15:4. The proposed agreement consisted of
a single contract covering a number of Microsoft operating system
products, including Windows 95, Windows 3.11, and Windows NT.
Through this agreement, Microsoft conditioned substantial
discounts to Windows 95, and a license to Microsoft's newest
version of Windows NT, on IBM abandoning its favorable rate for
Windows 3.11 and accepting a much higher rate (initially proposed
at $62). Norris, 6/7/99pm, at 8:13 - 9:16, 13:16 - 14:4. IBM could
sign a Windows 95 license without giving up its favorable Windows
3.11 rate, but if it did, Microsoft would withhold MDA discounts on
Windows 95 amounting to $75 million a year. Norris, 6/7/99pm, at
9:4-9, 10:21-25.
37.1.1. Microsoft sought to raise IBM's Windows 3.11 royalty, and
thereby to migrate its installed base to Windows 95, in order to ensure Microsoft's
continued market dominance.
- Norris testified that Microsoft told IBM that it conditioned discounts
vital to the IBM PC Company's business on IBM abandoning its
favorable rate for Windows 3.11 because Microsoft "wanted more
customers to move to Windows 95, and more customers to move to
Windows NT." Norris, 6/7/99pm, at 12:1-7, 39:20 - 40:2.
- As Dr. Warren-Boulton testified, one way Microsoft reinforces the
applications barrier to entry is to "migrate" its "installed base" of
users -- those already using Windows operating systems -- to
newer versions of its operating system. Increasing the number of
Windows 95 users increases the incentives of ISVs to develop for
Windows 95, thus reinforcing the applications barrier to entry.
Warren-Boulton, 11/23/98pm, at 75:13 - 77:7.
37.1.2. Microsoft also told IBM that, even if it signed the agreement
Microsoft proposed, it would not get as good a deal as IBM's rival, Compaq, because
IBM (unlike Compaq) competed against Microsoft.
- See infra Part V.C.2.b(3); ¶ 209.2.1.
37.1.3. IBM ultimately acquiesced in Microsoft's demands and gave up its
$9 royalty for Windows 3.11 because it lacked any viable commercial alternative to
Windows 95 and the discounts Microsoft threatened to withhold were necessary in
order for IBM to compete against OEM rivals.
- Norris testified that IBM gave in to Microsoft's demands because
IBM "did not have a choice. We had no place else to go. We had
to have Windows 95 in order to be in the PC business," and
Microsoft was threatening to increase IBM's costs "by $75 million" a
year. Norris, 6/7/99pm, at 13:9-25, 40:3-15 (same).
- Norris further testified that a Microsoft account manager told IBM
that accepting these terms was the "Cost of doing business with
Microsoft." GX 2186; Norris, 6/7/99pm, at 74:20 - 75:10.
37.1.4. Although IBM was able to negotiate the originally offered $62
royalty Microsoft proposed for Windows 3.11 down to an effective royalty of
approximately $19.50, Microsoft's threatened withholding of $75 million in MDA
discounts, and its ability to increase the price charged IBM for its inferior Windows 3.11,
demonstrate substantial market power.
37.1.4.1. IBM's assent to Microsoft's demands demonstrates that
Microsoft possesses substantial pricing discretion with respect to Windows 95.
Microsoft threatened to withhold $75 million in discounts to IBM without concern that
IBM would shift its business to another operating system vendor or that charging a high
price to IBM would hasten the day when a viable alternative to Windows would arise.
- Professor Fisher testified that Microsoft's monopoly power is
evidenced by the fact that its "customers do not believe that
they have serious commercial alternatives to Windows."
Fisher, 6/1/99am, at 11:9-19.
37.1.4.2. The price IBM paid for Windows 3.11 would have
increased even more had IBM not kept its shipments of Windows 3.11 below 8% of all
Microsoft operating systems that it shipped. Microsoft's ability to change the Windows
3.11 royalty depending on the extent to which IBM facilitated Microsoft's objective of
moving users to Windows 95 is further evidence of monopoly power.
- GX 2186 (document discussing IBM's royalty payments to
Microsoft states as "Special Condition one" that "If win 3.11
vol. <8% of total volume for the contract period IBM receives
rebate of $5m + $6 per copy of all win 3.11 shipments").
- Garry Norris testified that there was a two-part agreement
under which "Microsoft offered IBM an incentive, and the
incentive was that if IBM's shipments of Windows 3.11 fell
below eight percent of its total Microsoft operating
shipments, Microsoft would rebate to IBM 5 million U.S.
dollars, which had been agreed upon in a previous
settlement agreement in 1995. The second aspect of that
was that after shipments did, in fact, fall below eight percent,
then the price for Windows 3.11 would receive an additional
$6 rebate." Norris, 6/7/99pm, at 37:10-20; see also Norris,
6/9/99pm, at 48:9 - 49:5 (same).
3. Other aspects of Microsoft's pricing of Windows are
consistent with monopoly power
38. Other aspects of Microsoft's pricing of Windows are consistent with
Microsoft's possession of monopoly power.
38.1. The increasing price of Windows. In contrast to other components
of a personal computer (where prices have substantially decreased), the price of
Windows has increased in both absolute and relative terms in the past several years.
38.1.1. The price OEMs pay for Microsoft's operating systems has
risen in absolute terms in the past several years.
- Professor Fisher testified that he has "looked at what's
happened to Microsoft's operating system price over time,
and it isn't falling, and I don't believe it's falling even on a
quality corrected basis. And for that matter, it isn't even
constant. It's rising." Fisher, 1/11/99pm, at 41:24 - 42:3;
see also GX 1404 (sealed) (chart sponsored by Professor
Fisher showing
- The royalty
- redacted - Rose, 2/17/99pm, at 30:9-18 (sealed
session).
-
- redacted - GX 1430 (sealed).
-
- redacted - See supra Part
II.C.2.a; ¶ 32.
- Kempin acknowledged that
- redacted - Kempin, 2/25/99pm,
126:5 -128:13 (sealed session); GX 1506 (sealed); GX 1508
(sealed).
38.1.2. The price OEMs pay for Microsoft's operating systems has
risen in relative terms in the past several years.
- Kempin wrote to Gates in December 1997 that the price of
Microsoft's operating systems to OEMs has increased "over
the last ten years" while "other components" of PC systems
"have come down and continue[] to come down." GX 365,
at MS7 007194.
- See Romano Dep. (played 12/16/98pm), at 33:20 - 34:21
(testifying that the prices of all components of the PC have
decreased except the operating system, the price of which
has increased); Warren-Boulton, 12/1/98am, at 26:16 - 30:9;
GX 439 (Microsoft chart demonstrating price increase); GX
1430 (chart based on GX 439) (sealed); Warren-Boulton Dir.
¶ 61.
- Compare Schmalensee, 6/23/99am, at 14:15 (testifying that
"hardware costs are falling") with DX 2301 (admitted in
sealed session) (chart sponsored by Dean Schmalensee
showing
38.1.3. Although Dean Schmalensee asserted that
- redacted - + (Schmalensee, 1/25/99am,
at 11:11 - 15:18) (sealed session), that assertion is not supported by the evidence:
- Microsoft, as explained, raised the price of Windows 95 to
the same level as Windows 98. But Microsoft did not at the
same time increase the quality of Windows 95. See supra
Part II.C.2.a; ¶ 36.1.
- As Professor Fisher testified, this relative increase in the
price of Windows "ought to at least make one suspicious" of
Microsoft's assertion that its price increases merely reflects
adjustments for increased product quality. Fisher,
1/11/99pm, at 43:14-23; Fisher, 1/13/99am, at 39:13 - 41:9.
- See supra ¶ 38.1.2 (price of operating system has increased
relative to other components of PC system).
38.2. Microsoft's pricing of its Windows 98 upgrade. Microsoft's pricing
of its Windows 98 upgrade also is consistent with Microsoft's possession of monopoly
power.
38.2.1. The evidence shows that Microsoft had substantial
discretion in setting the price of its Windows 98 upgrade product, the operating system
product it sells to existing users of Windows 95.
- A contemporaneous Microsoft study shows that it could
have charged $49 for the product -- and there is no reason
to believe that price would have been unprofitable -- but
concluded it could earn greater profits by charging $89.
Warren-Boulton, 12/1/98am, at 24:8 - 25:13; GX 1371, at
MS7 003730, MS7 003748.
- The existence of a range of prices over which Microsoft
believed it could profitably sell its upgrade product is, at a
minimum, consistent with its possession of substantial
market power. Warren-Boulton, 12/1/98am, at 24:8 - 25:13.
38.3. Microsoft's ability to price discriminate. Microsoft's sustained ability
to price discriminate is probative of market power and, therefore, consistent with
monopoly power.
38.3.1. Price discrimination is probative of the existence of market
power and, therefore, is consistent with monopoly power.
- Price discrimination is the practice of setting different prices
for the same product to different customers. Schmalensee,
1/21/99am, 30:11-16. Dean Schmalensee testified that he
continues to agree with his statement in his 1982 article from
the Harvard Law Review, that it is "a standard textbook
proposition that for a seller to practice price discrimination
profitably, it must have some control over price, some
monopoly power." GX 1514. By "some monopoly power" in
that quotation, Dean Schmalensee says he meant "market
power." Schmalensee, at 1/14/99pm, 47:7-14; see also
Schmalensee, at 1/21/99pm, 4:22 - 5:4.
- Professor Fisher testified that Microsoft's ability to price
discriminate indicates its ability to earn supranormal profits
from OEMs which do not pay the low price. Fisher,
1/11/99pm, 41:17 -23. Because monopoly power is a "high
and sustained degree of market power," Professor Fisher
testified, evidence that Microsoft has market power helps
form the basis for his opinion that Microsoft has monopoly
power. Fisher, 1/13/99am, at 26:16-22.
- Microsoft introduced an excerpt from a current economics
textbook, which states that for a firm to be able to engage in
price discrimination, the firm must have some market power.
The book further comments: "Even though all firms would
like to price discriminate, many are not able to do so." DX
2271, at page 434.
38.3.2. Microsoft engages in price discrimination by charging
different OEMs different prices for Windows.
- Professor Fisher testified that - redacted -
Fisher,
1/11/99am, at 18:14 - 19:8 (sealed session).
- Summarizing the charts he sponsored, Professor Fisher
testified that the price differences among OEMs cannot be
explained except in light of Microsoft's exercise of market
power. Fisher, 1/13/99am, 57:16 - 58:9.
- As Professor Fisher's charts show,
- redacted -
For instance, GX 1403 (sealed), GX1432 (sealed), and GX
1433 (sealed) show
- redacted -
GX 1405 (sealed), GX 1406 (sealed), GX 1407
(sealed)
- redacted - GX 1416 (sealed), GX 1417 (sealed), GX 1419
(sealed), and GX 1420 (sealed) show
- redacted - GX 1408 (sealed), GX 1409
(sealed), GX 1410 (sealed), GX 1412 (sealed), GX 1414
(sealed), and GX 1415 (sealed) show
- redacted - GX 1422 (sealed), GX 1423
(sealed), GX 1426 (sealed), and GX 1428 (sealed) show
- redacted -
- Professor Fisher testified, referring to his charts
- redacted -
Fisher, 1/11/99am,
at 20:12-18 (sealed session).
- Further, Professor Fisher testified, Microsoft's price
discrimination is part of a system which tends to increase
Microsoft's future revenues and reinforce the barriers to
entry protecting Microsoft's monopoly. Fisher, 1/11/99pm, at
44:3 - 45:13.
- redacted -
Fisher, 1/11/99pm, at 44:3 - 45:13.
- redacted - Fisher, 1/11/99pm, at 30:8-11 (sealed
session).
-
- redacted -
DX 2307.
- Dean Schmalensee repeatedly emphasized
- redacted -
DX 2306; Schmalensee, 1/25/99am, at 29:6-11
(sealed session). But his chart DX 2307 shows
- redacted -
DX 2307;
Schmalensee, 1/25/99am, at 22:17-22 (sealed session).
38.3.3. Among the five largest OEMs,
- redacted -
- According to a chart sponsored by Dean Schmalensee,
- redacted -
DX
2307.
- According to Dean Schmalensee,
- redacted -
DX 2307. In October 1997,
Gates wrote to Kempin, Microsoft's Vice President in charge
of OEM relations, "[o]verall, we will never have the same
relationship with IBM that we have with Compaq, Dell and
even HP because of their software ambitions. I could deal
with this just fine if they weren't such rabid JAVA backers."
GX 257.
-
Professor Fisher showed, focusing on the same language
mix and time period as Dean Schmalensee, that
- redacted -
GX 1432 (sealed).
- redacted - GX
1432 (sealed).
- redacted -
D. Dean Schmalensee's contrary analysis is unreliable
39. Dean Schmalensee testified that Microsoft lacks monopoly power. Refusing
to define a relevant market, Dean Schmalensee opined that Microsoft cannot be a
monopolist because it does not behave like a monopolist. Dean Schmalensee's
analysis is deeply flawed. It is based on suppositions that are contrary to both the
evidence and common sense and contradicts his prior writings and testimony.
1. Dean Schmalensee's approach to market definition is flawed
40. Dean Schmalensee testified that there is no purpose for which defining a
market in which Microsoft sells operating systems is relevant (Schmalensee,
1/13/99pm, at 37:12-22). The reasons Dean Schmalensee gave for refusing to define a
market, and his objections to the market the plaintiffs defined, are not credible and are
unreliable.
40.1. First, Dean Schmalensee testified that assessing market share is
"not helpful in an industry like software" because "entry is possible from many known
and unknown sources" and a software industry is too "dynamic" to apply the traditional
tools of antitrust analysis (Schmalensee Dir. ¶ 187). This reason for refusing to define a
market is inconsistent with the testimony Dean Schmalensee gave in the Caldera case,
his prior writings, and sound analysis.
40.1.1. In the Caldera case, in which Microsoft is being sued by a
producer of a rival operating system, Dean Schmalensee defined a market for Intel-compatible desktop operating systems -- the very market he testified here has no
purpose.
- Schmalensee, 1/13/99pm, at 29:9-14.
40.1.2. Dean Schmalensee's refusal to define a relevant market in
this case also conflicts with his prior writings.
- In a paper entitled "Diagnosing Monopoly Power in Antitrust
Cases," Dean Schmalensee wrote that "market share has
long been the legal touchstone for deciding whether a firm
has market power" and that any weaknesses in that
approach "do not make a case for abandoning the traditional
concern with market share." GX 2335, at page 1.
- In a Harvard Law Review article entitled "Another Look At
Market Power," Dean Schmalensee quoted an article by
Landes & Posner as saying that the "standard method of
proving market power in antitrust cases involves first
defining a relevant market in which to compute the
defendant's market share, next computing that share, and
then deciding whether it is large enough to support an
inference of the required degree of market power." GX
1514, at 5. Schmalensee endorsed "the basic approach of
Landes & Posner" and said that computing market share
"can provide information about the importance of market
power, but markets differ considerably and shares should be
interpreted in light of evidence on market demand elasticities
and other conditions." GX 1514, at 9.
40.1.3. Dean Schmalensee's analysis is, in any event, unsound.
Defining markets and assessing shares is appropriate in this case, and Dean
Schmalensee's refusal to do so leads to analytic errors.
- Professor Fisher testified that, although "the question of
what is a relevant market in this case, and in most cases, is
not a question with very definite answers," it is nonetheless
useful because it "is a way of starting to summarize what are
the things you have to understand" to determine "the
constraints on the alleged monopolist." Fisher, 6/1/99am, at
7:17 - 8:5.
- Professor Fisher testified that, because the critical question
in this case is whether Microsoft has "monopoly power in PC
operating systems" -- the product Microsoft sells -- it is
sensible to begin the analysis by determining whether other
products can constrain Microsoft's ability to exercise power
over PC operating systems; that is, to determine whether PC
operating systems are a relevant market. Fisher, 6/1/99am,
at 7:23 - 8:10; see also Fisher Dir. ¶¶ 8-9.
- By contrast, Dean Schmalensee's refusal to define a
relevant market led him to engage in a flawed assessment
of barriers to entry. By "not focusing on market definition to
begin with," Professor Fisher testified, Dean Schmalensee
improperly focused on ease of entry "into the microcomputer
software industry" rather than the difficulty of entry into Intel-based PC operating systems. Fisher, 6/1/99am, at 9:3-12.
Whether entry into the microcomputer "industry" is easy says
nothing about whether it is easy to offer a product that can
effectively compete against Microsoft's operating system.
Fisher, 6/1/99am, 8:21 - 11:8.
40.2. Second, Dean Schmalensee asserted that the market definition is
not useful here because it is "illogical" to exclude other "platform" products that threaten
Microsoft's position in operating systems -- including Internet browsers and Java -- and
platforms are "too heterogenous" to be a market (Schmalensee Dir. ¶ 336;
Schmalensee, 1/13/99pm, at 32:3-17; Schmalensee, 6/23/99pm, at 58:15 - 59:21).
This argument is badly flawed.
40.2.1. It is Dean Schmalensee's analysis that is illogical. By his
reasoning, one could never define a market -- even if it included all of the products (like
PC operating systems) that are substitutes for and compete against one another -- as
long as there are complements for those products (like browsers or other platform
software) that other firms could use to develop new or strengthen existing substitute
products.
- Under Dean Schmalensee's reasoning, it would be illogical
not to place in the same relevant market:
- an oil refiner in California and a railroad company that
is planning on building a new line into California, if the
railroad could threaten the oil refiner's position by
facilitating the entry into the California market of oil
refined in other States. Fisher, 6/1/99am, at 15:13 -
17:21 (giving example of producer of bulky
commodity); or
- a manufacturer of automobiles and a producer of
methanol, if methanol threatens the automobile
manufacturer's position by facilitating the
development of cars that run on methanol. Fisher,
6/1/99am, at 16:5-12.
- In these examples, as Professor Fisher testified, a product
(like railroads or methanol) is properly not included in the
relevant market -- because it is not a reasonable substitute
for products in the market (oil and automobiles) -- even
though it threatens to increase competition within that
market because it is an important complement that can
facilitate growth or entry by products that compete with
products in the market. Fisher, 6/1/99am, at 15:7 - 18:11.
By contrast, under Dean Schmalensee's reasoning, defining
a market in such circumstances would not be a useful
enterprise. Schmalensee, 6/22/99pm, at 25:7 - 26:7. Dean
Schmalensee's position is untenable because, for example,
a market for oil refining plainly can be defined even though
railroads may threaten an oil refiner's market power. Fisher,
6/1/99am, at 15:7 - 18:11.
40.2.2. Although platform products such as Netscape and Java
are complements to operating systems, they are not substitutes for operating systems.
Thus, even though they pose a threat to Microsoft's dominant position in the personal
computer operating system market, they are not in that market.
- See supra ¶ 19.1.
- Dean Schmalensee conceded, "conceptually, there is a
difference, and an important difference" between operating
systems and platforms. Schmalensee, 6/21/99am, at 20:7-10. "An operating system operates the computer...runs the
disk drive, runs the printer, manages the interfaces and so
forth." Schmalensee, 6/21/99am, at 20:4-6. By contrast, a
"platform" exposes "a set of APIs" that can "be used by other
software developers." Schmalensee, 6/21/99am, at 19:15.
- Thus, although "operating systems, typically, are platforms"
and "many platforms are operating systems" Schmalensee,
6/21/99am, at 20:7, platforms cannot fully substitute for
operating systems; see also Gosling Dir. ¶ 8.
- Java and Internet browsers threaten Microsoft's position in
operating systems, not because they can develop into
another operating system, but rather because the platform
they supply could erode the applications barrier to entry and
facilitate the entry and expansion of another operating
system. Schmalensee, 1/13/99pm, at 35:5-12 (agreeing that
"middleware" is a competitive threat to Windows even
though a firm supplying middleware is "not a potential
entrant into the business of supplying operating systems that
would compete with Microsoft").
- Just as a railroad cannot threaten a monopoly oil refiner
unless there is another oil refiner whose entry the railroad
can facilitate, so Java and Internet browsers cannot threaten
Microsoft's position in operating systems unless there are
other operating systems on which those "middleware"
products can be run. Fisher, 6/1/99am, at 18:5-11 ("In the
present case, the growth of the Netscape browser or the
widespread use of original Java might have perfectly well
have broken down the applications barrier to entry and
allowed other operating systems to compete. But it would
be the other operating systems that were then on the
market, not . . . either Netscape, the browser market, or Sun
because of Java."); Schmalensee, 6/23/999am, at 57:14 -
58:3 (conceding that, at present, an operating system is
essential to access web-based applications).
40.2.3. There is no evidence that Java and Netscape constrain
Microsoft's ability to exercise monopoly power today. Thus, even if the market should,
as Dean Schmalensee improperly insists, include "every significant constraint" on "the
alleged monopolist" (Schmalensee, 6/24/99pm, at 60:10-20), Java and Netscape
should not be included in the market.
- Dean Schmalensee conceded that what he characterizes as
Microsoft's existing competitors are not a significant
constraint on its ability to exercise market power.
Schmalensee, 1/14/99am, at 23:5-18, 24:16-21.
- Dr. Warren-Boulton testified that, "under the particular
economic conditions in this market, I would not expect the
prospect of such a threat" to Microsoft's monopoly "in the
future to significantly affect current pricing by Microsoft."
Warren-Boulton, 11/19/98pm, at 33:6-14.
40.3. Third, Dean Schmalensee asserted that market share is not useful in an
industry characterized by significant intellectual property protection and low marginal
costs (Schmalensee, 1/20/99pm, at 63:21 - 65:4). This argument ignores both the
relevant issue -- whether Microsoft's conduct is constrained by competition from others
-- and the importance of other entry barriers.
- Professor Fisher testified that "the applications barrier to entry protects
Microsoft" "independent" of its intellectual property rights in Windows.
Fisher, 6/2/99am, at 14:24 - 15:4. Although a copyright-protected movie
cannot prevent new movies from being written, the applications barrier to
entry inhibits the entry and expansion of other Intel-based PC operating
systems. Fisher, 6/2/99am, at 13:20 - 15:4.
2. Dean Schmalensee's opinion that Microsoft lacks monopoly
power because of low barriers to entry is flawed
41. Dean Schmalensee testified that Microsoft lacks monopoly power because
"Microsoft does not have the protection of substantial barriers to entry" (Schmalensee,
1/14/99am, at 8:22 - 9:9). Dean Schmalensee's reasons for finding the absence of
economically meaningful barriers to entry are flawed and inconsistent with the
evidence.
a. Dean Schmalensee is wrong that the applications
barrier to entry is low
42. Dean Schmalensee asserted that "the facts are inconsistent" with the
existence of a high applications barrier to entry (Schmalensee, 6/22/99pm, at 56:9-12).
But the evidence is to the contrary.
42.1. Dean Schmalensee conceded virtually all of the critical facts that
underlie the applications barrier to entry.
- Dean Schmalensee conceded that operating systems seeking to
substitute for Windows face a "chicken-and-egg problem . . . .
Consumers will not use an operating system if there are not
enough applications written to it. ISVs will not write applications
software for an operating system unless they expect enough
customers to use that operating system." Schmalensee Dir. ¶ 100;
Schmalensee, 6/23/99pm, at 58:10 - 59:24.
- Dean Schmalensee conceded that most applications are "written
for Windows first and sometimes only" for Windows. Schmalensee,
1/13/99pm, at 61:22 - 62:4.
- Dean Schmalensee conceded that Windows has a much larger
stock of applications than are available for other PC operating
systems and "that the rich set of applications available for Windows
contribute significantly to the attractiveness of that platform, and
that . . . by itself gives it an advantage over other platforms."
Schmalensee, 1/19/99am, at 50:3-12.
- Dean Schmalensee conceded that "to attract as much attention as
Microsoft attracts, for a brand new entrant, might require" spending
more than Microsoft does. Schmalensee, 1/14/99am, at 16:10-25.
- Dean Schmalensee conceded that, because of the absence of
sufficient applications available for other operating systems, there
is no operating system to which a large OEM presently could switch
and that Microsoft could raise the short-term price of Windows.
Schmalensee, 1/13/99pm, at 42:16-22, 46:10-12; 6/23/99pm, at
60:9 - 61:4; Schmalensee, 1/20/99pm, at 38:13-17 (agreeing that "if
Microsoft were to increase its prices by 10 percent or 15 percent or
20 percent now, it would increase its short-term profits").
- Dean Schmalensee conceded that "switching costs and network
effects may be larger for some operating systems than for many
applications programs." Schmalensee Dir. ¶ 130.
- Dean Schmalensee conceded that ISVs will not write to a particular
operating system unless they believe the expected return will cover
the costs ISVs must sink. Schmalensee Dir. ¶ 105; Schmalensee,
1/13/99pm, at 61:10-13 (stating that the "reasons for not porting or
not writing to for particular operating system" are "normally
business reasons. You write for an operating system if you think
it's likely to be profitable to do so.").
- Dean Schmalensee conceded "that the applications programming
barrier to entry . . . is something that does, in fact, make it more
difficult for people to enter the business of supplying operating
systems." Schmalensee, 1/14/99am, at 9:10-18.
42.2. Despite these concessions, Dean Schmalensee argued that the
applications barrier to entry is low because there is no evidence that rivals face higher
costs to compete effectively than does Microsoft and that any cost disadvantage is not
significant (Schmalensee Dir. ¶¶ 105, 132; 1/14/99am, at 16:14-25; 6/23/99pm, at
11:22). The evidence is inconsistent with this argument.
- As explained, because of Microsoft's massive installed base, the
expected return to ISVs from writing to other operating systems is
lower than the return from writing to Windows; other operating
system vendors thus face higher costs in inducing a large number
of ISVs to write to their operating systems. See supra Part II.B.3.,
¶ 27.
- Because of the collective action problem referred to above, ISVs
are very unlikely to write to other operating systems in sufficient
numbers to enable those operating systems to become viable
substitutes for Windows. See supra Part II.B.3., ¶ 27.2.3.2 -.3.
- Dean Schmalensee did not analyze "what it would take someone
with a hypothetical attractive operating system" to obtain sufficient
developer support to duplicate the applications available for
Windows. Schmalensee, 1/14/99am, at 14:23 - 15:22.
42.3. In support of his argument that other operating systems do not face
a cost disadvantage in attracting ISVs that prevents effective competition against
Microsoft in PC operating systems, Dean Schmalensee pointed to the recent success of
several niche operating systems, including Linux and BeOS (Schmalensee Dir. ¶¶ 138-40, 158). But the ability of Linux and BeOS in attracting both developer attention and
consumer interest has been limited and thus confirms, rather than undermines, the
existence of the applications barrier to entry.
42.3.1. BeOS is marketed as a specialized complement to
Windows because it lacks the range of applications necessary to substitute for
Windows.
- BeOS's founder, Jean Louis Gassée, stated: "'We don't
want to compete directly with Microsoft to be the only
operating system on the PC . . . but we can be
complementary.'" GX 568 (quoting Gassée). Dr. Warren-Boulton testified that BeOS is a complement, rather than a
substitute, for Windows. Warren-Boulton, 12/1/98am, at
45:5 - 49:10.
- Thus, BeOS is being loaded by OEMs not instead of
Windows, but together with Windows "as a 'dual boot,' letting
users switch between the two as needed." GX 568.
- Although Dean Schmalensee asserted that BeOS's strategy
of becoming a complement to Windows through "dual boot"
was merely a stepping stone to challenging Windows
(Schmalensee, 1/13/99pm, at 54:8-25), that testimony is
undermined by his later testimony that there is no
substantial demand for dual boot systems. Schmalensee,
6/23/99pm, at 62:2-23.
42.3.2. Linux is principally marketed as a server operating system,
and its employment as a desktop operating system is confined to specialized tasks
because its lacks applications comparable to Windows'.
- The CEO of Red Hat, an important Linux vendor, stated that
Red Hat Linux "is almost exclusively being used today to run
specialized server computers that distribute data on the
Internet or internal corporate networks." GX 1568. He
further added: "Just because we exist doesn't mean
Microsoft doesn't have a monopoly with desktop machines.
It's like a telephone company executive holding up a walkie-talkie and saying this is a competitor to local phone service."
GX 1568.
- The President and CEO of Caldera, another Linux vendor,
testified that Caldera's OpenLinux product does not compete
with Windows 95, and that Caldera does not "have the
application base to really compete as a desktop" with
Windows. Warren-Boulton, 12/1/98am, at 50:4 - 51:15 (play
Sparks deposition); see also Warren-Boulton, 12/1/98am, at
56:17 - 57:16 (to the extent Linux is competing with
Microsoft, it is competing in the server market; Caldera does
not view itself as a competitor in the desktop market
because it does not have the necessary stock of
applications).
- An IBM executive stated: "The limiting factor for Linux
breaking into the desktop area right now is simply the lack of
available applications written for the operating system." GX
2091. He explained that "users tend to deploy Linux for
smaller, simpler tasks rather than for huge, enterprise-scale
transactions." GX 2091. Another IBM executive added that,
although "it is technically possible to install Linux on an IBM
thinkpad," there "are just not enough applications to make it
worthwhile." GX 2091.
- Dean Schmalensee conceded that Linux is "not a major
competitor today." Schmalensee, 1/13/99pm, at 45:23.
Although Dean Schmalensee also asserted that "the
majority of sales of Linux" are "for desktops" (1/13/99pm, at
73:18-19), he later contradicted that testimony, conceding
that the "bulk" of Linux users "at present are" using Linux on
"servers." 6/23/99pm, at 66:5 - 67:5.
- Although a small number of OEMs are offering Linux on
some portions of their line (DX 2434 (reporting that Dell is
offering Linux)), a representative of another prominent OEM
stated: "We see Linux as a server phenomenon right now
more than a desktop phenomenon." GX 2091.
42.3.3. Thus, although Linux and BeOS have attracted some
developer attention, consistent with the applications barrier to entry, they have not
attracted sufficient developer attention to provide an effective substitute for Windows for
a large number of users.
- As explained, BeOS and Linux have thousands of fewer
applications available than Windows. See supra Part II.B.3.,
¶ 26.1.3.
- Dr. Warren-Boulton testified that although BeOS is a viable
"specialized" niche operating system, it cannot effectively
substitute for users because it lacks the "extraordinary width
of applications available . . . on Windows." Warren-Boulton,
11/23/98am, at 18:8-22. Dr. Warren-Boulton further testified
that the absence of applications prevents Linux from gaining
substantial market share, and that only the advent of a large
stock of cross-platform applications could Linux present
substantial competition to Windows. Warren-Boulton,
12/1/98am, at 57:8 -59:4.
- Bill Gates reportedly stated regarding Linux: "Like a lot of
products that are free, you get a loyal following even though
it's small. I've never had a customer mention Linux to me."
GX 1378.
- Bryan Sparks testified that Linux cannot effectively compete
with Windows because it "just" doesn't "have the
applications base to really compete as a desktop." Warren-Boulton, 12/1/98am, at 51:12-15 (playing Sparks
deposition).
- An IBM executive explained: "The limiting factor for Linux
breaking into the desktop area right now is simply the lack of
available applications written for the operating system." GX
2091. Another added that, although "it is technically
possible to install Linux on an IBM thinkpad," there "are just
not enough applications to make it worthwhile." GX 2091.
- Professor Fisher testified that "Linux is going to remain a
quite successful niche operating system for some time to
come, and it's not in fact going to offer a serious threat to
Microsoft." Fisher, 6/3/99pm, at 25:14-17.
42.3.4. The existence of niche operating systems, such as Linux
and BeOS, is entirely consistent with Microsoft's possession of monopoly power; and
Dean Schmalensee is wrong when he argues that, if the applications barrier to entry is
high, other operating systems vendors or vendors of other platform products that also
can be complements to Windows are "wasting their time" seeking to attract developers
(Schmalensee, 6/23/99am, at 23:16 - 27:10; 1/13/99pm, at 55:1-22).
- Professor Fisher testified: "It's well-accepted that a firm can
have monopoly power with a fringe of competitors." Fisher,
6/1/99am, at 22:4-17.
- Professor Fisher further testified that it is not sufficient to
overcome the applications barrier to entry "that there may be
some ISV's or even many ISV's that will write to operating
systems other than Windows" because "what makes the
applications barrier to entry so severe" is "the breadth and
depth of the numerous applications that are written or
Windows." Fisher, 6/1/99am, at 55:15 - 56:19.
- Dr. Warren-Boulton testified that the fact firms are porting to
Linux shows that they are betting Linux will be profitable, not
that Linux will substitute for Windows. Warren-Boulton,
11/19/98pm, at 99:7 - 100:4.
- Dr. Warren-Boulton testified that "the existence of fringe
competitors that are in the operating system market does
not mean in any way that Microsoft does not have monopoly
power" because of the applications barrier to entry. Warren-Boulton, 11/19/99am at 19:16 - 20:3.
42.3.5. Any threat Linux and BeOS pose to Microsoft's position is
speculative and does not prevent Microsoft from enjoying monopoly power today.
- The CEO of Red Hat, a leading Linux vendor, stated: "We
are absolutely not a viable competitor" to Windows "at this
time. We have every intention of being one, but how long
will that take? Realistically, it will be 20 years." GX 1568.
- Dean Schmalensee conceded that Linux is not a significant
constraint today on Microsoft's ability to exercise power and
cannot predict when it will exert such a constraint.
Schmalensee, 1/13/99pm, at 52:25 - 53:8; 1/14/99am, at
23:16-25. He conceded that he had made no estimate of
how many PCs have Linux preinstalled now or will have
Linux preinstalled in the future. Schmalense, 6/23/99pm, at
65:17-24. Dean Schmalensee testified that he didn't
"pretend to be able to forecast" whether there will be
substantial demand for Linux in the future. Schmalensee,
6/23/99pm, at 73:7-12.
- Dr. Warren-Boulton testified that: "I have absolutely no
evidence that Microsoft's pricing" of Windows "is constrained
by perceived or actual competition" including "the availability
of Linux." Warren-Boulton, 11/19/98pm, at 96:20 - 97:1.
42.4. Dean Schmalensee is wrong that, even if other operating system
vendors face substantially higher costs than Microsoft faces today, that does not
amount to an entry barrier because an entry barrier exists only if the costs to a rival
operating system today are higher than the costs Microsoft incurred when it entered
(Schmalensee, 6/22/99pm, at 62:8-20).
42.4.1. For one thing, this definition of a barrier to entry contradicts
the approach to entry barriers taken by Dean Schmalensee elsewhere in his testimony
and in his prior writings.
- Dean Schmalensee described as "broadly consistent" with
his definition of barriers to entry the proposition that a barrier
to entry is any factor that "permits a firm already in the
market to earn returns above the competitive level while
deterring others from entering." GX 1516; Schmalensee,
1/14/99, at 6:17 - 7:19. And he testified that a barrier to
entry exists if there are factors that "disadvantage . . . firms
that otherwise would be capable of competing efficiently."
Schmalensee 1/21/99am, at 33:2-5; 6/22/99pm at 70:3-24
(testifying that a barrier to entry exists if the rival cannot
"attract the resources to expand and to become
competitive").
- Dean Schmalensee previously wrote that: "In general, a
clear signal of low barriers is provided only by effective,
viable entry that takes a nontrivial market share . . . ." GX
1513 ((Richard Schmalensee, Ease of Entry: Has the
Concept Been Applied Too Readily, 56 Antitrust L.J. 41,
42 (1987)).
42.4.2. Moreover, successful entry into PC operating systems is
much more difficult today than 15 years ago. The network effects that underlie the
applications barrier to entry are much larger today than when Microsoft entered
because PC penetration (the percent of potential PC users who already use PCs) is
higher and Microsoft is a well-established incumbent with a dominant market share.
- Professor Fisher testified: "When Microsoft won the network
battle, when Windows became the dominant operating
systems, there were . . . many fewer P.C.'s, and there was
no incumbent operating system of equal power and
importance. There were, of course, other operating systems
to fight and there were other operating systems to, as it
were, overcome. One of them, of course, was Microsoft's
own operating system, DOS. The cost after . . . after
Microsoft's victory . . . of persuading ISV's to build such a
stock rather than write for Windows has got to be much
more substantial than it was for Microsoft to persuade them"
to write for Microsoft operating systems "in the first place."
Fisher, 6/1/99am, at 53:6 - 54:1. In other words, "the
economy of scale" that underlies the applications barrier to
entry "is bigger now." Fisher, 6/1/99am, at 54:2-10. See
also Fisher, 6/1/99am, at 56:14 - 58:18 (Although there
might be some incentive for ISVs breaking into the market to
write for new operating systems, that is not enough to
induce ISVs in general to write to other operating systems
such that they can substitute for Windows.)
- Demonstrating the increased penetration of PCs, Microsoft's
own documents show that its shipments of operating
systems rose from 11.4 million units in 1990 to 51.9 million
units in 1996. GX 439.
- Dr. Warren-Boulton testified that, when Microsoft entered the
operating system market, the applications barrier to entry
was not comparable to that which potential entrants face
today. He explained: "[C]ompare the difficulty there with the
difficulty today where you are faced with an incumbent with
tens of thousands of API's, a huge stock of applications--trying to play catch-up at that point, it's just very difficult."
Warren-Boulton, 11/24/98am, at 48:17 - 49:6.
42.5. Dean Schmalensee's assertion that the history of competition for
operating systems shows that the category is easily contestable and that "inflection"
points that displace rivals occur frequently is also belied by the evidence and his prior
writings.
42.5.1. Dean Schmalensee previously observed that the "fact that
entry has occurred in the past does not imply there are no barriers to entry or that entry
is necessarily easy."
- GX 1513 (Ease of Entry Article).
42.5.2. The evidence shows not, as Dean Schmalensee claims,
frequent displacement of a dominant firm, but rather Microsoft's demonstrated ability to
perpetuate its market power.
- Microsoft, according to Dean Schmalensee's own analysis,
has had the dominant PC operating system since at least
the late 1980s. Schmalensee Dir. ¶¶ 118-119.
- Microsoft has maintained that dominance notwithstanding
the development of, among other things, (i) the graphical
user interface; (ii) the migration of PC operating systems
from 16-bit to 32-bit chip architecture; and (iii) the advent of
the Internet, all of which Microsoft claims to be "inflection
points." Maritz ¶ 15.
- Professor Fisher testified after being asked about the history
of users switching operating systems that while "it's true that
users would switch to [another] operating system if they
perceived there to be a significant advantage," the "problem
is that because of the network effects or what's sometimes
been termed the applications barrier to entry, users are not
very likely to perceive that in the present circumstances of
Windows. And Microsoft does its best to see that they
won't." Fisher, 1/6/99am, at 81:25 - 82:10.
42.6. That Microsoft, like other operating system vendors, must continue
to attract ISV attention and improve its product (Schmalensee Dir. ¶ 160; Maritz,
1/28/99pm, at 6:13 - 7:9) is entirely consistent with a high applications barrier to entry
and with market power.
42.6.1. Because of its large installed base, the costs to Microsoft
to attract sufficient ISVs to make its operating system broadly attractive to users are far
less than the costs to its rivals.
- See supra Part II.B.3., ¶¶ 25-27.
42.6.2. Because of its ability to ensure "backward compatibility,"
Microsoft can migrate its installed base between its operating system releases, thus
perpetuating its advantage and, hence, the applications barrier to entry.
- Rational's Mike Devlin testified that, "because Microsoft
strives to make its operating system product 'backwardly
compatible,' we (and our customers) know that a program
we write using the APIs for one Microsoft operating system
will likely run on its successor." Devlin Dir. ¶ 15.
- Microsoft executive Ben Slivka wrote: "Regardless of all the
cool, sexy features in OS/2 (multi-tasking, better graphics
API, memory protection), it was not a no brainer upgrade
from MS-DOS -- customers had to give something up in
order to switch to OS/2: their existing software! Only with
Windows 95 (where we have focused on compatibility to an
amazing extent) are we finally going to enable to move
customers away from MS-DOS." GX 21, at MS98 0102396
(emphasis in original).
42.6.2.1. Microsoft's efforts to attract ISVs are consistent with
monopoly power because monopoly power does not mean unlimited power, because
even a monopolist has an incentive to increase demand for its product, and because
attracting ISVs reinforces the applications barrier to entry.
- See infra ¶ 50.
b. Dean Schmalensee's contention that entry into the
microcomputer software industry is easy is a red
herring
43. Dean Schmalensee argues that "there are no barriers in the microcomputer
software industry that prevent" new entry (Schmalensee Dir. ¶ 37). But whether entry
into the microcomputer software industry as a whole is easy is beside the point because
the relevant question is not whether entry into the "industry" is easy or even whether
producing a PC operating system is easy, but rather whether producing an operating
system with sufficient applications to challenge Windows is easy.
- Professor Fisher testified: "This case . . . centers on monopoly power in
the market for PC operating systems. The question of entry into the
microcomputer software industry in general is not relevant." Fisher,
6/1/99am, at 9:3-17; Fisher 6/1/99am, at 23:6-20.
- As Professor Fisher further explained, there is no evidence that the
microcomputer industry in general does or could constrain Microsoft's
ability to exercise substantial market power over PC operating systems.
"To take a simple but illuminating example, Nintendo produces games.
Games are in the microcomputer software industry," but they are "not a
constraint on Microsoft's power in . . . pricing its Windows operating
system." Fisher, 6/1/99am, at 10:3-7.
- Nor is the fact that others in the microcomputer industry could hire
programmers and produce a PC operating system relevant. Those firms
are "not going to be able to produce an operating system with those
programmers, or with other programmers, which can overcome the
economies of scale and the network externalities that are required."
Those firms are "not going to be able to produce an operating system
which attracts a very large number of applications writers, enough to
overcome Microsoft's very commanding lead." Fisher, 6/1/99am, at 10:23
- 11:6.
44. The factors that, according to Dean Schmalensee (Schmalensee Dir. ¶ 95),
make entry into the "microcomputer software industry" easy are not enough to
overcome the applications barrier to entry into personal computer operating systems.
44.1. That the microcomputer software industry has abundant skilled
programmers and a ready supply of capital cannot, as Microsoft implies (Schmalensee
Dir. ¶¶ 39-44), overcome the economies of scale that create the applications barrier to
entry.
- As explained, the evidence shows that, despite the ready
availability of programmers and capital, the economic incentives to
write for niche operating systems are insufficient to warrant sinking
the huge costs necessary to create an operating system and set of
applications capable of substituting for Windows for a large number
of users. See supra Part II.B.3.b; ¶¶ 25-31.
- Professor Fisher testified that "if there were no other barrier to entry
into operating systems . . . acquiring programmers and financing
and so forth wouldn't be a problem" but there nonetheless "is a
very substantial barrier to entry. I suppose it would be harder to
get in if it weren't easy to get programmers, but getting good
programmers is not near enough to get into the P.C. operating
system business." Fisher, 6/1/99am, at 23:21 - 24:4.
- Dr. Warren-Boulton testified that although there appears to be no
capital entry barrier (Warren-Boulton, 11/19/98pm, at 65:25 - 66:6),
the applications barrier to entry presents a huge entry barrier.
Warren-Boulton Dir. ¶ 59.
44.2. Microsoft's argument that rivals can overcome the applications
barrier to entry by mimicing the Windows user interface and cloning the Windows APIs
is inconsistent with the evidence. To the contrary, cloning the Windows APIs is
infeasible because the number of APIs is very large and constantly changing.
- John Soyring of IBM testified: "Not only is it difficult to reliably
duplicate the function of each API, another company can not
realistically duplicate the function of all of the APIs since Microsoft
continues to introduce new APIs. Applications will not work
correctly if they use APIs whose functions have not been
duplicated. Therefore, there will always be a risk that some
application important to a user now -- or in the future -- will fail.
This uncertainty places a heavy drag on any chance for long-term
success. Given the expense, time and uncertainty involved, I do
not think supporting Windows applications on another operating
system for desktop or mobile PCs offers any reasonable
opportunity for a positive financial return, and I would not
recommend that IBM attempt to provide additional support for
Windows applications in OS/2." Soyring Dir. ¶ 13. Soyring further
testified that, because IBM "lacked the technical capability or the
legal rights" to Microsoft's Windows 95 source code, it could not
ensure that Windows applications would run on OS/2. Soyring,
11/17/98pm, at 76:4-20.
- Bryan Sparks of Caldera, a Linux vendor, testified that "writing a
Windows compatible operating system that's capable of running
Windows applications without Microsoft's supplied operating
system is very difficult. We tried that for sometime in a sister
company when I was at Novell, and we just determined that the
breadth of API's is astonishing" and that Microsoft "adds API's at
what we perceive as an incredible rate, and keeping up with that
API and developing a compatible product is very, very difficult. And
even if you created that, you'd have a hard time branding it as an
acceptable platform because of the breadth of the API." Sparks
Dep. (played 12/1/98am), at 52:15 - 53:25.
- Microsoft's Joachim Kempin noted in December 1997 that cloning
the Windows APIs "would be a lot of work and potentially" pose
"patent problems for someone attacking us." GX 61. Bill Gates
understood that the more difficult a technology is to clone, the more
control over it Microsoft would have; in discussing Microsoft's
strategy for its HTML rendering engine (code named "Trident"),
Gates wrote: "I think we want to make Trident extremely hard to
clone. I think we want to patent elements of Trident. I think we
want to make extensions to Trident on an ongoing basis." GX 351.
- Dr. Warren-Boulton testified: "Certainly, at this point, cloning . . . in
the sense of developing an operating system which would provide
the complete set of API's that is in Windows 98, is physically almost
impossible and, as a practical business matter, is not reasonable."
Warren-Boulton, 11/19/98pm, at 29:13-21.
c. Dean Schmalensee is wrong in arguing that the
existence of potential threats to Windows shows that
barriers to entry are low
45. Dean Schmalensee argued that the threat to the applications barrier to entry
posed by Internet browsers and Java is inconsistent with the conclusion that entry
barriers are high (Schmalensee, 6/22/99pm, at 71:6 - 74:17). This testimony is
misconceived.
- As Professor Fisher testified, the fact that barriers to entry might someday
be eroded, whether by Internet browsers, Java, or other threats, known or
unknown, does not affect whether Microsoft has monopoly power today.
Fisher, 6/1/99am, at 14:9- 15:6; 6/1/99am, at 25:25 - 26:18.
- Dean Schmalensee's position, as Professor Fisher testified, proves too
much. It implies that "any monopolist who took action to preserve its
monopoly and saw a threat worth taking action would be able to argue
successfully that the fact it took the actions means that it can't have
monopoly power." Fisher, 6/1/99am, at 13:12-20.
- Microsoft has taken steps to ensure that these threats cannot overcome
the applications barrier to entry, and its conduct has reinforced the already
substantial entry barriers. Fisher, 6/1/99am, at 12:9-17; Fisher, 6/1/99am,
at 60:4 - 62:2; Fisher, 6/1/99am, at 66:9-25.
46. The possibility that other information applications might eventually wrest
some business away from personal computers similarly does not show, as Microsoft
argues (Maritz ¶¶ 104, 275-77), that entry barriers are low.
46.1. First, other devices, as explained, do not constrain Microsoft's
ability to exercise power over PC operating systems and thus do not affect whether
Microsoft has monopoly power.
- See supra Part II.B.2; ¶ 19.
46.2. Second, even if other devices were to become better substitutes for
some PC uses and gain wider use, that would affect only the value or size of Microsoft's
monopoly power, not its existence. In any event, the evidence shows that demand for
PCs, and thus the value of Microsoft's monopoly, will if anything increase.
- See supra Part II.B.2; ¶ 19.
- Steve Ballmer recently stated that the "PC will remain a very
important central device to the way computing happens, in our
view, over the course of the next ten years." GX 2301, at 4. He
further commented that he could "accept the notion of new devices.
I just don't accept the idea that the PC goes away. And so while
other things, other environments may grow up faster, the PC stays
important." Id. at 5.
- Bill Gates wrote in May 31, 1999, opinion piece for Newsweek that,
"despite pundits who had predicted the end of personal computers,
sales continue to rise." He concluded: "For most people at home
and at work, the PC will remain the primary computing tool." GX
2059.
- The very report Microsoft introduced in support of its contention
that information appliance shipments will soon overtake PC
shipments in fact shows the opposite. It states: "When viewed in
its all-encompassing scale, with all form factors and all customer
segments, PC's far out-ship information appliances on a unit basis .
. . and dwarf the market on a value basis." DX 2423, at page 6. As
Professor Fisher testified, the report shows "that the PC isn't going
away" but, to the contrary, will "remain extremely important" and
that Microsoft's "[m]onopoly over PC operating systems will,
therefore, continue to be important." Fisher, 6/3/99pm, at 69:14-18;
see also GX 2082 (IDC chart showing that number of PC units
shipped is expected to continue to grow significantly until at least
2002, and that despite slightly faster growth in shipments of
information appliances, in 2002 there will still be several tens of
millions more PC units shipped); GX 2083 (IDC chart showing that
the expected value of PC units shipped will remain vastly larger
than the expected value of shipments of other information
appliances until at least 2002).
- Steve Case testified, "'It's hard[] to imagine that PCs won't be the
dominant way people connect with the internet for many years to come
and Microsoft has a pretty amazing lock on that business . . . . Other
devices will emerge, but I doubt any will challenge Windows.'" Fisher,
6/4/99am, at 44:17 - 45:4 (quoting Case Dep. (quoting Ct. Ex. 1) (citation
omitted)). Case further testified that AOL "'[h]as no intention of battling
Microsoft's core business'" and "'no flight of fancy that [AOL] can dent
in any way, shape or form what is a Microsoft monopoly in the operating
system business.'" Fisher, 6/4/99am, at 43:19 - 44:16 (quoting Case Dep.
(quoting Ct. Ex. 1)).
46.3. Third, other devices could threaten Microsoft's monopoly only if PCs
were effectively eliminated as an important computing device. The evidence shows
precisely the opposite: that demand for PCs will remain robust for the foreseeable
future.
- Professor Fisher testified: "So long as PCs remain an important
computing device, and a device which has the property that you
need them to do certain applications," that "[o]ne cannot imagine"
that "a small change in the price of the Windows operating system
is going to cause a lot of people to abandon PCs and go to these
other devices." Fisher, 6/3/99pm, at 82:4-19, 65:23 - 66:6.
Professor Fisher rejected the proposition "that the possible
innovations in various other devices" will "reduce the problem of
Microsoft's monopoly." Fisher, 6/1/99am, at 28:12-15.
- Professor Fisher further testified: "Microsoft has monopoly power
over operating systems for PCs. The question of the influence of
other devices, in this case information appliance devices, would
only become relevant to Microsoft's monopoly power over PC
operating systems if it did one of two things, and I don't think either
one is going to happen. One is that information appliance unit
shipments would become so big and so widespread that people
would drive . . . PCs out." "This chart" DX 2423 "shows PC
shipments growing and continuing to grow, and it matches the
obviously sensible proposition that PCs are going to continue to be
important and indeed very important." Fisher, 6/3/99pm, at 65:9-22.
- Further evidence that PCs will continue in importance is the fact
that non-PC devices cannot be used to accomplish tasks for which
PCs are necessary. For instance, Microsoft pointed to gaming
console as a source of possible competition to PCs (Fisher,
6/2/99pm, at 72:21 - 76:19); but the very exhibit Microsoft
introduced states that "'the new Sony machine will not process text
or calculate a budget.'" Fisher, 6/3/99pm, at 72:15-17 (quoting DX
2553). It further reports: "Sony executives went to some pains
today to assert that their new machine was not a competitor to
Wintel, the combination of Microsoft corporation's Windows
operating system and Intel's pentium microprocessors that
dominates the personal computer industry." DX 2553.
- Dean Schmalensee does not have a basis to opine that Microsoft's
monopoly will be extinguished by the existence of other devices.
When asked whether he had reached a judgment about "the extent
to which" the "personal computer operating system will continue to
be an important business going forward into the future," he
responded that he was being tempted "to prophesy again," and that
"from everything" he had "seen, at least for some number of years -- and it would be hard to say how many -- . . . a lot of work will be
done on the desktop using desktop equipment. How much, how
fast, how the trends will go, I don't know, but it seems apparent to
me that for some time to come," the PC operating system "will be
an important business." Schmalensee, 6/23/99pm, at 41:15 -
42:14.
47. Dean Schmalensee's speculation that operating-system neutral, web-based
applications developed on the Internet could some day erode the applications barrier to
entry (Schmalensee, 6/23/99am, at 36:15 - 41:22) also does not mean that Microsoft
lacks monopoly power.
47.1. First, the possible development of a range of web-based
applications even roughly comparable to the set of applications available for Windows is
entirely speculative.
- Bill Gates wrote, with regard to AOL's acquisition of Netscape, "Platform threat - AOL
doesn't have it in their genes to attack us in the platform space." GX 2241, at MS98
0231890 (sealed; cited portion published).
- Dean Schmalensee conceded that he performed no study or
analysis to determine how many web-based applications exist or
how much investment in that area has been made. Schmalensee,
6/23/99am, at 49:16 - 50:23; Schmalensee, 6/23/99pm, at 37:15 -
38:10.
- Dean Schmalensee conceded that he performed no study of the
number of web-based applications that require Windows.
Schmalensee, 6/23/99am, at 54:21 - 55:9.
- Dean Schmalensee, when asked if "there will come a time in the
future when people will spend as much effort developing web-based applications as they do developing applications for
Windows" responded: "I'm not a prophet . . . . I cannot, as I sit
here, represent that I know what will happen in this regard in the
future." Schmalensee, 6/23/99pm, at 38:18 - 39:2; Schmalensee,
6/23/99pm, at 39:13 - 40:1 ("one extrapolates current trends with
some hazard in this business, and as I say, I'm not a prophet").
- Dean Schmalensee conceded that he did not, and could not,
determine the number of web-based applications would exist in the
next couple of years. Schmalensee, 6/23/99am, at 50:24 - 51:7.
- Professor Fisher testified that he conducted no study of the number
of web-based applications because "however interesting those
applications are they are nowhere near enough to overcome the . .
. applications barrier to entry into operating systems for PCs."
Fisher, 6/3/99pm, at 81:6-15.
47.2. Second, because web-based applications require a browser,
Microsoft could vitiate this potential threat by gaining a substantial share of browsers
and then using proprietary extensions.
- See infra Part VII.D.
48. Dean Schmalenee is also wrong in arguing that the possibility of entry
should be assessed "over a long period of time," beyond the next several years
(Schmalnesee Dir. ¶ 184).
- As Professor Fisher testified, this argument confuses the question of the
period over which Microsoft could recoup predatory investments designed
to preserve its monopoly power with whether that power exists. For
example, under Dean Schmalensee's reasoning, one could not determine
whether AT&T was a monopolist in 1980 without considering "the
telephone industry well into the next millennium because it is possible that
if it succeeded in driving out MCI, it would still recoup money 30 years
later." Fisher, 6/1/99am, at 18:12 - 20:10-15.
3. Dean Schmalensee's contention that "long term threats"
prevent Microsoft from exercising monopoly power today is
flawed
49. Dean Schmalensee bases his argument that barriers to entry are low, and
thus that Microsoft lacks monopoly power, principally on his contention that Microsoft's
pricing of Windows is severely constrained by largely unknown long-term threats to its
position. Dean Schmalensee reasons that, if Microsoft were a monopolist, it would be
charging more than $1,800 for Windows, instead of the approximately $70 it in fact
charges, and infers from this that Microsoft is engaging in massive limit pricing
designed to exclude threats that have not yet arisen (Schmalensee, 1/21/99am, at
11:17-18, 13:11-19, 23:25 - 24:5). The evidence, however, is at odds with Dean
Schmalensee's argument.
49.1. First, limit pricing -- lowering price and thus sacrificing revenues
today in order to deter entry tomorrow -- is irrational if potential rivals know that the firm
can lower price later, if and when competition emerges. In that event, rivals will be
deterred by the prospect of price reductions in response to competition, and there
would be no reason for the monopolist to sacrifice revenues by cutting prices today.
Dean Schmalensee's limit-pricing analysis thus must assume that Microsoft cannot
credibly threaten to lower price in the future. Microsoft, however, plainly has the power
to lower prices in the future, if and when competition emerges.
- Professor Fisher and Dr. Warren-Boulton both testified that it is not
plausible that Microsoft keeps the price of Windows significantly
lower than Microsoft otherwise would in order to deter entry
because Microsoft can lower its price should such entry occur.
Potential entrants evaluate the profits they would earn after entry,
and they recognize that Microsoft's price now is not a guide to what
Microsoft would charge -- and what profits are therefore available
to the entrant -- if entry actually occurred. Fisher, 6/2/99am, at 6:2
- 7:14; Warren-Boulton, 12/1/98am, at 43:14 - 45:5.
- Microsoft can credibly lower price tomorrow in response to entry
because, as Dean Schmalensee himself testified, the marginal cost
to Microsoft of producing and selling additional copies of Windows
through an OEM is "zero." Schmalensee, 1/20/99pm, at 68:5-20;
Warren-Boulton, 11/19/98am, at 58:25 - 59:3; Schmalensee Dir. ¶
85.
- Dean Schmalensee asserted that the greatest threats to Windows'
dominance are not other PC operating systems, but rather
"paradigm shifts." Schmalensee, 1/13/99pm, at 65:7-24. But there
is no reason to think that the possibility of "paradigm shifts" is
affected by the prices Microsoft charges today. Fisher, 1/11/99pm,
at 47:19 - 48:17.
49.2. Second, Dean Schmalensee's hypothesis that Microsoft is
engaging in massive limit pricing is also inconsistent with how Microsoft views the
constraints on its pricing of Windows.
- Kempin testified that he did not consider competing operating
systems or "'competition more generally'" in setting the Windows 98
royalty. See supra Part II.A; ¶ 15.1.5.
- Kempin's memorandum on Microsoft's pricing of Windows 98, sent
to Bill Gates, does not identify long-term threats as a constraint on
Microsoft's pricing of Windows. Long-term threats are described
instead as possibilities that could "derail" Microsoft's strategy. GX
365.
- Based on this evidence, Professor Fisher testified that long-term
entry is not a significant consideration in Microsoft's choice of a
price for Windows. Fisher, 1/13/99am, at 23:5-14 (it is doubtful
"long-term entry . . . is . . . at the forefront of the Microsoft corporate
mind").
49.3. Third, the analysis Dean Schmalensee advanced says nothing
about whether Microsoft possesses monopoly power. To the contrary, Dean
Schmalensee's analysis (Schmalensee, 6/23/99am, at 6:3 - 9:17) shows, at most, that
Microsoft is not seeking to maximize its short-term profits exclusively through operating
system royalties.
- Professor Fisher testified that Dean Schmalensee's analysis at
most could show only that Microsoft is not taking out its monopoly
power in the short-run price of Windows. Fisher, 1/12/99pm, 16:12
- 17:17. But it "wouldn't tell you anything about the power itself. It
wouldn't tell you whether Microsoft had power. It would tell you
whether it was exercising power in a particular way." Fisher,
6/1/99pm, at 9:3-12; Fisher, 1/11/99pm, at 48:13 - 50:19 (even if
one concluded that Microsoft had priced to deter future entry, that
would not necessarily mean that Microsoft lacks monopoly power
today).
49.3.1. An analysis that focuses entirely on short-run prices is
inappropriate because it ignores the fact that Microsoft may charge what seems like a
"low" short-term price in order to maximize its profits in the future for reasons unrelated
to deterring entry.
49.3.1.1. By keeping price low today and "growing" the
market, Microsoft earns greater complementary revenues in the future.
- Paul Maritz testified: "Microsoft broadly licenses
operating system products to computer
manufacturers at attractive prices (typically less than
5% of the price of a new computer). Such broad
licensing promotes the adoption and use of
Microsoft's operating system products, which in turn
promotes the development of a wide range of useful
complementary hardware and software products that
are compatible with Windows and thus with other
Windows-related products." Maritz Dir. ¶ 132.
- Professor Fisher testified that a monopolist like
Microsoft has a greater incentive than a
nonmonopolist would to set a low price with the
purpose of furthering the general popularity of
computing because only the monopolist reaps the full
future reward of the greater popularity. Fisher,
1/12/99pm, 66:4 - 67:9, referencing colloquy at
Fisher, 1/12/99am, 24:13 - 25:21.
- Professor Fisher further testified that, because
Windows users often buy upgrades and other
complementary products from Microsoft in years after
their initial Windows purchase and because the
number of copies of Windows sold has grown every
year, Microsoft earns greater complementary
revenues per copy of Windows than can be captured
in Dean Schmalensee's equation. Fisher, 6/4/99am,
at 13:23 - 15:3. Dean Schmalensee improperly
compared current Windows revenues to current
revenues from complementary products. Fisher,
6/4/99am, at 13:23 - 15:3.
- In fact, Dean Schmalensee did not investigate the
complementary revenues Microsoft receives from the
sale of Windows. Rather, he accepted his staff's
representation that Microsoft "record[s] operating
system sales by hand on sheets of paper" and, for
that reason, lacked "a sophisticated internal
accounting system" from which he could estimate
anticipated complementary revenues. Schmalensee,
1/20/99pm, at 46:3 - 49:8.
49.3.1.2. Dean Schmalensee ultimately conceded that
Microsoft may be pricing low today to obtain long-term benefits that depend on network
effects.
- Dean Schmalensee testified that Microsoft "keeps
price low so that a lot of people use Windows, and I
can attract applications vendors for both reasons,
both because a lot of people use it and because there
are more applications for it." Schmalensee,
6/22/99pm, at 39:13-18.
- In this regard, Dean Schmalensee's testimony is
consistent with Professor Fisher's testimony that
Microsoft has "an overriding interest in preserving the
applications barrier to entry and taking advantage of
the network effects. When it sells Windows, the more
Windows it sells, the more the network effects are.
That, by the way, is a reason for keeping the price of
Windows lower than would otherwise be the case,
and there are other reasons as well." Fisher,
1/12/99am, at 21:8-14.
49.3.1.3. Dean Schmalensee's focus on short-term price
also overlooks the fact that Microsoft takes a portion of its monopoly returns, not in
cash payments, but rather in the form of costly restrictions upon its customers and
commitments by them to behave in ways that augment and maintain Microsoft's
monopoly power.
- Professor Fisher testified that Microsoft "takes some
of its profits in the form of protection of its monopoly."
Fisher, 1/12/99am, at 19:20-21. Professor Fisher
further testified that there are examples in other
industries of sellers with monopoly power choosing to
exercise that power by means other than charging as
high a price as possible for the monopolized product;
in the late 1970s, for example, the two airlines that
owned computer reservations systems found it more
profitable -- before the Civil Aeronautics Board
intervened -- to raise rival airlines' costs by biasing
the systems' flight displays than to raise the price to
those airlines of participating in the systems. Fisher,
1/12/99am, at.14:11 - 17:3.
-
- redacted -
GX 1498, at GW 019843 (sealed).
- Garry Norris of IBM testified that some of Microsoft's
MDA milestones require IBM to take acts that exclude
Microsoft's potential rivals. Indeed, Norris testified,
referring to the language in his contemporaneous
notes of their March 6, 1997 meeting, that Microsoft's
Bengt Akerlind told IBM "no Netscape and receive
more MDA dollars across the P.C. company" and
threatened IBM with "MDA repercussions" unless IBM
agreed to promote IE exclusively. Akerlind told Norris
that Microsoft might impose these repercussions, i.e.,
raise the price of Windows to IBM, either by modifying
MDA milestones themselves or by exercising its
discretion to decide whether IBM had met its MDA
milestones. GX 2164; Norris, 6/8/99am, at 29:19 -
30:23; Norris, 6/8/99am, at 31:24 - 32:12.
-
- redacted -
GX 1436
(sealed).
- redacted - Fisher 1/12/99pm, 41:19 - 43:20
(sealed session).
- Microsoft offered IBM substantial MDA discounts to
reduce support for OS/2; had IBM accepted the
provisions offered by Microsoft, Microsoft's annual
Windows revenues from IBM would have dropped by
$40 to $48 million, given IBM's volume of Windows
shipments at that time. Norris, 6/7/99am, at 22:16-18.
Norris testified that Microsoft offered to reduce the
price IBM paid for Windows 95 if IBM, in Microsoft's
words, agreed to "adopt Windows 95 as the standard
operating system for IBM" and to make it "the only OS
mentioned" in advertisements and marketing
materials. Norris 6/7/99am, at 20:1 - 23:5 (quoting
GX 2132). IBM did not agree to these provisions and
others because they would have had the effect in the
marketplace of effectively putting its own OS/2
operating system product "to the grave." Norris
6/9/99am, at 10:18-24.
- Kempin recognized that one tactic Microsoft could
use to effectively decrease the cost to Windows
would be to "Reduce some of the more rigid licensing
requirements, which increase costs to the OEMs."
GX 365.
49.3.2. Dean Schmalensee's analysis is also flawed because it
leads to absurd results.
- Professor Fisher demonstrated that, at the price that would
maximize Microsoft's short-run profits, given Dean
Schmalensee's undisputed assumption that Microsoft's
short-run marginal cost of Windows equals zero, the
elasticity of demand for Windows must equal one. Ct. Ex. 2-A; Ct. Ex. 2-B; Fisher, 1/12/99pm, at 13:16-16:19.
- If Microsoft believes that it is operating at a point on the
demand curve at which the elasticity of demand for Windows
equals one, Microsoft must believe that a 10% increase in
the price of Windows -- about $5 -- would lead to about a
10% decrease in the number of copies of Windows it sells.
As Professor Fisher testified: "If you look at the testimony of
the OEMs and you just think about it, that can't possibly be
right. You can't believe that. It would believe it would lose
10% if it raised the price only $5, and the OEMs have no
other place to go." Fisher, 1/12/99pm, at 16:16 - 17:8
(testifying about Microsoft's pricing).
49.4. Fourth, even taken on its own terms, Dean Schmalensee's
calculation of a "short term" monopoly price for Windows of more than $1,800 is wrong.
49.4.1. Dean Schmalensee's calculation depends on his
assumptions about three variables, all measured in the same year:
- the average
hardware price of a PC less the price of Windows;
- the elasticity of demand for PCs, and
- the average revenues Microsoft earns from sales of other complementary products (GX 1960).
In each instance, Dean Schmalensee made arbitrary or flawed assumptions.
49.4.1.1. Average price of a PC. Dean Schmalensee used
$2,000 as the average price of a PC, even though that average was calculated by
including higher-priced computers, such as workstations, and does not reflect the fact
that, in setting its prices, Microsoft takes into account the downward trend in PC prices.
- Dean Schmalensee acknowledged that his $2,000
figure includes significantly more expensive servers
and that "presumably" the right number to use would
be one that includes only desktop PCs.
Schmalensee, 6/24/99pm, at 71:23 - 73:9.
- Professor Fisher testified that the $2000 figure Dean
Schmalensee used in January for the average price
of a PC (including Windows) significantly overstates
today's true average price, whether or not that price
properly includes monitors. Fisher, 6/4/99am, at 6:5-21; DX 2492 (citing $953 price).
- The average price of PCs has clearly fallen in recent
years and continues to fall. Fisher, 6/4/99am, 11:10 -
12:3. In February 1999, PC Data reported, sub-$600
PCs (not including monitor) were the fastest-growing
retail segment and constituted 19.9% of all retail
sales. DX 2493. Even the IDC study cited by Dean
Schmalensee as the source for his estimate of
average PC price shows PC prices falling historically
and for the forseeable future. GX 2300; see also DX
2498, at 22. Dean Schmalensee acknowledged that
it is appropriate to take the decline in hardware prices
into account and lowered the average price of a PC in
his formula from $2,000 in his January testimony to
$1,800 in his June testimony. Schmalensee,
6/23/99am, at 14:11-17.
- Contrary to Dean Schmalensee's $2,000 figure,
Microsoft executives looked to the future expected
price of the PC in setting the Windows royalty, and
Joachim Kempin's December 1997 memorandum to
Bill Gates discusses how Microsoft's pricing should
take into account the growth of the sub-$1,000 PC
market segment. GX 365.
49.4.1.2. Elasticity of demand for PCs. Dean Schmalensee
assumed that the elasticity of demand for PCs is 2 (Schmalensee, 1/21/99am, at 10:19-20; Schmalensee, 6/24/99pm, at 62:17-19), but the reasons he gave for doing so are
arbitrary and unsound.
- Dean Schmalensee asserted that an elasticity of 2
followed from plaintiffs' assertion that PCs were a
market (Schmalensee, 1/21/99am, at 10:5-7;
Schmalensee, 1/20/99pm, at 39:1-3; Schmalensee,
1/20/99pm, at 40:22 -23). But neither plaintiffs nor
their experts took the position that PCs are a market.
Neither Dr. Warren-Boulton nor Professor Fisher
testified that there is a market for operating systems
for Intel-based PCs, and Professor Fisher made clear
that conclusion does not require defining a market for
personal computers. Fisher, 6/2/99pm, at 30:2-13;
Fisher, 6/3/99pm, at 65:23 - 66:6.
- Dean Schmalensee gave inconsistent testimony
about his own views on the plausible range of
elasticities.
- In his October 1998 deposition, he testified
that there is a plausible range one could think
of, and that "'numbers below one are pretty
implausible. Numbers above five and six are
pretty implausible, based on elasticities one
encountered, but that's a pretty wide range,
economically, and I don't think I know enough
to narrow it.'" Schmalensee, 6/24/99pm, at
63:16-20 (quoting Schmalensee's deposition).
- In his trial testimony, Dean Schmalensee
testified he had done no work since his
deposition to estimate the price elasticity for
PC systems and had seen no estimates in the
literature. Schmalensee, 1/20/99pm, at 39:8-11.
- Yet in his rebuttal direct testimony, when asked
whether he had previously testified "that a
range of up to 6 was plausible," he answered:
"No, it isn't. I went back and looked at
everything I've said in this proceeding on the
subject, and I don't think that's consistent with
what I said. . . . I never said 4 was plausible
and I don't believe it." Schmalensee,
6/23/99am, at 18:16-24.
- Dean Schmalensee sought to reconcile this
inconsistent testimony by characterizing his
deposition testimony as an "outlier" among his
testimony on the topic. He contended that his
January testimony was consistent with his
current contention that an elasticity of four in
this market is "totally implausible."
Schmalensee, 6/24/99pm, 64:12-23; id. at
67:23 - 69:6.
- Dean Schmalensee arbitrarily assumes an elasticity
of demand for PCs of 2 despite having testified at his
deposition that the elasticity could plausibly range up
to five or six and despite having cited no studies of
the PC industry by himself or others to justify his
assumption. Schmalensee, 6/24/99pm, at 63:16 -
65:15.
49.4.1.3. Complementary revenues. Dean Schmalensee
used an incorrectly low and arbitrarily-derived estimate of Microsoft's complementary
revenues from Windows sales.
- Dean Schmalensee derived his estimate of
complementary revenues by arbitrarily dividing the
revenue of Microsoft's applications group (which he
regarded as the repository of complementary
revenues) by the revenue of Microsoft's platforms
group (which he regarded as representing revenue
from the sale of Windows), and then doubling that
figure to reach what he called a "generous" estimate
of approximately $100 in complementary revenues
per copy of Windows. Schmalensee Dir. App. B, at
B-4 n.11 ; Schmalensee, 1/21/99am, at 11:23 - 12:1
(characterizing this estimate as "generous");
Schmalensee, 1/21/99am, at 17:25 - 18:25
(explaining methodology).
- Correcting for Dean Schmalensee's errors, Professor
Fisher estimated Microsoft's true complementary
revenues as $160, before any doubling for
conservatism -- that is, more than three times Dean
Schmalensee's estimate. Fisher, 6/1/99pm, at 15:16
- 17:5. And this is quite apart from Dean
Schmalensee's failure to take full account of future
complementary revenues because of his formula's
limitation to the short-term. See supra ¶ 49.1.
49.4.2. Despite the conceptual defects in Dean Schmalensee's
formula, as Professor Fisher testified, it is nonetheless possible using plausible
estimates of each of the variables in the formula to estimate a short-term profit-maximizing price for Windows that is close to the price Microsoft actually charges.
- Professor Fisher testified that using an elasticity of demand
for PCs of four (within the range that Dean Schmalensee
testified is "plausible"), a current price per PC of $1,000, and
a corrected complementary-revenues estimate of $160,
Dean Schmalensee's equation shows that the price for
Windows that would maximize Microsoft's profit is $65 --
very close to the actual price of Windows. Fisher, 6/1/99pm,
17:17 - 18:2. Using an elasticity of five -- also within Dean
Schmalensee's range -- would produce, according to Dean
Schmalensee's analysis, a profit-maximizing price of $40,
which is in fact below the actual price of Windows. Fisher,
6/1/99pm, at 18:3-6; see also Fisher, 6/1/99pm, 11:8-23.
- Professor Fisher further testified that performing the same
exercise with the significantly higher price per PC that was
typical in 1996 or 1997 still produces estimates, according to
Dean Schmalensee's analysis, that are within a few hundred
dollars of the actual price of Windows in 1996 or 1997.
Fisher, 6/2/99am, at 31:13-21.
4. Dean Schmalensee is wrong that Microsoft's other behavior is
inconsistent with monopoly power
50. Dean Schmalenssee argued that Microsoft is not a monopolist because it
does not "behave like a firm with monopoly power" (Schmalensee Dir. ¶ 180 (emphasis
omitted)), but his analysis is flawed.
50.1. As an initial matter, Schmalensee's approach is flawed because it
implicitly but wrongly assumes that monopoly power means unlimited power and
ignores the fact that a monopolist has an incentive to increase its monopoly profits by
improving product quality.
- Dr. Warren-Boulton testified that, "to an economist, every
monopolist faces competition. Every monopolist faces potential
entry. But the reason why he faces competition or potential
competition is because profit-maximizing behavior is to raise your
prices until you run into that competition. . . . So something is out
there, whether it's entry, whether it's just simply demand falls off, or
whatever reason, there is a reason why" a monopolist "doesn't
increase the price further than he is already increasing it." Warren-Boulton, 11/19/98am, at 38:23 - 39:18.
- Dr. Warren-Boulton also testified: "There's nothing about
monopoly power that indicates that a profit-maximizing monopolist
has some incentive not to listen to its customers." Warren-Boulton,
11/30/98am, at 29:22 - 30:11.
>
- Professor Fisher testified that even a monopolist has incentive to
increase demand for its product. Fisher, 1/12/99pm, at 19:1 -
20:15.
50.2. Microsoft's general efforts to innovate are thus consistent with
monopoly power, even if absent innovation Microsoft might eventually lose its monopoly
power (see Maritz Dir. ¶ 153).
- Professor Fisher testified that one "can't look at an industry or a
market, and . . . from merely the fact that innovation is going on,
conclude that there can't be monopoly power." Fisher, 1/12/99pm,
at 19-20; 6/3/99am, at 8:11-14.
- Dr. Warren-Boulton testified that "a company always has the
option, if you like, of simply stopping technical innovation . . . I just
don't understand why anyone would want to do that . . . there is
nothing I conclude from that as to whether or not" Microsoft "is a
monopoly or not. A monopolist also has the same incentive to
innovate as a competitive firm." Warren-Boulton, 11/19/98am, at
79:12-25.
- Dr. Warren-Boulton further explained: "if Microsoft were to simply . .
. shut down its R&D version . . . it would probably lose its monopoly
power within a reasonable time period," but that is entirely
consistent with Microsoft's possession of monopoly power today.
Warren-Boulton, 11/19/98pm, at 41:8 - 43:14.
50.3. Microsoft's efforts (amounting to several hundred millions of dollars
a year) to induce ISVs to write applications that run on Windows are also consistent
with monopoly power.
50.3.1. Inducing ISVs to write more and better applications makes
Microsoft's operating system more attractive, thus increasing the monopoly profits
Microsoft can earn.
- Paul Maritz testified that Microsoft's efforts to work with
developers result in "great applications for Microsoft's
Windows family of operating system products," which in turn
increases Windows' attractiveness to consumers. Maritz
Dir. ¶¶ 127, 136.
- See supra ¶ 26.1.
50.3.2. Inducing ISVs to write more and better applications to
Windows also increases the applications barrier to entry because it increases the
attractiveness of the Windows platform, which reinforces ISVs' incentives to write first
and foremost to Windows, and reduces the resources ISVs can devote to writing to
other operating systems.
- Dr. Warren-Boulton testified that inducing ISVs to develop
for Windows is "an investment in creating the applications
barrier to entry." Warren-Boulton, 11/24/98am, at 39:13-14.
50.4. Microsoft's argument that the existing installed base of Windows
users and piracy together prevent Microsoft from exercising monopoly power
(Schmalensee, 1/14/99am, at 25:4-22; Maritz Dir. ¶ 123) is also flawed.
50.4.1. The evidence shows that, whatever constraint piracy
imposes on Microsoft's pricing, it is not substantial and does not prevent Microsoft from
enjoying monopoly power.
50.4.1.1. Microsoft discourages piracy by penalizing OEMs
through MDAs for shipping naked machines.
- Dean Schmalensee testified that Microsoft's MDAs
penalized OEMs for shipping naked machines, and
that the purpose of the penalty is to reduce piracy.
Schmalensee, 6/23/99pm, at 67:13 - 70:17; 69:7 -
70:18.
-
- redacted -
50.4.1.2. There is no evidence that piracy prevents
Microsoft from exercising substantial monopoly power. To the contrary, the evidence
shows that Microsoft has substantial and durable discretion over its pricing of Windows
notwithstanding the possibility of piracy.
- See supra Part II.C., ¶¶ 33-38.
50.4.2. The evidence similarly shows that the modest constraint
created by its installed base does not prevent Microsoft from enjoying monopoly power.
- As Professor Fisher testified, Microsoft prohibits licensees
from transferring operating systems to new computers; there
is thus no "secondary market" in operating systems. Fisher
Dir. ¶ 77.
- The installed base cannot affect the price of operating
systems acquired in connection with OEM sales. "New
operating systems are principally acquired in connection with
the purchase of new computers and only secondarily in
connection with upgrades. At best, Microsoft's installed-base argument relates to its pricing of upgrades. It does not
apply to the more important channel of new computers."
Fisher Dir. ¶ 75; Warren-Boulton, 11/19/98am, at 64:18 -
66:8 (testifying that consumers buying an operating system
with a new PC and an "upgrade" operating system have
different demand characteristics).
- And, as Dr. Warren-Boulton explained, the constraint on
Microsoft's pricing of upgrades is modest because, although
software "never wears out" (Maritz Dir. ¶ 202), it can become
obsolete. Warren-Boulton, 11/19/98am, at 64:7-17. Indeed,
as explained, Microsoft's pricing of its Windows 98 upgrade
product evidences substantial pricing discretion and thus
monopoly power even in that market segment. See supra ¶
36.
III. Alternative Platform-Level Technologies, Especially Internet Browsers and
Java, Threaten Microsoft's Operating System Monopoly
A. Middleware technologies have the potential to reduce the
applications barrier to entry and facilitate operating system
competition
51. The applications barrier to entry, as explained, is the result of a chicken and
egg problem: An operating system cannot attract a sufficiently large set of applications
o challenge Windows without a large installed base with which to attract ISVs, and
cannot obtain a large installed base without a large and attractive set of applications.
- See supra Part II.B.3; ¶¶ 23-32.
52. Middleware technologies -- principally Internet browsers and Java, which are
designed to run on top of an operating system -- threaten to facilitate the creation of
competition to Windows by reducing the importance of Windows APIs and thereby
eroding the applications barrier to entry.
- Chris Jones described in an August 1995 e-mail: "We are so dominant in
all other aspects of the market that we can never be displaced by a full
frontal assault. However, when we do leave a hole in our strategy, there
are many companies eager to move in and try to leverage this hole to
grow into our other businesses. And they have: you only have to browse
the Web to realize that NetScape, Sun, Apple, Adobe, and MacroMedia
are establishing a presence. The real threat to our business is solutions
like Java, which present a different programming model than Windows
and take developer and content provider mind share. This platform
offering is quickly evolving, with two key players moving forward with their
offerings and evangelism. In addition to Java, NetScape has announced
an interface for plugging in different document types, while in turn Apple is
building a programmable browser using OpenDoc. The Result -- People
Aren't Writing to Our Interfaces. The solutions people have implemented
today do not benefit Windows uniquely -- they work on all platforms
equally well. More importantly, these solutions are being driven by other
companies rather than our own -- specifically, NetScape and Sun.
Without an alternative to this platform we will lose control of a critical
segment of the developer (and customer) market." GX 523, at MS98
0103658.
- Paul Maritz testified: "If a middleware product provides a set of APIs to
software developers that makes them more productive and enables them
to create better software products, the value of any underlying operating
system will, of course, be greatly reduced." Maritz Dir. ¶ 236.
- Dr. Warren-Boulton testified that a competitive threat to Microsoft's operating
system monopoly is less likely to come from other operating system products than
from extensions to complements of Windows that also can serve as platforms to
which ISVs write applications programs . . . The wide dissemination of the
complement among PC end users means that application developers can reach a
broader base of potential customers by writing to it than by writing to an
operating system that competes directly with Windows 95/98 and starts with very
low market penetration and installed base." Warren-Boulton Dir. ¶¶ 65-66; see
also Schmalensee Dir. ¶ 136; Tevanian Dir. ¶ 46.
B. The widespread use of non-Microsoft Internet browsers threatened
to erode the applications barrier to entry and Microsoft's monopoly
power
1. The nature of the browser threat
53. Internet browsers, including Netscape Navigator, possess three key
middleware characteristics that make them threats to Microsoft's operating system
monopoly in ways that traditional operating systems, without middleware assistance,
are not.
53.1. First, by contrast to traditional operating system competitors to
Windows, Internet browsers can gain (and have gained) widespread usage based on
their value as a complement to Windows, without having first to compete against
Windows as a substitute.
- Dr. Warren-Boulton testified: "Although a PC operating system cannot
successfully compete against Microsoft's operating systems without first
overcoming formidable barriers to entry, the situation is different for a
product (e.g., browsers or Java technology) that is both initially a
complement from an end user perspective and a potential substitute for the
Windows 95/98 platform to which applications developers can write.
Because applications written to such a complement are compatible
with Windows, their developers can sell their applications to users
of the Windows operating system. Eventually, a sufficient number
of such applications may become available to support an
alternative platform to Windows. " Warren-Boulton Dir. ¶ 65; see
also Warren-Boulton, 12/1/98am, at 67:19 - 68:8.
53.1.1. With the advent of widespread popular use of the Internet
in 1994-95, browser products became a widely-used complement to Windows.
Netscape Navigator emerged as the browser market leader and quickly attracted a
large installed base of users.
- As Netscape CEO James Barksdale testified: "The commercial
release of Netscape Navigator 1.0 occurred on December 15, 1994.
By the end of the second quarter of 1995, Netscape had collected
over $10 million in revenue generated by the browser alone. By
the end of 1995, Netscape had collected approximately $45 million
in revenue from browsers," (Barksdale Dir. ¶ 18) had "over 70
percent market share for Internet clients and had distributed
15 million browsers around the world through a variety of
channels including ISPs, OEMs, and resellers as well as
over the Internet." Barksdale Dir. ¶ 66.
- James Clark, founder and former Chairman of Netscape,
testified that Netscape attained an "85 percent market
share." Clark Dep., 7/22/98, at 39:3-9 (DX 2562).
53.1.2. Netscape enjoyed early success with its innovative
browser.
- Barksdale testified that Netscape Navigator "hid the
technological complexities of the Internet from the end user. Its
introduction into the marketplace had a profound effect; the
product was an immediate and huge success precisely because of
its ease of use and its ability to bring so much new multimedia
information to the consumer." Barksdale Dir. ¶ 12; see also
Schmalensee, 6/23/99pm, at 47:23 - 48:3.
- In a presentation in April 1996, Microsoft Senior Vice-President
Brad Silverberg made clear that Netscape and Sun "are smart,
aggressive, and have a big lead. This is not Novell or IBM we are
competing with." GX 40 (emphasis in original).
- Indeed, in May 1996 Mr. Gates had made clear to Microsoft's top
executives his impressions of Netscape as a strong competitor:
"During this Thinkweek I had a chance to play with a number of
Netscape products. This reinforced the impression that I think all
of us share that Netscape is quite an impressive competitor. They
are moving at full speed." GX 41, at MS6 6012952.
53.2. Second, because Internet browser products, including Netscape
Navigator, expose APIs to which ISVs can write, Internet browsers can serve as a
"platform" for other software used by consumers.
- Gates recognized that Netscape Navigator exposed APIs:
- redacted -
Gates Dep., 8/27/98, at 54:4-12 (DX 2568A)
(sealed); see also Gates Dep. (played 12/2/98am), at 21:25 -
22:18.
- As Apple's Avadis Tevanian explained: "Internet-related
technologies such as browsers are important in the development of
future software platforms which could operate 'on top' of different
operating systems. These software platforms could be used to run
various applications such as programs that display, edit,
manipulate and transmit various types of content." Tevanian Dir. ¶
45.
- Microsoft's James Allchin testified that middleware products such
as browsers running on top of a conventional operating system can
serve as a platform for other software. Allchin Dir. ¶ 35; Dertouzos
Dep., 1/13/99, at 427:18 - 428:4; Slivka Dep., 1/13/99, at 712:21 -
715:6.
- Allchin acknowledged that browser products such as Netscape's
expose "certainly hundreds, maybe thousands" of APIs to application
developers without being included in any operating system. Allchin,
2/3/99pm, at 10:1; see also Maritz, 1/25/99pm, at 29:22 - 30:19
(distinguishing Netscape's browser from browser "shells" built on
top of Internet Explorer in that Netscape's browser had the
capability of developing into an alternative platform); GX 489, at
MS6 6000311 ("Navigator/NetOne provides a new API set -- in
near/medium term, Navigator provides the volume platform for ISVs
& Corps to target.").
- Barksdale testified that Netscape sought to "allow people to build
applications on top of our browser using what is called the NSAPI,
the Netscape Application Programmer Interface," Barksdale ,
10/27/98am, at 73:11-25. As a result, "the browser is not only
useful for browsing the Web but also can serve as a platform for
the development of all sorts of network-centric software
applications, such as online-banking software products. These
network-centric applications, in essence, sit on top of the browser
and take advantage of its Web-oriented functionality." Barksdale
Dir. ¶ 15; see also Colburn Dir. ¶ 8; Andreessen Dep. (played
12/1/98am), at 63:22 - 66:1; Clark Dep., 7/22/98, at 44:25 - 46:16
(DX 2562); Schell Dep., 9/15/98, 103:17 - 104:22 (DX 2562).
- Professor Fisher testified: "Netscape's browsers contain their own
set of APIs (as well as a set of Java APIs) to which applications
developers can write applications. As a result, applications can be
developed that will run on browsers regardless of the underlying
operating system." Fisher Dir. ¶ 84; see also Warren-Boulton Dir. ¶
69; Warren-Boulton, 11/23/98pm, at 34:12 - 35:13.
53.3. Third, Internet browsers, including Netscape Navigator, have been
ported to multiple operating systems, thereby enabling application developers to write
cross-platform applications using browser APIs. Applications written for the browser will
run on multiple operating systems.
- Dr. Tevanian described the importance of Internet-oriented
platforms, including browsers: "Importantly, applications written for
such platforms would be able to run on any computer that has the
software platform, regardless of the underlying operating system."
Tevanian Dir. ¶ 45.
- As Professor Fisher summarized: "The browsers produced by
Netscape run on many different operating systems, including
Windows, the Apple Macintosh operating system, and various
flavors of the UNIX operating system." Fisher Dir. ¶ 83; GX 13
(listing 22 operating systems on which Netscape Navigator runs);
see also Schmalensee, 6/21/99am, at 20:10 - 21:7 (explaining how
the Web and the browser serve as a platform).
- Microsoft's Paul Maritz, among others, recognized that Netscape's
browser represented an alternative platform to which ISV's may
write cross-platform programs. Maritz, 1/25/99pm, at 28:7-11; see
also McGeady, 11/9/98pm, at 56:4-25 (describing Maritz's comments to
Intel about how Netscape's browser posed a "cross-platform threat").
- Andreessen testified "that because Navigator or Communicator
tend to support more operating system platforms, it's easier to write
a cross-platform application." Andreessen Dep., 7/15/98, at 165:11
- 166:6 (DX 2555); see also Clark Dep., at 7/22/98, 48:21 - 49:21
(DX 2562) (explaining that Netscape's objective was "to provide a
computer- and operating system-independent layer for applications
that were network based to be developed").
- Dr. Warren-Boulton testified: "The issue is not Netscape as a stand-alone
alternative to Windows. The issue is . . . the existence of an independent
browser industry supporting cross-platform standards in encouraging a set
of applications which is large enough so that someone will provide a
platform." Warren-Boulton, 11/23/98am, at 80:8-13.
54. Internet browsers, in particular Netscape Navigator, thus posed a threat to
Microsoft's operating system monopoly because they threatened to reduce the
applications barrier to entry; in the words of Bill Gates, non-Microsoft browsers
threatened to "commoditize" Windows.
- Bill Gates, "The Internet Tidal Wave," May 26, 1995. GX 20, at MS98
01128763.
- Barksdale summarized the threat posed by Netscape to Microsoft's
Windows monopoly: "These innovations arising from the development of
browser technology, particularly Navigator, were eventually noticed at
Microsoft. The possibility of a vast library of applications written in Java or
other OS-neutral languages coupled with independent user interfaces and
platforms, such as those provided by Navigator, posed a serious threat to
the Windows monopoly." Barksdale Dir. ¶ 85.
- As Barksdale pointed out, given the APIs, whether extensive or limited,
exposed by Netscape Navigator, the "big threat" to Microsoft "would be that
if developers began developing for the browser and because it was across these
19-some platforms as I mentioned, it then has the potential that OEM's could put
different types of operating systems on their machines because the other programs
and applications out in the general market would be able to run on top of the
browser and not be particular as to which operating system was installed with the
PC." Barksdale, 10/27/98pm, at 4:19 - 5:9; see also Barksdale, 10/27/98am, at
74:10-16 (explaining that if Netscape's browser were successful, it could
potentially "marginalize or commoditize the platform characteristics of the
operating system beneath it").
- Professor Fisher testified: "To the extent that browsers support
applications independent of the operating system, they could erode the
applications programming barrier to entry that protects Microsoft's
monopoly in operating systems." Fisher Dir. ¶ 82; Fisher Dir. ¶¶ 85-86, 90
(collecting internal MS documents; citing GX 354, GX 473, GX 510, GX
1016); Fisher, 1/12/99pm, at 68:20 - 69:2 (explaining that OS and Java
threaten to facilitate a substitute's entry).
55. Non-Microsoft browsers posed an especially serious threat because
network-based computing in general, and the Internet in particular, quickly blossomed
into a very important way users employ their PCs; if Microsoft were unable to control
the standards and interfaces that are central to network-based computing, other firms
could develop rival platforms using those standards and interfaces and would be able to
challenge the applications barrier to entry.
- In his May 1995 memo, "The Web is the Next Platform," Microsoft's Ben
Slivka wrote that "we should be extending the web with as many Microsoft
technologies as possible, even if we have to modify those technologies in
ways not original [sic] intended by their designers." He concluded: "If
Microsoft doesn't enhance the Web, there is a nightmare scenario where
an OS-neutral Web platform arises, and then a company like Matsushita
or Siemens could come out with a $500 'Web Box' that runs web
applications (with no need for Windows, or MS-DOS compatibility, or Intel
compatibility), and consumers make the obvious choice between a $2000
Windows PC and the $500 Web Box. Say good-bye to Windows." GX
21, MS98 0102397.
- A June 1996 Microsoft marketing report, "Winning @ Internet Content,"
states: "The rise of the Internet has been driven by the success of a
series of 'platforms' that utilize these protocols at their core and provide a
set of APIs for ISVs to develop on top of. By far the most successful
platform to date has been Netscape's, with Netscape Navigator on the
browser and Netscape Suite Spot on the server. The core threat for
Microsoft is the potential for this platform to abstract the Win32 API. For
example, if Netscape continues its success in getting ISVs and ICVs to
develop applications for Netscape's client/server Api's, these API's could
be the most important API's in the future, putting Win32 and Microsoft's
platform position in jeopardy." GX 407, at MS6 5005709.
- See also infra Part VII.D.
2. Microsoft recognized the threat that Internet browsers, in
particular Netscape Navigator, posed to its operating system
monopoly
56. Microsoft recognized that Internet browsers not controlled by Microsoft could
threaten its monopoly by eroding the applications barrier to entry.
56.1. The contemporaneous documents show that Microsoft's executives
recognized the browser threat and developed their business strategy to respond to it.
- In a May 26, 1995, memo entitled "The Internet Tidal Wave," Gates
announced to the rest of Microsoft that he assigns "the Internet the
highest level of importance. In this memo I want to make clear that
our focus on the Internet is critical to every part of our business.
The Internet is the most important single development to come
along since the IBM PC was introduced in 1981. It is even more
important than the arrival of the graphical user interface (GUI)."
Gates identified "a new competitor 'born' on the Internet" --
Netscape. "Their browser is dominant, with 70% usage share,
allowing them to determine which network extensions will catch on.
The are pursuing a multi-platform strategy where they move the
key API into the client to commoditize the underlying operating
system." GX 20, at MS98 0112876; see also GX 16; GX 17; GX
336, at MS7 007443; Gates, 1/13/99, at 460:15 - 461:10, 407:9-18
(Gates stated that Netscape was "creating a product that would either
reduce the value or eliminate demand for the Windows operating system if
they continued to improve it and we didn't keep improving our product.")
- iMcGeady described what Microsoft executives, including Mr. Gates, told
Intel about its view of Netscape in 1995: "If you begin to get a few
leading-edge application developers that are developing for the
Netscape environment, then that makes that environment that
much more attractive both for end users and for other applications
developers. And so more applications developers come to up [sic]
which brings more users to it and more application developers,
that's the positive feedback loop. That's what he wanted to prevent
happening, that kind of a feedback loop which everyone seeks in
this industry . . . If independent software developers began to write
applications or plug-ins that worked directly with the browser, then,
first of all, they may not--they may no longer write them to work
directly with Windows, but more importantly, then Netscape begins
to be the one who is setting--who is defining those application
programming interfaces we discussed earlier, and Netscape then is
much more in control of the rate of innovation and the kinds of
innovations that happen for those applications, and Microsoft is,
correspondingly, less in control." McGeady, 11/9/98pm, at 59:22 -
60:11; see also McGeady, 11/9/98pm, 57:10 - 58:8; GX 279, at MS
CID 00077 (Notes of an August 2, 1995 meeting with Mr. Gates).
- Maritz wrote in May 1995 to other senior Microsoft executives that
"we all agree . . . that the Internet represents a big
threat/opportunity to our current businesses" and that "Priority #1 is
to not lose control of key interfaces and protocols that
applications/titles use. O'Hare needs to evolve into an extensible
client that encourages 'online applications' to take full advantage of
Windows and other MS assets." GX 148. Maritz, 1/28/99am, at
56:20 - 57:1 (Maritz explaining that Navigator is a threat to
Windows "if more and more application programs get their services
from Navigator and not from Windows, the perceived value of
Windows is going to decline, and the ability to have those
applications moved to other platforms will also be increased"); see
also GX 503, at MS6 6008248.
- In his May 1995 memo, "The Web is the Next Platform," Ben Slivka
wrote that "The Web is an application platform (complete with APIs,
data formats, and protocols) that threatens Windows -- many
corporate developers and ISVs could develop and deliver their
solutions more quickly, to a wider audience, with the Web than they
can with Windows or MSN as it exists today." GX 21, at MS98
0102395; see also GX 329; GX 399, at MS98 0103343 (Ben Slivka
wrote: "The Web could make Windows irrelevant in the next few
years."); GX 521, at MS98 0103337; Slivka Dep., 1/13/99, at 724:1-8 (Slivka testified: "You know, whether it was Navigator 1 or Navigator
2 or Navigator 3, the point was not that that thing as it stood then would
immediately kill Windows. . . . The point was that that thing could grow
and blossom and provide an application development platform which was
more popular than Windows.").
- Brad Chase described in an April 1996 planning memo how
Microsoft would lose "the Internet platform battle" if it did not
increase consumer usage of Internet Explorer: "The industry would
simply ignore our standards. Few would write Windows apps
without the Windows user base. . . ." GX 39, at MS6 5005720. He
goes on to say that, "Netscape is already entrenched in our
markets all over the world. The situation today is scary." GX 39, at
MS6 5005724 (emphasis in original); see also GX 510, MS7
004127 (Chase warned that competing Internet browsers could
eventually "obsolete Windows"); GX 59 (Chase observed in April
1997 that "IE share is critical. Without it, we lose the desktop,
which translates to Windows and Office revenue over time."); GX
828, MS98 0118367 (In March 1998, Chase notes:
- redacted - ")
(sealed); GX 40, at MS6 6005550 (Silverberg writes: "Our
competitors are trying to create an alternative platform to
Windows."); GX 407, at MS6 5005716; GX 475.
- In a May 1997 Internet Explorer 5 Planning document, Chris Jones
analyzed Netscape's approach as follows: "Netscape Communicator
defines a new platform, taking advantage of the lessons learned from
Visual Basic, Visual C++, Java and Web content. They are completely
focused on turning their applications framework (HTML, object model,
scripting, and JFC) into the primary way developers deliver Internet-centric applications." GX 494, at MS7 004614. Mr. Jones also testified
that "as soon as the internet came around . . . it was clear that you could
take and create something that extended and enhanced what was on the
internet and a set of services that are HTML and create an alternate
environment that wouldn't need Windows anymore, that would abstract
away all the value that Windows provided and make it just a general
purpose--to quote a Netscape vice-president--partially debugged device
drivers. And boy, you know, I'm not in the business of shipping partially
debugged device drivers." Jones Dep., 1/13/99, at 574:24 - 575:22, 578:2-14 ("If you mean did we think that the Netscape browser was a platform
threat, the answer to that question is yes because the services that it
provided were compelling alternatives to the services on Windows.").
- Microsoft's Yusuf Mehdi agrees that "having users use our software . . . is
an important goal for us to defend the Windows market share and provide
a platform for those developers to write to. And to the extent at that
Netscape would have a more popular platform that people wrote to and
used instead, that would be a threat to the business for the Windows
business for Microsoft." Mehdi Dep., 1/13/99, at 637:14 - 638:22.
- As Dr. Warren-Boulton summarized: "Microsoft clearly regarded
Netscape, particularly initially, as a direct threat to its operating system in
the sense that Netscape might, in fact become . . . a complete and direct
competitor." Warren-Boulton, 12/1/98am, at 42:14-20; see also
Warren-Boulton Dir. ¶ 87 (collecting quotes from Microsoft
personnel, citing GXs 20, 39-40, 503, 510).
56.2. At trial, Microsoft's witnesses acknowledged that Netscape
Navigator posed a competitive threat to Windows because it provided an application
platform that threatened to erode the applications barrier to entry.
- Dean Schmalensee testified that "Netscape apparently envisioned
pursuing a middle ware strategy to compete with Windows.
Netscape Navigator relied on APIs in Windows and in that sense
was an application. In addition to expanding its features, Netscape
promoted its client products as 'platforms,' and encouraged ISVs to
write to them by providing APIs and other 'hooks,' and offered
services and software tools . . ." Schmalensee Dir. ¶ 137;
Schmalensee, 1/13/99pm, at 33:21 - 34:5 (agreeing Netscape and
Java are threats to Microsoft because applications written to those
platforms "can be run cross-platform"); Schmalensee, 1/13/99pm,
at 35:5-14 ; Schmalensee, 6/21/99am, at 23:10-19 ("I believe that
Netscape was a potential platform competitor, and Java was
certainly by -- was and is, by any definition, an actual platform
competitor.").
- Allchin agreed that Netscape's browser posed a platform threat to
Windows. Allchin, 2/1/99pm, at 55:22; Allchin, 2/1/99pm, at 60:23-25
(conceding that the "web application platform" was a threat to Windows
and that integrating the browser into Windows was a response to that
threat); Allchin, 2/1/99pm, at 60:3-4 ("they were a platform competitor,
absolutely"); Allchin, 2/3/99pm, at 8:20-22 (discussing GX 47: "by this
time it was obvious to me that Netscape was certainly adding enough
APIs, that that was the competitor to Windows."); Allchin, 2/3/99pm, at
9:1-8, 10:9-15, 28:12-15.
- Maritz stated that he considered Netscape both an actual platform
competitor, "in terms of how people could structure applications,"
and a "potential" platform competitor. Maritz, 1/26/99am, at 28:13-23; Maritz, 1/26/99am, at 30:4-6 (Microsoft's "initial concerns about
Netscape focused on their ability to expose API's and their ability to
expose new facilities to web pages."); Maritz, 1/25/99pm, at 26:20 -
27:19 ("During the first half of the calendar year 1995," Microsoft
came to believe that "Netscape was becoming a platform . . . that
other software could depend upon, and they were extending it's
capability as a platform. And one of the natures of a software
platform is that it exists to enable other software and if the other
software is depending upon your competitor's platform, even if it's
running on top of your own platform, over time the value of the
platform can become diminished . . . .").
C. Cross-Platform Java also presented a middleware threat to
Microsoft's operating system monopoly
57. Cross-platform Java is another middleware technology that has the potential
to erode the applications barrier to entry by gaining widespread usage of APIs without
competing directly against Windows as an operating system.
1. The nature of the Java threat
58. James Gosling and others at Sun Microsystems developed Java in
significant part to provide developers a choice between writing cross-platform
applications and writing applications that depend on a particular operating system.
58.1. Java consists of a series of interlocking elements designed to
facilitate the creation of cross-platform applications, i.e. applications that can run on
multiple operating systems.
- Gosling testified: "The Java technology is intended to make it possible to
develop software applications that are not dependent on a particular
operating system or particular computer hardware . . . . A principal goal of
the Java technology is to assure that a Java-based program -- unlike a
traditional software application -- is no longer tied to a particular operating
system and hardware platform, and does not require the developer to
undertake the time-consuming and expensive effort to port the program to
different platforms. As we said in the Preface to The Java Programming
Language, 'software developers creating applications in Java benefit by
developing code only once, with no need to 'port' their applications to
every software and hardware platform.' . . . Because the Java technology
allows developers to make software applications that can run on various
JVMs on multiple platforms, it holds the promise of giving consumers
greater choice in applications, operating systems, and hardware. The Java
technology has the potential not only to free individual consumers from
concern about whether the software they want to run is supported by a
given operating system, but also to permit corporations and Internet users
more easily to mix different types of computing systems across a
network." Gosling Dir. ¶¶ 20-29(b); see also Gosling, 12/3/98am, at
6:3-6.
- Gosling stated this theme in internal documents as early as August 1995:
"The issue of making developers CPU and OS independent is that they can
port to Sun or to Windows. . . . Sun's or any alternate CPU company
(MIPS and SGI) key to success is apps. Apps are the key to volume. Java
allows developers to decrease their dependence on Intel and Microsoft."
DX 1285; see also DX 2012, at SUN 87 001685 ("Sun is attempting
to establish Java as a viable computing platform which is hardware
and operating system independent.").
- In his June 1999 rebuttal testimony, Dean Schmalensee recognized "Java
is used for a wide range of things. It's how my son first learned computer
programming. It's used to run on a wide array of platforms. That, of
course, is one of its most important selling features from Sun."
Schmalensee, 6/23/99pm, at 50:5-11.
- As the District Court for the Northern District of California found: "Sun's
JAVA Technology comprises a standardized application programming
environment that affords software developers the ability to create and
distribute a single version of programming code which is capable of
operating on many different, otherwise incompatible, system platforms
and browsers. Most computer systems implement platform-dependent
programming environments, such as Microsoft's Win32 programming
environment. Programs created to run on a particular platform will not
function on a different platform. Thus, a software developer must choose
the platforms for which it will develop and support different versions of
the same program. Sun's platform-independent JAVA Technology, which
can be implemented on many different system platforms and browsers,
obviates the need for creating and supporting different versions of the
same program." Sun Microsystems, Inc. v. Microsoft Corp., 999 F. Supp.
1301, 1302 (N.D. Cal. 1998).
58.1.1. Java provides ISVs a programming language with which to
write applications. Java also includes a set of "class libraries," a collection of programs
written in Java, that offer APIs that ISVs can use to develop software applications.
- Microsoft's Paul Maritz summarized the different aspects of Java:
"Java, the programming language; Java the virtual machine, which
you need to execute Java programs; and then there's this collection
of other programs written in Java, which I call the Java classes.
And it's that collection of software that is being put forward by
Sun, and that Netscape has announced their intent to cooperate
with Sun, that forms another body of middleware that I am
concerned about." Maritz, 1/26/99am, at 18:22 - 19:23.
- As the District Court for the Northern District of California
summarized: "The Java programming environment allows
software developers to create a single version of program
code that is capable of running on any platform which
possesses a compatible implementation of the Java runtime
environment. The Java programming environment
comprises (1) Sun's specification for the Java language, (2)
Sun's specification for the Java class libraries and (3) the
Java compiler." Sun Microsystems, Inc. v.
Microsoft Corp., 21 F. Supp. 2d 1109, 1112 (N.D. Cal.
1998).
58.1.2. The Java programming environment also provides software
developers a Java "virtual machine" (JVM) that, when ported to different operating systems,
serves as the "host" or "adaptor" to which programs written in the Java language can be run,
independent of the underlying operating system. Together, the Java class libraries and virtual
machine are often referred to as the "Java runtime environment" (JRE).
- As Gosling explained, Java makes it much easier for ISVs to
develop platform-independent software because Java programs
"need not run by interacting with a particular operating system's
APIs. Instead, they typically interact with a Java virtual machine
('JVM'), which is an intermediate software layer that translates the
Java-based program for the particular operating system and
hardware platform that the Java virtual machine runs on. In
essence, the Java-based program views the JVM as an operating
system, and the operating system views the JVM as a traditional
application . . . once a JVM is developed for a software platform, if
the JVM is fully compliant with the Java specifications, . . . it
should run most Java-based programs without the need to
recompile or otherwise modify the programs. . . . Such programs
thus have the potential to run on any PC, other type of computer, or
even devices not traditionally thought of as computers (such as
cellular telephones), provided that the machines have compatible
JVMs installed on them."; see also Gosling Dir. ¶¶ 24, 25, 28;
Gosling, 12/2/98am, at 55:5-11 (any Java program, if written
properly and properly compiled into bytecodes, should run
equivalently on any properly-designed and implemented Java
virtual machine, regardless of the underlying platform is); Gosling,
12/10/98pm, at 21:8-14 (describing the JVM as an adaptor).
- Sean Sanders, an executive with Novell Corporation, testified:
"The Java virtual machine is essentially yet another software layer
that allows people to run Java-based applications and to help them
-- provide them the tools that they could use to yet build and
develop and to -- for optimization of any other Java applications
that they might want to develop." Sanders Dep., 1/13/99, at
188:18 - 189:15.
- The District Court for the Northern District of California described
the Java architecture as follows: "Sun's JAVA Technology is a
so-called 'class-based' language in that its functionality is
determined by the Java classes available to the programmer.
Therefore, new functionality requires developing new Java classes.
Programs written in the Java programming language are compiled
into intermediate instructions called bytecodes or Applets. These
bytecodes or Applets are then 'interpreted' by another computer
program which emulates a hypothetical CPU called the Java
Virtual Machine. The Java Virtual Machine translates the Applets
into instructions understood by the specific computer CPU on
which the Java Virtual Machine is running. Therefore, a specific
interpreter or virtual machine is needed for each computer CPU on
which the Java program is run." Sun Microsystems, Inc. v.
Microsoft Corp., 999 F. Supp. 1301, 1302-03 (N.D. Cal.
1998) (citations omitted.).
58.2. Because Java offers alternative APIs, applications written using standard
Java programming tools and class libraries can run on any operating system for which there is a
Java virtual machine. The widespread adoption of a cross-platform Java programming
environment could reduce computer users' dependency on the Windows operating
system.
- As Gosling explained: "As more new Java-based programs are
developed, distributed and used, new operating systems may be
developed to take advantage of the existing body of Java-based
software. In other words, potential developers of new operating
systems and hardware platforms need not be deterred by the
absence of platform-specific programs for their new systems, so
long as there is a JVM available to enable existing Java programs
to run on the systems. This may give new operating systems and
hardware platforms a chance to compete in markets previously
dominated by a particular vendor." Gosling Dir. ¶ 29; Gosling,
12/10/98pm, at 28:20 - 29:2 ("Once the APIs that developers develop to
are ones that are realized on many different operating systems, then those
operating systems can compete with Windows. And that would lead to
sort of a lesser -- lesser role for Windows in that they would have to
compete with these operating systems on the merits of the operating
system rather than on the lock that tends to be inherent in the APIs and the
binary compatibility.").
- Soyring explained IBM's rationale for supporting Java: "Primarily
because the value that it provides to IBM's customers and the value it
provides to IBM. As you probably know, IBM has a variety of operating
systems, primarily four different ones. Many of our customers have many
of these different -- several -- one or more of these operating systems
installed. It's less expensive for them and less time-consuming for them to
be able to buy one application or one software product that they can buy,
maintain, and support but run it on different operating system platforms."
The success of cross-platform Java would enhance the ability of other
operating systems "to compete." . . . What drives demand for the sales of
operating systems is the availability of applications. And if there is a large
install base of Java that's consistently implemented, what it does is create
an economic opportunity for commercial software developers to be able to
develop a commercial software application using Java and then make it
available to sell and be run on many different operating systems rather
than just on one." Soyring, 11/18/98pm, at 54:8 - 55:10; see also Soyring
Dir. ¶ 28 ("The Java technology from Sun is designed to allow
Java-compatible application programs to run on a wide variety of different
hardware and operating systems. This would provide users with the
benefits of increased number of applications and would reduce the cost of
ISVs of developing applications for multiple operating systems. This
characteristic of Java also has the potential to undermine the Windows
application advantage . . . .").
- Barksdale testified that "the cross-platform benefits of Java, allowed for
the development of software applications that were directed more to the
Internet than to the desktop, and thus had the potential to serve as a partial
substitute for the Windows OS as a development platform." Barksdale
Dir. ¶ 15; see also Sasaki Dep. (played 12/16/98pm), at 31:24 - 32:7
(explaining that Java has the potential to level the playing field among
operating systems).
- Dean Schmalensee readily acknowledged that cross-platform Java
technology poses a competitive threat to Windows by potentially
rendering underlying operating systems less significant: "Sun's Java
poses potentially serious competitive implications for Windows. . . . If
Java achieves its advocates ambitions, operating systems would become
less important in the marketplace, and the important standards would come
to be determined by Sun, which vigorously defends its control over the
Java language." Schmalensee Dir. ¶¶ 141-142; see also Warren Boulton,
11/19/98pm, 31:8-12 (concluding "an increasing number of users may be
able to simply do without Windows entirely").
2. Microsoft recognized the Java threat
59. Microsoft understood the threat Java posed to its monopoly power. Java offered
ISVs the ability to create a robust set of cross-platform applications that might reduce the
applications barrier to entry.
59.1. Java provided software developers with a platform to create
applications that could run on different operating systems and hardware platforms.
- Eubanks testified: "One of the great things with Java is that when
you create a Java application, it will run on any machine that has a
Java virtual machine." Eubanks, 6/16/99am, at 68:11-20.
- Gosling, whose responsibilities include working with numerous
application developers and who himself has a career of experience as a
developer, made clear that Java's theme of "'write once, run anywhere'
was terrifically attractive to developers. Developers want this more than
just about anything you can imagine." Gosling, 12/3/98am, at 32:10-12.
- Soyring testified: "'Write once run everywhere' . . . has been the
holy grail of programming for many years to be able to write an
application once and then run it on many different operating
systems or hardware platforms, and we find that Java is a
technology that most closely approaches this by a long distance
versus any other technology, and we have been able to
successfully demonstrate with a set of our clients that it is possible
using the Java technology to write an application once, compile it
once and then run that exact same code on a variety of different
operating systems, giving our customers the choice to choose
different operating systems and different hardware platforms."
Soyring, 11/18/98pm, at 51:18 - 52:6; see also Sanders Dep.,
1/13/99, at 186:20 - 187:3 (stating that Java "provides the benefit of
an application-running environment that would allow people to run
applications independent of any kind of operating system or cpu
type of restraints they may currently be facing").
- Barksdale testified that "Java allows software developers to write cross-platform applications that will run on any operating system, increasing
consumer flexibility and ease of use, while reducing development costs
associated with writing an application and then porting it to run on various
different operating systems . . . . The Java programming technology
solves the platform dependency problem that has so long plagued software
development. Programs written in Java can be run on any platform that
has a Java virtual machine and Java class libraries, which Navigator does."
Barksdale Dir. ¶¶ 15, 83.
59.2. Microsoft recognized, and continues to recognize, the competitive
threat that Java poses to Windows by providing an attractive cross-platform
programming environment that could erode the applications programming barrier to
entry.
- Dr. Warren-Boulton summarized the evidence of Microsoft's
perception of the competitive threat posed by Java: "Microsoft has,
almost from the beginning, recognized that the clearest threat to
that monopoly power is the emergence of Java technologies
combined with an independent browser market. Their response to
that threat has been to attempt to take that technology, and instead
of making it multi -- cross-platform, has been to transform that
technology into a technology that is Windows-specific so as to
prevent the emergence of a large stock of applications that could
be used on any operating system . . . ." Warren-Boulton,
12/1/98am, at 19:24 - 20:8.
- Dr. Fisher, similarly, summarized the evidence that Microsoft treated Java
as a significant competitive threat to Windows. Fisher Dir. ¶¶ 204-207.
59.2.1. Microsoft executives have throughout the past four years
treated cross-platform Java as a serious threat to Microsoft's operating system
dominance.
- In a June 1996 e-mail to Microsoft executive staff, Paul Maritz
focused on the need to "fundamentally blunt Java/AWT
momentum and to re-establish ActiveX and non-Java approaches
as a viable strategy for structuring software." The reason Mr.
Maritz provided for this objective was to "protect our core asset
Windows -- the thing that we get paid $'s for. While Java per se is
not the problem, if everything & everybody moves to Java as a
language, then it will be so much more easy for AWT to become
the API, and Windows is damaged." GX 42, at MS6 6010347
(emphasis in original); see also GX 473, at MS6 6006237 ("Java.
Gaining as scripting language . . . Class libraries define 'API.'
Becoming the 'brand' for software components."); GX 504, MS98
0169096 (Maritz writing that:"Java. Sun's goal is: -- Java class
library/runtime = new OS API -- leverage this new API to replace
Windows by JavaOS.").
- In August 1996, Bob Muglia wrote: "When a Java developer writes
to AWT, they are writing to Sun APIs, and their application can be
easily run on competitive platforms." GX 466, at MS6 5003781;
see also Muglia, 2/26/99pm, 10:5 - 11:22 (by offering an
alternative platform, Sun could get developers to write to the Java
platform and not to Windows, and therefore the applications that
they wrote would not be focused on Windows).
- In September 1996, Adam Bosworth sent Bill Gates and others an
e-mail discussing Java. Bosworth noted, "I think it is important to
understand that Java is not just a language. If it were just a
language, it would not be a threat to us. We would and could
easily just build the best implementation of this language and be
done. It is, however, much more. It is an alternative to COM. . . .
Java is on Unix and requires no dealing with setup, install, de-install, or anything else. Thus it is really easy to understand how a
system for dynamically authoring Web pages on the server that
depended upon Java objects rather than COM ones would have
wider appeal."Gates responds: "This scares the hell out of me. Its
still very unclear to me what our OS will offer to Java client
applications code that will make them unique enough to preserve
our market position. Understanding this is so important that it
deserves top priority." GX 983, MS7 032895.
- In January 1997, an internal Microsoft analysis described the
"platform challenge" posed by Java: "possible emergence of a set
of APIs and underlying system software that lead to a lesser or no
role for Windows." GX 51, MS7 005534.
- In a February 1997 e-mail to Jim Allchin, Mr. Gates again
addressed the cross-platform threat posed by Java: "What will we
have that the Java Runtime will not have? . . . The fact is that
applications can be run on the server against an HTML client. . . .
Most applications will have very little client code in the future. . . .
The fact is there will be lots of machines where HTML/some level
of Java is all they will have in common. Cheap devices and old
PCs will be like this. It makes it very easy for people to think they
should just program to this. . . . Lets work together to find the
solution to this. I can say I am more scared than you are but that is
not what will help us figure out where we should go." GX 475; see
also GX 590 (Gates writing: "Java is the biggest threat to us and I
certainly shouldn't be doing Apple events unless we are getting
some help from us on this.").
59.2.2. Microsoft's witnesses in this litigation conceded that Java
presented a significant potential threat to Windows.
- Gates testified repeatedly that he perceived Java to be a threat to
Windows: Gates Dep., (played 12/2/98am), at 22:19 - 23:1.
Gates stated: "we did think of" Java APIs "as something that
competed with us for the attention of ISV's in terms of
whether or not they would take advantage of the advanced
features of Windows." Gates Dep. (played 12/2/98am), at
24:15-22; see also Gates Dep., 8/27/98, at 90:12-19 (DX 2568).
- Muglia also testified that Microsoft considered Java a
serious cross-platform threat: ""Although Java was a new and
unproven technology, Microsoft took Sun's claims seriously. . . .
Sun has adopted a business strategy that seeks to transform the
Java programming language into a full operating environment and
software development platform. A key requirement of Sun's
strategy is delivering on its WORA claim -- that programs written
in Java, to the Java development platform, will run without
modification on any underlying platform for which there is a
JVM." Muglia Dir. ¶¶ 8, 10; see also Muglia, 2/26/99pm, at 4:8-18 (Muglia believed in 1995 and 1996 that Java represented a
serious threat to Microsoft's operating system business); Muglia,
2/26/99pm, at 7:2-19 (explained that the cross-platform threat
consisted of the JVM and Java class-libraries) ; Muglia,
2/26/99pm, at 9:3-21 (explained that: "what they were trying to do
was get developers to write to that alternative platform. So, even if
-- even if a developer wrote a Java program and that program runs
on Windows, even in the case where it runs on Windows, it's not
written to Microsoft's programming interfaces. So when I said
slide in their platform, what I meant is that they could, in essence,
make what everything else that our platform did irrelevant, thus
enabling to replace Windows and make it obsolete, so to speak.").
- Maritz also testified: "If successful, software developers could
write programs to run on Sun's technology, and neither Windows
nor any other operating system would provide significant value to
customers." Maritz Dir. ¶ 243; Maritz, 1/26/99am, 20:23 - 21:3
(the Java foundation classes posed a potentially serious platform
threat); Maritz, 1/28/99am, 59:10 - 60:17, 62:3 - 63:17 (Maritz
explained Java is a form of middleware. Sun's goal was to provide
most of the OS services through the Java runtime. The browser and
Java have the potential to serve as a virtual operating system.).
- Dean Schmalensee also acknowledged the cross-platform threat
Java poses to Windows: "Sun would like ISVs to write pure Java
so that their applications can run anywhere, in principle. Microsoft
would like ISVs to design applications that would run on
Windows. It matters to those companies what choice the ISV
makes, assuming it's a good application." Schmalensee,
6/22/99pm, at 23:23 - 24:7.
- Slivka also testified regarding the Java threat: "my recollection
was that this cross-OS Java platform stuff that we were attempting
to do with AFC, he [Bill Gates] was very unhappy with that." "He
thought that was a big threat to Windows." Slivka Dep., 9/4/98, at
367:13 - 369:3 (DX 2591); see also, Slivka, 1/13/99, at 735:13 -
736:4 ("All this comes back to Windows and the threat, you know,
Sun's very direct threat to our Windows platform, and the success
of Windows on the client. So, this seemed like if the library space
fragmented, the 'write once, run anywhere,' I guess, actually is
what Sun called it, that would be a lot less probable . . . I guess the
end was to protect the Windows franchise, not to defeat the 'write
once, run everywhere.'").
D. The threats to Microsoft's monopoly posed by Internet browsers and
Java are mutually reinforcing, and they could be essential to the
emergence of other platform-level threats to Microsoft's operating
system monopoly
60. The competitive threats posed by non-Microsoft Internet browsers and
cross-platform Java are, to a significant degree, interdependent.
60.1. Dissemination of Java virtual machines and Java runtime
environments not controlled by Microsoft hinges in significant measure on the
widespread distribution of non-Microsoft Internet browsers.
60.1.1. Industry witnesses recognize that Internet browsers are the
principal distribution vehicle for Java Virtual Machines and JREs and that, because
Microsoft distributes only its own (as will be discussed below, non-cross-platform)
implementation of the JRE with its browser, Netscape Navigator was the principal
distribution vehicle for cross-platform Java.
- IBM's John Soyring testified that Netscape has been a
significant distributor of Java virtual machines: "Netscape is
a very high-volume distribution vehicle for Java virtual
machines on operating systems other than OS/2." Soyring,
11/18/98am, at 89:8-12; see also Soyring Dir. ¶¶ 28 ("The
reason this relates to browsers is that Netscape Navigator
has been the prime distribution vehicle for Sun's Java
technology while Internet Explorer contains the Microsoft
version of Java.").
- Barksdale testified that "the widespread distribution of
Netscape Navigator facilitated widespread distribution of the
Java programming language developed at Sun
Microsystems." Barksdale Dir. ¶ 15; see also Sasaki Dep.
(played 12/16/98pm), at 31:6-8; 32:8-11.
60.1.2. Microsoft, both in contemporaneous documents and
through its witnesses at trial, recognized that Internet browsers are essential to
distribute JVMs and Java class libraries and, in particular, that Netscape was the
principal distribution vehicle for a cross-platform Java runtime environment.
- Muglia acknowledged at trial that Netscape has been "one of the
largest volume distributors of JVMs." Muglia Dir. ¶ 15.
- Maritz conceded that Netscape, in May and June 1995, "was
an important distribution vehicle for Java APIs." Maritz,
1/26/99pm, at 59:21 - 60:6; Maritz, 1/26/99am, at 30:10 -
31:2.
- Documents written by Maritz in 1997 expressly link
Netscape nad Java as a threat. GX 52, MS7 003270
(January 1997 Microsoft presentation identifies as a
"Scenario: Emergence of a new API" and notes that "Sun
AWT provides base cross-platform API" and further, that
"Navigator/NetOne provides: additional API's" and "a volume
platform for ISVs & Corps to target, since runtime gets
shipped with Navigator"); GX 113; GX 514, at MS7 007509
("If we look further at Java/JFC as being our major threat,
then Nscp is the major distribution vehicle.").
60.2. Conversely, the ability of Internet browsers to supply an attractive
set of APIs is enhanced by the viability of cross-platform Java APIs. The browser and
Java APIs sets can together provide the foundation for developers seeking to write
cross-platform applications, particularly network- and Internet-oriented applications.
- Contemporaneous Microsoft documents describe the
interdependence of competitive browser and Java products. E.g.,
GX 466, at MS6 5003781 ("Without question, the Java platform
API's have surpassed the Macintosh as the #2 platform for
software development. In concert with this, Netscape has its own
offering of platform API's called Netscape One which is also built
on Java. Collectively, these two initiatives represent the most
serious threat to our core Windows business which Microsoft has
seen in years. The Windows franchise is fueled by application
development which is focused on our core APIs. When a
developer writes an application to AWT, even if they are using
Windows and Visual J++, they are not supporting our platform.
Instead, they are furthering Sun's momentum, potentially opening
up the opportunity for our competitor to slide in its own operating
system offering." ); GX 485, at MS6 5005195 ("The Internet
challenge is critical as Netscape, Sun and others try to build a non-Microsoft platform alternative.").
- Gosling also summarized how browsers and Java technology together can
be particularly significant for Internet-oriented applications: "Because the
Java technology is particularly useful for running software that is
downloaded over a network, such as the Internet, we adapted the Java
technology to work in conjunction with web browsing programs known as
'browsers.' . . . Java technology in essence permits certain software
programs to run within browsers. Java-based programs can be
downloaded from the Internet or other network to a user's computer
without regard to what operating system or hardware is installed."
Gosling Dir. ¶¶ 34-35.
- Dr. Warren-Boulton also explained that competitive browsers may over
time competitive browsers tend to threaten the Windows monopoly more
as a complement to, and distribution vehicle for, Java, rather than as an
independent platform in its own right. Warren-Boulton, 12/1/98am, at
42:7 - 43:10; see also Warren-Boulton, 11/19/98am, at 48:13-24
(Java an implicit complement to browsers).
61. Because of the growing importance of network computing (over the Internet
and otherwise), Internet browsers and Java in combination posed a serious threat to
the applications barrier to entry.
- See infra Part VII.D; ¶¶ 398-400.
62. The success of cross-platform browser and Java products could also
facilitate innovation in new forms of computer hardware.
- As Professor Fisher explained: "Similarly, browsers could reduce
the power of the operating system monopoly by facilitating the
expansion of network computing, in which users with 'thin clients'
use a network to access applications residing on a server computer
rather than hosting the application on the PC itself." Fisher Dir. ¶
87.
- In an April 1997 Memo entitled "Preserving the desktop paradise,"
Brad Chase commented that Netscape and Sun might not only
reinvigorate operating system software competition, but also
facilitate the success of low-cost hardware: "Our competitors are
still hard at work trying to obsolete Windows. More people than
ever now believe they will. Netscape and Sun endeavor to
commoditize the OS and drive developers to adopt their
technologies and APIs. This is more true today than ever and
these technologies are precisely those that may make the NC
viable." GX 512, at MS7 004149; see also DX 1490, at MS7
007476 (identifying network computer as a "competitive threat").
- Maritz also focused on the potential for new hardware
development, facilitated by browser and Java, in his trial testimony.
Maritz Dir. ¶¶ 31, 259 ("impending competition from so-called
'network computers'").
- As Microsoft's Ben Slivka stated in his deposition, a "nightmare
scenario is that the web grows into a rich application platform in an
operating-system neutral way, and then a company like Siemens or
Matsushita comes out with a $500 'WebMachine' that attaches to a
TV." Slivka Dep., 1/13/99, at 712:6-11 (commenting on GX 1016).
- AOL's Barry Schuler also testified that
- redacted -
In order to achieve that,
- redacted -
Schuler Dep.,
5/5/99, 159:12 - 160:4 (DX 2810A) (sealed).
IV. Microsoft Attempted To Enter Market-Division Agreements To Eliminate
Platform-Level Software That Threatened Its Operating System Monopoly
A. Microsoft tried to eliminate the browser threat by proposing a naked
market-division agreement to Netscape
63. Microsoft initially tried to eliminate the threat non-Microsoft browsers posed
to the applications barrier to entry by attempting to bribe, and later threatening,
Netscape into giving up its core Window 95 web-browsing business. Had Netscape
accepted Microsoft's market-division proposal, Microsoft would have succeeded in
killing the browser threat in its infancy and likely would have acquired a monopoly over
browsers.
1. Microsoft first unsuccessfully sought to purchase or license Netscape's
browser software code
64. Before it fully recognized the threat that Internet browsers posed to its
operating system monopoly, Microsoft unsuccessfully sought to purchase or license the
software code for Netscape's Navigator browser.
64.1. When Microsoft decided that it wanted to offer its own Internet
browser product in late 1994, it opened discussions to license browser software code
with several companies, including Netscape.
- Microsoft's Thomas Reardon "contacted Netscape in the early fall of 1994 and
explained his desire to explore whether Netscape would be willing to consider some
sort of licensing arrangement for the first version of its web browsing software." Rosen
Dir. ¶13 (citing Reardon Dep., 9/9/98, at 153-54, 224).
64.2. Netscape representatives rejected Microsoft's proposal.
- Barksdale testified that Netscape did not want to sell their software
"at the price they [Microsoft] offered. They offered a flat fee of a
couple of million dollars to take us out of the game. And that would
have killed our product in their space." Barksdale, 10/21/98am, at
28:6-10; Barksdale Dir. ¶ 96 ("those discussions did not prove
fruitful because Netscape was not interested in Microsoft's
proposal, which was to purchase the Navigator code for what
Netscape considered to be a low flat fee payment.").
64.3. In late December 1994, in what he described as "a moment of
weakness," Netscape then-CEO James Clark -- without the knowledge of the rest of the
Netscape board -- attempted to reopen those discussions, but he was rebuffed by
Microsoft. The discussions went no further and were not renewed.
- Barksdale Dir. ¶ 97.
- According to Barksdale, Clark said he turned to Microsoft because
of "fear on the part of a small company looking into the eyes of the
world's most powerful software company and kind of feeling that
they might somehow help us if we licensed it to them." Barksdale,
10/27/98am, at 49:22 - 50:1.
- Clark Dep., 7/22/98, at 58:6 - 60:18 (DX 2562)
2. When Microsoft recognized the threat that Netscape's browser
posed to its monopoly, Microsoft set out to eliminate the threat
by seeking Netscape's agreement not to compete and to
divide the browser market
a. Microsoft recognized that it could cripple the browser
threat by eliminating Netscape as a browser supplier for
Windows 95
65. With Netscape's dramatic success in early 1995, Microsoft realized that
Internet browsers, in particular Netscape, posed a serious threat to the applications
barrier to entry that protects Microsoft's monopoly power. Microsoft thus also
recognized that inducing Netscape to abandon its efforts to develop platform-level
browsing software for Windows 95 would eliminate the browser threat to its operating
system monopoly.
65.1. By June of 1995, Microsoft recognized that Netscape was a
significant competitor in Internet browsers and therefore posed a threat to Microsoft's
operating system monopoly.
- See supra Part III.B.2; ¶ 56.1.
65.2. In the first half of 1995, Microsoft and other industry participants
expected that Windows 95 would quickly become the dominant PC operating system
when it was released.
- Barksdale Dir. ¶ 110 (In the June 21, 1995, meeting between
representatives of Netscape and Microsoft, "We all anticipated" that
Windows 95 "would shortly be the dominant operating system.")
65.3. Microsoft understood, therefore, that it could eliminate the platform
threat posed by Netscape's browser by convincing Netscape not to make its browser
available as an alternative platform for Internet software development on Windows 95.
Indeed, convincing Netscape not to offer a competitive, platform-level browser on
Windows 95 was Microsoft's primary goal for its negotiations with Netscape.
- On April 14, 1995, Daniel Rosen, the person at Microsoft most
responsible for negotiating a deal with Netscape (Rosen Dir. ¶7),
wrote to Paul Maritz and other Microsoft executives about "the
potential structure of a [Microsoft] relationship with Netscape."
Rosen wrote that the "acceleration of Internet related activities
makes it critical that we begin to make these choices ASAP, or they
will be made for us . . . . With Netscape, we face a small,
aggressive, FOCUSED competitor or ally. Their focus is both good
news and bad. They can be successful, simultaneous with our
success, if we can decide what we want . . . . Will they adopt our
tools? Will they cede the client and its standards to us?" GX 18
(emphasis added).
- This goal was reiterated on May 15, 1995, when Rosen wrote to
Maritz and others that "our goal should be to wrest leadership of
the client evolution from [Netscape]." GX 331, at MS98 0103672.
On May 25, Thomas Reardon was even more direct, writing in an
email to Rosen and others that any agreement to exchange
"protocol specs" with Netscape "would really be a veiled effort on
our part to move them off of the Windows client." GX 952
(emphasis added).
- One week later, Maritz wrote to Bill Gates and Rosen that he
viewed Microsoft's "imperative to be first and foremost to ensure
that we keep control of the standard Internet client api's and
protocols," that in order to do so Microsoft would have to "coopt
Netscape," and that he was open to any strategy that would
achieve that result. GX 953.
- Microsoft's intention to eliminate competition from Netscape was
expressed again that same day by Thomas Reardon, who sent a
list of "working goals" for Microsoft's relationship with Netscape to
Maritz and other Microsoft executives. Those goals included
"move Netscape out of the Win32 Internet client area" and "avoid
cold or hot war with Netscape. Keep them from sabotaging our
platform evolution." Reardon also wrote that Dan Rosen "points
out that we must offer them [Netscape] some story as to how they
can slowly shift away from the core client business, or at least the
core Win32 client business" and that "Dan feels there is reasonable
hope for engaging Netscape in long term strategic cooperation,
where Netscape might run with the Mac and Win16 clients." GX
24.
- There is no meaningful difference, in these discussions, between
the terms "Internet client" and "browser." Rosen, 2/22/99am, at
30:10-17 ("The client . . . was focused on all of those things one
needed to connect a PC to the Internet to do useful things."); GX
20, at 5 (lengthy May 26, 1995 Gates memo refers to "O'Hare,"
which became Internet Explorer 1.0, in the same paragraph as both
"our Internet client" and "our browser"); GX 22 (May 31, 1995, e-mail to Maritz and others states that "Netscape is very influential on
what happens with Clients"); Rosen 2/22/99pm, at 8:23-25 ("I think
it would be fair to say" that Microsoft's Internet client "would include
certain components that were then shipping with the browser, but
not necessarily all of them."); Rosen 2/23/99am, at 6:20-7:3
(defining the Windows "client code" as "TCP/IP, Internet shortcuts,
HTTP and HTML rendering, among others").
b. Microsoft first suggested that Netscape not compete
with it in the Windows 95 browser business at a June 2,
1995, meeting
66. Microsoft began discussing Netscape's Windows 95 browser plans in a June
2, 1995, meeting with James Barksdale. At that meeting, Microsoft suggested that
Netscape consider abandoning its independent base of platform-level browsing code on
Windows 95 and merely build on top of Microsoft's code.
66.1. James Barksdale visited Microsoft and met with Daniel Rosen,
Nathan Myrhvold, and Paul Maritz. Both sides discussed potential points of technical
and other cooperation.
- The meeting "explored possible areas of collaboration between the
two companies." Barksdale Dir. ¶ 101. "Microsoft principally was
interested in getting" Netscape "to consider adopting certain
security protocols, incorporating into Navigator certain viewers that
would enhance Microsoft content, and other technologies . . . ."
Barksdale Dir. ¶ 101; see also GX 25 (Rosen summary: Microsoft
wants Netscape "Support of STT"); GX 26 (Microsoft offered to
provide Netscape "with early disclosure of any 'standards and
protocols' that would allow" Netscape to improve its browser so that
it could work with MSN content, and Microsoft suggested that
Netscape "could be the preferred or perhaps exclusive developer of
these sorts of products."). There were also discussions of
Microsoft possibly bundling Netscape server software. GX 25, at
MS98 0009972; see also GX 26.
- The meeting was "very friendly, non threatening," GX 26; see also
Barksdale Dir. ¶ 101; GX 25 (Rosen's notes: "a cordial, open
discussion of issues and direction."); Rosen Dir. ¶ 64.
66.2. Microsoft's goal at the time of the meeting was to "move Netscape
out of the Win32 Internet client area." Rather than directly confront James Barksdale
with that plan on June 2, however, Microsoft initially suggested that Netscape consider
not competing and, instead, "using Microsoft's underlying code for its browser, adding
their value-added components on top."
- Rosen's notes of the meeting record, under the heading "What
Microsoft wants from Netscape," that Microsoft asked Barksdale for
"Strong support of Win95 and its evolution path. (Stronger support
than for other products like Acrobat or Java). This might include
using Microsoft's underlying code for its browser, adding their
value-added components on top." GX 25, at MS98 0009972.
- Thomas Reardon conceded that Netscape's "core client" already included
an HTML renderer but noted: "There's lots of value-added stuff beyond
HTML rendering in the Navigator product." Reardon Dep., 9/9/98, at
347:20-25 (DX 2606). A value-added application was "roughly
anything that takes advantage of an API. It's an application. That's
it. . . . But I should qualify it. I was making a distinction between
core client technology, HTML, HTTP, transport technology,
rendering technology, and programs that run above that or
alongside of that independent of HTML, the HTML rendering itself."
Reardon Dep., 9/9/98, at 343:23 - 345:10 (DX 2606).
66.3. James Barksdale responded by stressing the importance of
Netscape's upcoming Windows 95 browser to Netscape's plans. He did not fully
appreciate until the later June 21 meeting the import of Microsoft's statements that
Netscape should simply use Microsoft's underlying code and add value-added
components on top of it .
- Barksdale "wanted to stress the importance of the client to
Netscape's business strategy." Barksdale Dir. ¶ 101. He informed
the Microsoft participants that Netscape planned to release a retail
version of Navigator on Windows 95 shortly and that Windows 95
and Windows 3.1 were Netscape's primary browser distribution
platforms. GX 25, at MS98 0009972 (Rosen's notes). When the
parties discussed "browser cooperation," Barksdale responded that
"he would like to explore ways to cooperate, but he wants to
continue to add value in browser code." In fact, Barksdale even
suggested at one point that Microsoft "distribute his browser." GX
25, at MS98 009972.
- Barksdale told the Microsoft attendees that Netscape was planning
to charge for its retail browser product to all but student and non-profit customers, that they were "selling a lot of site licenses for
browsers," and that they were pleased with the revenues they were
receiving from licensing both browser and server software to
enterprises. GX 25, at MS98 at 0009972.
66.4. Microsoft's internal discussions following the June 2 meeting further
demonstrate that Microsoft's principal goal for any "strategic relationship" was
convincing Netscape to abandon the Windows 95 Internet client market.
- On June 5, 1995, Daniel Rosen circulated his notes of the
June 2 Netscape meeting to Bill Gates, Paul Maritz, and
other Microsoft executives, giving an optimistic assessment
of the potential for a "broad strategic relationship" between
Microsoft and Netscape. GX 25, at MS98 0009973. Maritz
responded that he did not believe "that Netscape is 'ready
for a broad strategic relationship.' It was clear that he/they
view the client as a key place to make money, since that
'hook is so important for selling additional software.'
Barksdale was primarily interested for us to distribute his
client and his server . . . ." GX 27.
c. At a June 21, 1995, meeting Microsoft expressly
proposed a naked market-division agreement to stop
Netscape from offering a competing platform
67. At a subsequent meeting held on June 21, 1995, under the pretext of
exploring forms of complementary collaboration, Microsoft proposed to Netscape that
the two firms divide the browser market, with Netscape to cease developing its
independent browser for the Windows 95 market.
67.1. On June 21, 1995, Microsoft representatives Daniel Rosen, Chris
Jones, Thomas Reardon, Richard Wolf, Anthony Bay, James Allard, and Barb Fox
came to Netscape for a follow-up meeting to the June 2, 1995, meeting.
- Rosen Dir. ¶¶ 70-75 (list of Microsoft attendees, who they
were, and what Rosen expected them to discuss at the
meeting); GX 32 (formal agenda for the meeting).
67.2. On the morning of June 21, 1995, prior to meeting with Netscape,
the Microsoft representatives had a pre-meeting at the St. Claire Hotel in San Jose to
discuss their objectives for the meeting. The consensus from that discussion was that
Microsoft's primary goals for the meeting were to determine whether Netscape intended
to compete with Microsoft at the platform level and, if so, to convince Netscape not to
do so.
- Rosen testified that he "had asked the group of Microsoft
representatives to meet in a conference room . . . rented in the St.
Claire Hotel in San Jose." Rosen Dir. ¶ 79. Jones attended the
meeting as the representative of the platform group at Microsoft
(Rosen, 2/22/99am, at 55:5-6), and "gave the best summary of the
purpose and goals of the meeting," during the pre-meeting
discussion, "so we nominated him to give our goals statement."
Rosen Dir. ¶ 82.
- Jones, when asked whether he recalled at the premeeting "any
discussion about a desire of anybody on the part of Microsoft who
was participating to be able to persuade or influence Netscape not
to compete with Microsoft," testified: "Absolutely." Jones Dep.,
1/13/99, at 582:10-18. See also Jones Dep., 10/27/98am, at 39:25
- 40:17.
- Discussion at the premeeting focused on the "line between
solutions and platform" and on how to persuade Netscape to stay
on the solutions side of that line on Windows 95. Jones Dep.,
1/13/99, at 581:4-19. When asked whether there was "any
discussion about trying to influence Netscape in any way to either
move toward or stay on one side of the line . . . as opposed to
simply finding out where it was that they intended to do business,"
Jones answered: "It was both." Jones Dep., 10/27/98am, at 39:24
- 40:11.
67.3. Following their premeeting, the Microsoft delegation met with Janes
Barksdale, Marc Andreessen, Mike Homer, and Ram Shriram for approximately four
hours. The discussion covered a wide range of technical issues from the pre-arranged
agenda.
67.4. In addition, the Microsoft team, led by its designated spokesman
Chris Jones, told the Netscape representatives that Netscape should not develop a
browser for Windows 95 because Microsoft intended to develop its own Windows 95
browser. Microsoft proposed that the two companies agree not to compete with each
other by drawing a "line" between what Microsoft called "the platform" and what it called
"solutions," with Netscape agreeing not to compete on the platform side of the line. In
exchange for Netscape's agreement to abandon a Windows 95 browser and to stay on
the solutions side of the line, Microsoft offered not to compete with Netscape in
browsers that ran on operating systems other than Windows 95 and in non-platform
solutions that ran on top of Windows 95 and Internet Explorer.
- James Barksdale testified: "I experienced something I had not ever
seen happen in my more than thirty years of experience with major
U.S. corporations . . . . Microsoft apparently came to Netscape
with a single focus: to convince Netscape not to compete with its
Windows 95 browser product, Internet Explorer. Microsoft
proposed a division of the browser market between our companies:
if Netscape would agree not to produce a Windows 95 browser that
would compete with Internet Explorer, Microsoft would allow
Netscape to continue to produce cross-platform versions of its
browser for the relatively small market of non-Windows 95
platforms: namely, Windows 3.1, Macintosh, and UNIX. Moreover,
Microsoft made clear that if Netscape did not agree to its plan to
divide the browser market, Microsoft would crush Netscape, using
its operating system monopoly, by freely incorporating all the
functionality of Netscape's products into Windows." Barksdale Dir.
¶ 25.
- "Microsoft's officials made clear that they believed that Netscape
should work with them on areas other than a browser for Window
95, but that we should not develop our own browser for Windows
95 because they intended to build a Microsoft browser for the
Windows 95 operating system. They proposed that a 'line' be
drawn between the area in which we developed products and
competed and the area in which they developed products.
Microsoft proposed that we build products that would run on top of
the Windows 95 operating system and browser. They offered to
allow us to continue to develop browsers for other operating
systems, as long as we did not try to compete with them in
developing a browser for the Windows 95 platform." Barksdale Dir.
¶ 110.
- Netscape's Marc Andreessen "who is an extraordinarily fast typist,"
took detailed notes on his laptop computer as the meeting was
going on. Barksdale Dir. ¶¶ 108, 112. Those simultaneous notes
record that Microsoft asked Netscape "would you be interested in
having a partnership where NS gets all the non-Win95 stuff and MS
gets all the Win95 stuff? If NS doesn't want to, then that's one
thing. If NS does want to, then we can have our special
relationship. THREAT THAT MS WILL OWN THE WIN95 CLIENT
MARKET AND THAT NETSCAPE SHOULD STAY AWAY." GX
33, at NSC 017100 (emphasis in original).
- Microsoft's Chris Jones, when asked whether there was "any
discussion in the meeting with the Netscape people that
essentially, under some form of a deal or partnership, Microsoft
could take, essentially would take the part of the business that
related to Windows 95 and Netscape could handle remaining parts
of the business, for example, the cross platform clients" answered:
"Oh, I believe there was a discussion of that nature, yes." Jones
Dep. (played 10/27/98am), at 40:19 - 41:3.
- Microsoft left no doubt of its intent to convince Netscape to agree
to stop competition. Andreessen testified: "I got the feeling that I
was being visited by a force that was extremely powerful in the
space where I chose to operate, and that I was basically being
given the terms and conditions by which an arrangement would be
struck, where, among other things, the market would be segmented
and our company would be allowed to succeed in certain limited
areas, and Microsoft would be allowed to succeed in areas that it
chose to define." Andreessen Dep., 7/15/98, at 429:10-20 (DX
2555). He wrote in a June 21 email, that: "Much of the
conversation centered on a discussion of how the line would be
drawn between the platform and their value added." GX 535.
- Chris Jones opened the meeting by commenting that Microsoft
"believes" that there is a "set of things that are provided in Internet
servers and browsers that will be in the core operating systems or
given away with the OSs, as a facility like the Win32 API. What MS
needs is someone -- a partner -- who is going to take those core
services to build on top of them and create solutions for customers.
MS has some ideas about what these features are, where the line
is, and who the partners should be. All of the relationship points
revolve around critical fact of -- is Netscape the kind of company
that's going to partner with MS on this or not? Will MS & NS be
able to cooperate & agree on the line, where it's drawn, etc. If not,
companies will compete. If so, then arrangement can be highly
beneficial, with 'aligned interests'." GX 33, at NSC 017098
(Andreessen's notes) (emphasis added).
67.5. In addition to offering not to compete with Netscape for non-Windows 95 browsers and for "solutions," Microsoft also offered to make Netscape a
"preferred" ISV and to give Netscape preferential access to technical information about
how to make its products work better with Microsoft's operating systems.
- "Microsoft officials said that, if we agreed to the 'special
relationship' they proposed, Microsoft would support us by making
Netscape a 'preferred' ISV. The Microsoft personnel made clear
that issues concerning the RNA API and related technical
information we had been seeking could be resolved '[d]epending
on how we walk out of this room today.' If we agreed to the
'special relationship they proposed, the Microsoft representatives
said that we would be the first ISV to receive the technical
information . . . ." Barksdale Dir. ¶ 110. Concerning Barksdale's
repeated requests for the RAS "dialer" APIs that Netscape had
been requesting, Rosen responded: "We can fix that problem. In a
perfect world, anyone can plug into that. With a special
relationship with you, you'll be the first to plug into it." GX 33, at
NSC 017101. Microsoft effectively stated that, if Netscape agreed
not to compete, "then we [Netscape] can have our special
relationship." Id.
- Microsoft offered to "consider licensing us the ability to turn
Navigator into a container," but only "IN THE CONTEXT OF THE
LARGER DEAL," and stated that the necessary APIs "have been
abstracted out into a layer above the normal Office Compatible
program and people who want to use them are picked specifically
by MS and are arbitrary -- MS reserves right to say who's in and
who's out. Again, part of the broader discussion . . . There are
things we can do if we're working together that we can't do
otherwise." GX 33, at NSC 017099 (Andreessen's notes).
Microsoft also offered to help Netscape do "tight integration" with
MSN, but only on the condition that the two companies have a
"tight relationship. . . . . If we didn't have a tighter relationship, you'd
be back to what a normal ISV can do." GX 33, at NSC 017100.
Microsoft also suggested that MSN content "could in fact be hosted
on a Netscape server," but "they refuse to talk about it until we
have the broader relationship established." GX 33, at NSC
017100-1.
67.6. Microsoft made clear that all of this preferential treatment and its
own willingness not to compete on non-Windows 95 browsers and "solutions" on top of
Internet Explorer depended on Netscape's agreement not to compete and to divide
markets and that, if Netscape declined the offer, "Microsoft would crush them."
- To Barksdale and Netscape, "the main goal for this meeting was to
get access to certain code and APIs necessary for . . . product
development . . . ." Barksdale asked "whether obtaining those
things was tied to . . . acceptance of this 'special relationship'
Microsoft had proposed . . . ." Microsoft answered that Netscape's
"obtaining the necessary technical information 'certainly isn't
independent'" of Netscape's acceptance. "Microsoft made it very
clear that they would attempt to crush" Netscape "by attempting to
own the client." Barksdale Dir. ¶¶ 110-112. See also GX 34 (June
22, 1995 AOL email detailing Andreessen's report, one day after
the meeting, of Microsoft's threat that "if Netscape didn't do the
deal, Microsoft would crush them.").
- Microsoft indicated "implicitly" that it "would bring full force of all of
its energy to crush us in the market . . . [p]resumably using many of
the same tactics they have used over the last three years, to seek
to make sure that we didn't have the opportunity to succeed as an
independent company." The "tactics" Microsoft would employ
included exclusionary contracts with ISVs and OEMs and "release
of full-on competitive or even cloned products that would be priced
at a level that would make it impossible for [Netscape] to compete,
which subsequently happened." Andreessen Dep., 7/15/98, at
429:10 - 431:8 (DX 2555).
67.7. The testimony of James Barksdale and Marc Andreessen about
Microsoft's June 21 market-division proposal is credible and consistent with the
available contemporaneous documents from Netscape, Microsoft and third parties like
America Online .
- See supra ¶ 67.2, 67.4, 67.8-.9.
67.8. Microsoft's internal discussions following the June 21, 1995,
meeting leave no doubt that Microsoft proposed an agreement not to compete and to
divide markets.
- On the evening of the June 21 meeting, Dan Rosen
prepared a summary and sent it to the other Microsoft
attendees for their comments. The summary, though "naive"
(GX 537), is substantially consistent with Marc Andreessen's
notes and James Barksdale's memory of the meeting and
makes clear that Microsoft's goal was to prevent platform-level competition. GX 535.
- Rosen wrote in part that "ChrisJo summed up the
purpose nicely: 'We need to understand if you will
adopt our platform and build on top of it or if you are
going to compete with us on the platform level.'" GX
535.
- Rosen also wrote that: "Much of the conversation
centered on a discussion of how the lines would be
drawn between the platform and their value added.
On the client end, we discussed 'sucking most of the
functionality of the current Netscape browser (but not
the toolbar, cool places or advertising) into the
platform; they seemed OK with this concept." GX
535.
- Thomas Reardon responded to Rosen's draft the next
morning with a more skeptical assessment, though one
which unequivocally confirms that moving Netscape out of
the Windows 95 browser market was Microsoft's central goal
for the meeting. He wrote that Netscape was "trying to
preempt ohare with a win95 product of their own. i do not
think they are so easily displaced from win95 client arena,
they continue to move down the path of selling/giving away
a 'premium' browser for win95." GX 535 (emphasis added).
- Richard Wolf's comments on Rosen's draft agreed with
Rosen that Netscape would not be "competing with us in
defining a platform . . . ." but expressed skepticism that
Netscape would be willing to "drop a broad based client in
favor of vertical markets." GX 501.
- Chris Jones began his response to Rosen's draft with a list
of "Microsoft's goals, in priority order." Number one on that
list was that Microsoft "Own client platform." Confirming
precisely Barksdale's and Andreessen's descriptions of the
meeting, Jones wrote: "The critical question is: Do they want
to align strategically with us or not? Are they willing to bet
that we'll be successful, and will they make the commitment
and changes necessary in their strategy to do this?
Because of our priority to own client and server platform, if
they can agree to use our client code on Win 95, and use
our BackOffice and NT APIs, and promote these as the
solutions, then they will have aligned with our businesses
and we have a deal." GX 557 (emphasis added).
- After gathering those comments from the other Microsoft
attendees, Rosen sent a revised draft to Bill Gates, Nathan
Myhrvold, Paul Maritz, and 17 others on the afternoon of
June 22, 1995. That draft incorporated Chris Jones's
assessment of Microsoft's objectives: "Our goals going into
the meeting were (in priority order): 1. Establish Microsoft
ownership of the Internet client platform for Win95." He also
echoed Jones's view that "the critical question" is whether
Netscape would be willing to "align strategically with us . . .
The test of this alignment will be Netscape's agreement to
use Microsoft's client code on Win 95." Rosen, ignoring
Reardon's more skeptical assessment, ended instead with
the optimistic conclusion that Netscape "seemed to embrace
this strategy" in the meeting. GX 536, at MS98 009585,
MS98 0009587 (emphasis added).
- Reardon disagreed and responded with an e-mail entitled:
"Netscape meeting: reality." Leading off a list of eight points
about the meeting that he felt Rosen had missed, Reardon
wrote: "1) Netscape is preempting O'Hare. We sent them a
list of about thirty talking points. They sent back ten, nearly
all revolving around shipping their Win95 browser." Reardon
also described Rosen's view that Netscape was "OK with
this concept" of Microsoft sucking browser functionality into
the platform as "bunk. . . . There was a noticeable increase
in the level of tension," Reardon wrote, "whenever this sort
of language came up. one clearly telling quote from
Barksdale: 'all we want is our god given 95% market share
for the browser'. he said this with a wink, but i don't know
what could be more clear." Reardon concluded that: "We
will compete on just about every technology. . . . maybe i am
being a dick, but there is no deal here." GX 536, at MS98
0009584 - 0009585.
- Top Microsoft management agreed that Rosen was
inaccurate. Brad Silverberg forwarded Reardon's "reality"
email to Bill Gates, who responded that "I think Thomas is
reading the situation pretty well. I think Dan is great but I
agree he is being a little naive in this case." GX 537.
- In a July 12, 1995, phone call with Dan Rosen, James
Barksdale discussed a possible server joint venture for
Windows . After a series of internal messages, Paul Maritz
summarized recent events regarding Netscape: "My
thoughts: we originally hoped that there was some way to
leverage a relationship with Netscape based on a business
model whereby they would be prepared to cede the client to
us or at least give us some major advantage, if we could
give them some major advantage in the server area. They
are not prepared to give us a significant client advantage
(either for O'Hare or for MSN), so we should treat them as
an ISV, but not much more." When Rosen ventured that
Netscape "may be willing to give us a client advantage,"
Maritz wrote that with respect to "them building on top of our
browser, I am very skeptical that they would agree to this.
Their history is to [sic] closely bound up with the browser -
this is after all Andreessen's company." Rosen replied that,
"Given your message, I understand that you don't see the
need to continue discussions with Netscape at a 'strategic'
level." Maritz responded with one word, "Right." GX 540
(emphasis added).
67.9. The account of the June 21 meeting contained in Microsoft and
Netscape records is also corroborated by a contemporaneous memorandum from the
files of America Online.
- On June 22, 1995, one day after the meeting took place,
AOL's David Kaiser reported being told by Marc Andreessen
that "Microsoft was at Netscape yesterday. . . . They wanted:
- equity
- a board seat
- Netscape to renounce the network as a platform
- Netscape to disclose all plans to microsoft
- Netscape to limit access to APIs
And in return, Netscape would be Microsoft's special
partner, get inside information, etc . . . and if
Netscape didn't do the deal, Microsoft would crush
them." GX 34.
3. Microsoft's after-the-fact assertion that its market division
proposal was simply exploring forms of legitimate cooperation
is pretextual and contrary to the evidence
a. Microsoft's contention that it was not trying to get
Netscape out of the browser business is erroneous and
rests on a misleading play on words
68. Microsoft's witnesses did not genuinely dispute the basic substance of
Barksdale's and Andreessen's testimony about the June 21 meeting; instead, they
attempted to cast the proposal not to compete in a less sinister light through a
manipulation of the word "browser." Although Microsoft's negotiators made clear to
Barksdale and Andreessen that Microsoft wanted Netscape out of the Win32 browser
business, Microsoft now contends that it simply wanted Netscape to build what
Microsoft calls a "browser" on top of "Microsoft technologies." This contention is
inconsistent with the contemporaneous documents and is, in any event, of no
consequence. Even on Microsoft's version of the facts, its proposal to Netscape was a
naked attempt to eliminate platform-level competition and to divide markets.
68.1. Mr. Rosen testified that Microsoft's objective for the June 21, 1995,
meeting was to convince Netscape to build software products on top of the "Microsoft
technologies" or "browsing software" that Microsoft was building into Windows.
- Daniel Rosen admitted that: "We wanted to persuade Netscape's
engineers to develop browser and other software offerings on top
of Windows 95" (Rosen Dir. ¶ 127) and that Microsoft "did make
clear that there would be browsing software in Windows 95 and
that we expected users to take advantage of that software." Rosen
Dir. ¶ 128; see also Rosen Dir. ¶ 50; Rosen, 2/22/99am, at 10:2-10.
- Thomas Reardon testified: "I knew that we were going to be
providing Internet APIs for HTML rendering, for HTTP, et cetera. I
wanted Netscape to use those APIs. That simple." Reardon Dep.,
9/9/98, at 343:23 - 349:9 (DX 2606) (sealed).
68.1.1. But Mr. Rosen's testimony on this point was inconsistent.
68.1.1.1. At times, Mr. Rosen testified that Microsoft wanted
Netscape to continue to build a Windows 95 "browser" as part of a joint venture with
Microsoft.
- Rosen Dir. ¶ 127 ("We wanted to persuade
Netscape's engineers to develop browser and other
software offerings on top of Windows 95."); Rosen,
Dir. ¶ 50 (Microsoft did not want Netscape to agree to
stop developing browser software for use on
Windows); Rosen, 2/22/99am, at 8:12-24 (allegedly
Microsoft "spent considerable effort during the spring
of 1995 trying to convince Netscape to build and
market" a browser to run on Windows 95.); see also
Rosen, 2/22/99am, at 11:13-18 (same).
68.1.1.2. At other times, Mr. Rosen testified that, because
Microsoft purportedly would be building "browsing software" into Windows 95 (Rosen
Dir. ¶ 128), the proposal was instead that Netscape should concentrate on "higher-level
client-side applications (like groupware and multimedia extensions)" (Rosen Dir. ¶ 59).
- Rosen testified that when James Barksdale said that
he "wanted to continue to add value in browser code"
at the June 2, 1995 meeting, Rosen understood that
comment "to be consistent with the sort of higher-level client-side applications (like groupware and
multimedia extensions) that I described above."
Rosen Dir. ¶ 59.
- Rosen testified that Microsoft encouraged Netscape
to develop "'cool looking and sounding app[lication]s'
that took full advantage of Internet-related features in
Windows 95. Examples of such applications included
so-called 'groupware' packages, multimedia
extension packages and the like." Rosen Dir. ¶ 50
(citing DX 734).
68.1.1.3. In the end, Mr. Rosen presented no clear
explanation of what kind of software he envisioned Netscape creating as part of any
deal and simply repeated the vague words "cool looking and sounding apps."
- Rosen testified that Microsoft's "principal interest --
making Windows more desirable -- was best served
by encouraging Netscape to stay in the client
software business and to develop what Mr. Reardon
had called 'cool looking and sounding apps' that took
full advantage of Internet-related features in Windows
95." Rosen Dir. ¶ 50 (citing DX 734).
- Rosen testified that "it was strongly in Microsoft's self-interest to encourage Netscape to continue to 'add
value' on top of Windows, that is, to accept the
invitation publicly extended by Thomas Reardon in
Germany in April 1995 to become one of the favored
'Web ISVs' for Windows 95. Having Netscape
continue to develop what Mr. Reardon called 'cool
looking and sounding apps' for Windows 95 would,
we believed, boost demand for Windows 95." Rosen
Dir. ¶ 124.
68.1.1.4. On cross-examination, Mr. Rosen seemed to
settle on a story that Microsoft wanted Netscape to build a "browser" of some kind on
top of the browser code that Microsoft now says it was building into Windows 95.
- Rosen, 2/22/99am, at 10:2-10 (there was "a lot of
discussion about whether they would adopt the
Microsoft technologies that we were building into
Windows 95 to build such a product or whether they
would try to do more things on their own").
68.1.2. The contemporaneous documents do not reflect any such
understanding within Microsoft at that time and make clear that none was
communicated to Netscape. Microsoft never said, and Netscape never understood,
what kind of software, if any, Microsoft wanted it to build on the Windows 95 platform.
- Barksdale testified, in reference to Microsoft's proposal that
Netscape build software in the form of vertical applications,
"that was never clear to me what that was." Barkdale,
10/21/98am, at 42:16-18.
- Barksdale testified that "it would be a little difficult to figure
out what that might be unless you go up to the server . . ."
and that "it got very, very confusing to me because the line
seemed to automatically exclude what we thought was going
to be one of our very best products." Barksdale,
10/22/98pm, at 37:13 - 38:2.
68.2. Mr. Rosen's apparent contention that Netscape would have been
free to build a "browser" on top of the "Microsoft technologies" in Windows 95 is just a
play on words.
68.2.1. There was no meaningful distinction between the
"technologies" Mr. Rosen now claims Microsoft intended to reserve for itself and
Microsoft's browser, Internet Explorer. The contemporaneous documents make it clear
that the "platform" technologies that Microsoft said it would build into Windows would
duplicate the entire functionality of Netscape's browser product, with the possible
exception of "the toolbar, cool places or advertising."
- Rosen's notes stated that: "Much of the conversation
centered on a dicussion of how the lines would be drawn
between the platform and their value added. On the client
end, we discussed 'sucking most of the functionality of the
current Netscape browser (but not the toolbar, cool places or
advertising) into the platform; they seemed OK with this
concept." GX 535.
- Andreessen's notes reflect that Microsoft asked whether
Netscape would be "interested in having a partnership
where NS gets all the non-Win95 stuff and MS gets all the
Win95 stuff? . . . THREAT THAT MS WILL OWN THE
WIN95 CLIENT MARKET AND THAT NETSCAPE SHOULD
STAY AWAY." GX 33 (emphasis in original).
68.2.2. In any event, building on top of those "Microsoft
technologies" would have required Netscape to abandon its independent platform-level
browser and associated APIs and, along with them, any hope of competing with
Microsoft at the platform level.
68.2.2.1. Several companies have recently written
specialized browsing programs on top of the so-called Internet "technologies" that
Microsoft distributes with Windows 95 and Windows 98; they are often called "shell
browsers," and consist of a small amount of user interface code that relies on Internet
Explorer to do the actual work of connecting to the Internet and displaying retrieved
information.
- See infra V.C.1.c; ¶ 187.2
68.2.2.2. Because the underlying code relied on by a shell browser
is Internet Explorer's, such a browser does not offer platform competition to Internet
Explorer.
- Rosen conceded that if Netscape adopted all of Microsoft's
underlying technologies for its browser, they would no longer
pose a platform threat. Rosen, 2/22/99pm, at 31:9-12;
Rosen, 2/23/99am, at 54:3-6. (similar).
- Maritz conceded that if Netscape adopted Microsoft's
technologies for its browser, that would "mitigate" the
platform threat. Maritz, 1/26/99pm, at 49:3-10; see also
Maritz, 1/26/99am, at 29:19-22 (Netscape's APIs made it a
threat); Maritz, 1/26/99am, at 30:4-6 (same); Maritz,
1/26/99pm, at 53:9-12 (Microsoft's goal was to keep control
of the browser APIs).
- When asked whether "a browser like Encompass that uses
the technologies of the Microsoft platform is not going to be
viewed as a serious competitive threat to Microsoft," Maritz,
answered: "Correct." Maritz, 1/25/99pm, at 30:5-8; see also
Maritz, 1/25/99pm, at 30:14-19.
- Professor Fisher testified that, for purposes of his
conclusions in this case, there was "absolutely no
difference" between shell browsers and Internet Explorer.
Fisher, 1/5/99pm, at 20:4-23.
b. Rosen's other testimony, both regarding the June 21,
1995, meeting and more generally, is evasive and
misleading
69. To the extent that Mr. Rosen's testimony about the June 21 meeting
diverged from James Barksdale's in more than a semantic way, Mr. Rosen's testimony
was incredible. Mr. Rosen's demeanor on the stand was evasive and unhelpful, and his
testimony was riddled with inconsistent and implausible assertions.
69.1. First, Mr. Rosen testified that he never viewed Netscape as a
competitive threat in early 1995 and was not aware of a single person within Microsoft
who did (Rosen, 2/22/99am, at 48:13). That assertion is flatly inconsistent with the
contemporaneous documents, the testimony of Microsoft's other witnesses, and Mr.
Rosen's role in the Netscape discussions.
- See supra Part III.B.2; ¶ 56.1.
- Rosen wrote a May 1995 memo that the "threat of another
company (Netscape has been mentioned by many) to use their
Internet WWW browser as an evolution base could threaten a
considerable portion of Microsoft's future revenue." GX 331.
- When confronted with that memo on cross examination Rosen
claimed first that he did not believe what he had written even at the
time he wrote it and that he had never sent the memo to the people
to whom it is addressed (Rosen, 2/22/99am, at 23:24 - 27:18).
Rosen's testimony was demonstrably false, and he was quickly
forced to admit that his testimony was untrue, conceding that he
must have sent the memo to Ben Slivka. Rosen, 2/22/99pm, at
4:22 - 5:9.
69.2. Second, Mr. Rosen testified that Netscape always considered its
leadership position in Internet client software to be a financial dead-end and a waste of
engineering resources and was therefore eager to cede to Microsoft responsibility for
client technologies on Windows 95 (Rosen Dir. ¶44; Rosen, 2/22/99am, at 41:3-4). But
the record shows that everyone at Microsoft, including Rosen, understood that client
innovation was an important part of Netscape's business strategy.
- Rosen's notes from the June 2, 1995 meeting with James
Barksdale report that Netscape was "pleased so far with their ability
to sell servers and browsers to enterprises . . ." that they "are
selling a lot of site licenses for browsers . . . . Windows 95 and Win
3.1 are their primary browser development platforms." Rosen also
reported that "Netscape feels that they must be free to support any
protocol, API, etc. that becomes popular, without restriction." And
on the subject of "browser cooperation," Barksdale said that "he
would like to explore ways to cooperate, but he wants to continue
to add value in browser code." GX 25.
- Maritz responded to Rosen's notes from the June 2 meeting with
the comment that "It was clear that he/they view the client as a key
place to make money, since that 'hook is so important for selling
additional software.' Barksdale was primarily interested for us to
distribute his client and his server." GX 27.
- Commenting on Rosen's draft of his notes from the June 21
meeting, Thomas Reardon suggested that he "add something
about barksdale's joking about 95% market share, i think it is
somewhat telling. . . . i do not think they are so easily displaced
from win95 client arena, they continue to move down the path of
selling/giving away a 'premium' browser for win95." GX 535.
- Rosen wrote on May 15, 1995, that Microsoft would have to "wrest
leadership of the client evolution from" Netscape. GX 331
(emphasis added).
69.3. Third, Mr. Rosen was deliberately evasive during cross-examination
about the meaning of the words "browser" and "client" in his testimony.
- Rosen testified that when, Richard Wolf used the word "client" in
DX 771, he meant "browser." Rosen, 2/23/99am at 30:7-21, 59:19-21.
- But when Rosen used the word "client" in GX 23, two hours later,
he meant something entirely different. Rosen, 2/23/99am, at 47:4-8.
- Rosen conceded that Internet Explorer was "Microsoft's Internet
client." Rosen, 2/22/99am, at 42:11-13. And when asked whether
"Netscape had leadership of the Internet client with respect to
various platforms . . ." Rosen answered: "Yes." Rosen, 2/22/99am,
at 40:1-3.
69.4. Fourth, Rosen claimed that Microsoft was not building a browser of
its own in mid-1995 (Rosen 2/22/99am, at 15:1-23), that he had never heard any
mention of a "browser battle" between Microsoft and Netscape (Rosen, 2/22/99am, at
17:1-18), and that he was not aware that gaining browser market share was a goal of
Microsoft (Rosen, 2/22/99am, at 18:2-5). Those statements are false. They are
contradicted by the rest of the trial record and the testimony of other Microsoft
witnesses.
69.4.1. Mr. Rosen testified that the June 21, 1995, meeting was
"primarily a technical 'brainstorming session' to seek areas of cooperation on
technologies, protocols and the like" (Rosen Dir. ¶ 68), rather than a "'relationship'
discussion" (Rosen Dir. ¶ 55) at which issues of broad business strategy and alignment
would have been discussed. That contention is belied by the contemporaneous
documents.
- Rosen's own notes from the earlier June 2, 1995, meeting
suggested that the "Next Steps" should be for each side to
"prepare a list of the things they want from and are willing to
give to form a relationship" and concluded that "Netscape is
open to a broad strategic relationship with Microsoft." GX
25.
- And when Rosen sent Microsoft's list to Netscape, it
included client terms, server terms, authoring terms,
technology terms, services terms, marketing terms, and
general terms including the possibility of a Microsoft
investment in Netscape and a seat on Netscape's board.
GX 556.
69.4.2. There is no contemporaneous evidence that anyone at
Microsoft, including Mr. Rosen, ever discussed or contemplated a proposal that
Microsoft would license Netscape's cross-platform browsers and distribute them under
Microsoft's brand name (Rosen Dir. ¶¶ 97, 99). In light of the documents, Mr. Rosen's
prior testimony, and the clear and credible testimony of James Barksdale, Mr. Rosen's
trial testimony on this point is neither credible nor persuasive.
- The internal documents suggest, at most, that Microsoft
discussed letting Netscape "run with the Mac and Win16
clients," (GX 24) on its own in return for Netscape's
agreement to vacate the Windows 95 market.
- By contrast, at trial Rosen testified that what Barksdale and
Andreessen remember as an offer to cede the non-Windows
95 browser business to Netscape was actually such a
proposal and attempted to portray that offer as merely a
"continuation" of Microsoft's December 1994 discussion with
Netscape about the possibility that Microsoft would license
Netscape's browser code for Windows 95. Rosen Dir. ¶ 97.
This characterization of events first emerged in Rosen's
testimony at trial.
69.4.3. Mr. Rosen's testimony that the discussion about the "line"
between Windows 95 and the browser at the June 21 meeting was initiated and
pressed by Netscape, not Microsoft (Rosen Dir. ¶¶ 85-88, 95), is also inconsistent with
the evidence.
- When asked at his deposition who were "the primary
speakers on this topic of the line," Rosen answered that "it
was primarily Thomas Reardon and potentially some
discussions with Chris Jones." Rosen Dep., 2/22/99pm, at
56:16-21.
- Chris Jones testified that the subject of "the line" was
extensively discussed at the Microsoft premeeting, where no
Netscape personnel were present. Jones Dep., 2/22/99pm,
at 40:14-22, 49:5-20.
69.4.4. Mr. Rosen's testimony that Microsoft wanted Netscape to
continue to make products that exposed APIs to independent software developers is
wholly inconsistent with the contemporaneous documents and the testimony of
Microsoft's other witnesses.
- See supra Part III.B; ¶ 56 (detailing Microsoft's recognition
that Netscape could develop APIs that could threaten the
applications barrier to entry).
c. Microsoft's contention that it engaged in legitimate joint
venture discussions with Netscape is contrary to the
evidence
70. Microsoft asserts more generally that the June 21, 1995, meeting was
merely an effort by Microsoft to find ways that Microsoft and Netscape could work
collaboratively together (Rosen Dir. ¶¶ 54, 77, 79). But the evidence shows that
Microsoft's offers to aid or collaborate with Netscape were merely a pretext designed to
secure Netscape's agreement to Microsoft's naked market division proposal and that
those offers were not genuinely intended to further any procompetitive collaboration
between the two companies.
70.1. First, contemporaneous Microsoft documents make plain that its
objective at the June 21, 1995, meeting was to eliminate the browser threat to the
applications barrier to entry and that any proposed "joint venture" was merely an artifice
to mask Microsoft's anticompetitive purpose.
70.1.1. Microsoft's internal correspondence both before and after
the June 21, 1995, meeting leaves no doubt that securing Netscape's agreement not to
compete with Microsoft in the Windows 95 browser market was Microsoft's primary
objective for that meeting. Those issues commanded the attention of Microsoft
executives at the very highest levels and were the subject of extensive discussion.
- See supra ¶¶ 66.2, 66.4, 67.2, 67.8.
- In a June 1, 1995 email to Bill Gates and other Microsoft
executives, Thomas Reardon wrote: "Dan points out that we
must offer them some story as to how they can slowly shift
away from the core client business, or at least the Win32
client business . . . . " GX 24.
70.1.2. Netscape understood the Microsoft proposal would put
Netscape out of business.
- Marc Andreessen testified at his deposition that "the
proposal . . . would result in crippling Netscape as a
potential competitor of Microsoft, and, indeed, as an
independent company." Andreessen Dep., (played
10/27/98am), at 25:11 - 18:14.
- Barksdale agreed with Andreessen's assessment that
Microsoft intended to divide the market and effectively
"cripple" Netscape. Barksdale, 10/27/98am, at 29:17.
- Barksdale testified: "Most of the matters on which they
would have had Netscape confine its work were not
commercially valuable." Barksdale Dir. ¶ 113.
70.1.3. Microsoft swiftly abandoned any plans for a collaboration
with Netscape when Netscape rejected the market division.
- Paul Maritz wrote that "we originally hoped that there was
some way to leverage a relationship with Netscape based
on a business model whereby they would be prepared to
cede the client to us or at least give us some major
advantage, if we could give them some major advantage in
the server area. They are not prepared to give us a
significant client advantage (either for O'Hare or for MSN),
so we should treat them as an ISV, but not much more." GX
540.
- Dan Rosen understood from Maritz's email that Maritz no
longer saw "the need to continue discussions with Netscape
at a 'strategic' level." GX 540.
70.2. Second, the one aspect of Microsoft's offer that was clear --
Microsoft's offer to refrain from competing with Netscape in the Macintosh, Unix, and
16-bit Windows segments of the browser market -- would not have aided any legitimate,
procompetitive collaboration between the two companies and was nothing more than a
naked bribe designed to secure Netscape's agreement to the market division scheme.
70.2.1. One of the inducements that Microsoft offered in exchange
for Netscape's agreement to abandon the Windows 95 browser business was a
promise not to compete with Netscape's browser on other platforms.
- See supra ¶ 67.4.
- Microsoft was thinking about a "long term strategic
cooperation, where Netscape might run with the Mac and
Win16 clients," at least as early as June 1, 1995. GX 24;
see also GX 33; Jones Dep., 10/28/98am, at 40:18 - 41:3.
70.2.2. The contemporaneous documents do not suggest that
Microsoft contemplated any technical or marketing efficiencies from such an
arrangement.
- See supra ¶¶ 67.2, 67.4, 67.8.
70.3. Third, that Microsoft's "joint venture" characterization is pretextual is
confirmed by the fact that eliminating Netscape as a browser supplier would have
reduced the value of, and thus demand for, Windows and was completely unnecessary
to achieve any legitimate purpose.
70.3.1. Microsoft makes money by selling copies of its operating
systems, and its legitimate business interests lie in maximizing consumer demand for
that software. Browsers are complements to operating systems; a good Windows 95
browser would therefore have increased demand for Windows, and thus Microsoft's
profits, regardless of whether that browser was produced by Microsoft.
- Fisher testified, "if browsers are complements to operating
systems . . . it should not matter who makes the
complement." Fisher Dir. ¶ 129(b).
- Warren-Boulton testified that "Microsoft has a legitimate
interest in ensuring that Windows users are able to acquire
high quality browsers at low prices, because that would
increase the demand for Microsoft's operating system."
Warren-Boulton Dir. ¶ 187.
- Rosen testified that "it was strongly in Microsoft's self-interest to encourage Netscape . . ." to build applications for
Windows 95 because those applications would "boost
demand for Windows 95." Rosen Dir. ¶ 124 (emphasis in
original).
70.3.2. That general principle was particularly true of Netscape's
web browser in late 1994 and 1995. Netscape's browser was the "killer app" of 1995
and generated tremendous consumer enthusiasm and demand for PCs.
- Barksdale specifically described Netscape's Navigator,
which would run on top of Windows 95, as "the killer app of
1995." Barksdale, 10/27/98pm, at 71:6-11.
- See supra Part III; ¶ 53.1.2..
- Barksdale testified that, at the time of the June 21, 1995
meeting, Netscape's browsing software "delivered
incomparably better performance than other browsers
available at the time," and that consumers preferred it by a
wide margin. Barksdale Dir. ¶ 231.
- Barksdale described Netscape's final version of Navigator
1.0 as a product which "delivered many times the
performance of other browsers available at the time," with
innumerable innovative features. Barksdale Dir. ¶ 54
- Barksdale testified that Netscape's browser was available
across many different platforms, including Mac, OS/2, and
various "flavors" of UNIX. Barksdale Dir. ¶ 80. Further, the
cross-platform availability of Netscape Navigator was of
great value to enterprises seeking to standardize their
software and training efforts, Barksdale Dir. ¶ 80; infra Part
V.B.1.b.(2); ¶ 107, and to independent software vendors
seeking to maximize the appeal of their programming efforts,
Barksdale ¶ 85.
70.3.3. By contrast, Microsoft's first version of Internet Explorer for
Windows 95 was still in development and was not released until July 1995. Cross-platform versions of Internet Explorer did not become available for the Macintosh and
UNIX operating systems until much later. And when Internet Explorer 1.0 for Windows
95 was eventually released, it was markedly inferior to Netscape's then-current
browsing software.
- Barksdale testified that the release of Navigator 1.0
"delivered incomparably better performance than other
browsers available at the time;" more specifically,
"Microsoft's 1.0 release of Internet Explorer was
substantially inferior to Navigator's 1.0 release." Barksdale
¶¶ 231-232.
- Reviews in the trade press consistently concluded that
Netscape Navigator was faster and more functional than
Internet Explorer. Schmalensee Dir. App. F Table F-1 (all
publications Dean Schmalensee listed selected Netscape
over Internet Explorer in 1995).
- Recognizing that performance gap, consumers preferred
Netscape Navigator to Internet Explorer by an overwhelming
margin in late 1995 (Myhrvold Dir. ¶ 26-27), even though a
license for Navigator cost $39 and Microsoft was giving
away Internet Explorer for free. Schmalensee Dir. ¶ 275;
Myhrvold Dir. ¶ 104. In fact, consumers still preferred
Navigator to Internet Explorer by almost four to one as late
as January 1997. GX 5.
- Barksdale testified that Netscape's browser share in January
1996 was approximately 84 percent. Barksdale,
10/21/98pm, at 33:1-4.
70.3.4. Moreover, by June 21, 1995, Netscape had already
invested millions of dollars to develop a new version of its browser for Windows 95;
these investments promised to further enhance the value of Windows and thus offered
consumers significant benefits.
- A May 11, 1995 email from Dan Rosen to Tom Johnston at
Microsoft, asking to "borrow/copy the Netscape Win95 new
client they gave us," makes it clear that Microsoft had a beta
version of Netscape's Windows 95 browser more than a
month before the meeting. GX 1892.
- Barksdale testified, "At the same time Microsoft was
releasing Internet Explorer, Netscape was well on its way to
its 2.0 release of Navigator." Barksdale ¶ 233.
70.3.5. The tremendous consumer enthusiasm for browsing
software and the Internet in late 1995, and the clear preference of consumers for
Netscape's browser at that time, demonstrate that consumers would have been harmed
if Netscape had acceded to Microsoft's demands and abandoned its Windows 95 web
browser.
- James Barksdale described Navigator as a "separate"
product which helps Microsoft sell Windows, adding, "a lot of
people buy computers because they want to run the
Netscape Navigator. It would help them to cooperate with
companies like ours." Barksdale, 10/27/98am, at 71:23 -
72:1.
- Barksdale testified, "that's why they wouldn't have asked us
to get out of it immediately. We were going to release
product in about a month. It was going to be great for
Windows 95. It was going to help them sell a lot more
because the Navigator is a hot application that ran on top of
Windows 95. It was the killer app of 1995." Barksdale,
10/27/98pm, at 71:6-11.
70.4. Fourth, the pretextual nature of any joint venture proposal is also
illustrated by the fact that Microsoft threatened to withhold technical support that
Netscape needed in order to "add value" on top of Microsoft's platform and retaliated
against Netscape by withholding that support when Netscape rejected the market
division proposal.
70.4.1. Microsoft retaliated against Netscape by withholding the
dialer API.
70.4.1.1. Microsoft knew that Netscape needed certain
critical technical information and assistance in order to complete its Windows 95
browser in time for the retail release of Windows 95; indeed, Netscape had repeatedly
asked Microsoft for some version of the dialer API and obtaining it was Netscape's
primary objective at the June 21, 1995, meeting.
- Beginning in March 1995,
Netscape's technical staff corresponded with Microsoft repeatedly in an effort to discuss
the technologies necessary for the development of a Netscape browser that would work
efficiently on Windows 95. Barksdale Dir. ¶ 95; Barksdale Dir ¶ 106; DX 728
- Reardon, in a June 23, 1995, email string to Maritz
and others entitled "Netscape meeting: reality,"
wrote: "We sent them a list of about thirty talking
points. They sent back ten, nearly all revolving
around shipping their Win95 browser." GX 536.
- On July 7, 1995, Mike Homer wrote to Daniel Rosen
that "I know that we have discussed this with you
before, but I wanted to reiterate how urgently we
need the following information: 1. RNA (remote
network access) phonebook API. Currently the only
way to create/edit phonebook entries is through the
WIN 95 interface. Our product will need to create an
RNA phonebook entry for the registration server (this
would happen during install) and also to configure the
customer's new account once the account has been
created. To make these calls, we need to have an
API that will allow him to configure the required
information. We first requested these in early June . .
. ." Barksdale added: "Dan - these are big deals,
please help us." GX 240.
- Barksdale testified, in reference to the meeting with
Microsoft, that: "Our top priority was to obtain APIs
and other technical information we needed from
Microsoft in order to release a browser compatible
with the Windows 95 operating system." Barksdale
Dir. ¶ 105.
70.4.1.2. At the June 21 meeting, Microsoft representatives
informed Netscape that Microsoft possessed a secret internal dialer API and that, if
Netscape was sufficiently cooperative on other matters, that technology could be made
available to Netscape immediately.
- Andreessen's notes reflect that when Barksdale
asked about the RAS "dialer" APIs that Netscape had
been requesting, Rosen responded: "We can fix that
problem. In a perfect world, anyone can plug into
that. With a special relationship with you, you'll be
the first to plug into it." Rosen also stated that: "We
need to give you code. Our alternative is to give you
stuff that wasn't developed for that purpose. There's
internal stuff that implements internal APIs, and those
APIs are known only within Microsoft." J. Allard then
added that: "'Depending on how we walk out of this
room today, we have a solution for your problem...' or
else in 3 months." Rosen concluded by saying that:
"If we had a special relationship, you wouldn't be in
this position." GX 33.
- Barksdale testified that Microsoft's message was, in
essence, "'depending on how we get along here
today, you can have the thing right now and we'll all
be good friends, and if it doesn't . . . go our way, then
you may have to wait a little while for it.'" Barksdale,
10/27/98am, at 53:3-8; see also Barksdale,
10/27/98am, at 54:20-23.
70.4.1.3. Microsoft's representatives at the June 21, 1995,
meeting repeatedly threatened that, if Netscape did not agree to the proposed "special
relationship" with Microsoft, Microsoft would provide that technology on a delayed
basis, if at all.
- Marc Andreessen's contemporaneous notes record,
for example, that Microsoft offered to "consider
licensing us the ability to turn Navigator into a
container," but only "IN THE CONTEXT OF THE
LARGER DEAL." GX 33 (emphasis in original).
Andreessen's notes also reflect that Microsoft
explained that "[t]here are things we can do if we're
working together that we can't do otherwise," and that
Microsoft offered to help Netscape do "tight
integration" with MSN, but only on the condition that
the two companies have a "tight relationship;" "If we
didn't have a tighter relationship, you'd be back to
what a normal ISV can do." GX 33.
- Barksdale testified that, over the weeks leading up to
the meeting, Netscape repeatedly asked Microsoft to
provide "dialer" APIs that would enable Netscape's
Windows 95 browser to access the Internet through a
dial-up ISP. Barksdale Dir. ¶¶ 106, 111.
- Andreessen's notes record that when James
Barksdale raised the issue at the June 21 meeting,
Daniel Rosen acknowledged that Microsoft had an
internal solution to the dialer problem and offered to
provide those internal APIs to Netscape, but only on
the condition that Netscape agree to the previously
outlined "special relationship." "We can fix that
problem. In a perfect world, anyone can plug into
that. With a special relationship with you, you'll be
the first to plug into it. Others will be in the future."
GX 33.
- Andreessen's notes reflect that Microsoft
representative J. Allard then clarified that
"'[d]epending on how we walk out of this room today,
we have a solution for your problem . . . ' or else in 3
months" and that Dan Rosen added, "[i]f we had a
special relationship, you wouldn't be in this position."
GX 33.
70.4.1.4. Microsoft followed through on its threat. When
Netscape rejected Microsoft's market division proposal, Microsoft -- despite Netscape's
repeated requests for assistance -- withheld a working version of the dialer API until
October, a little more than the "3 months" threatened by James Allard at the June 21
meeting. The delay pushed back the release of Netscape's browser until substantially
after the release of Windows 95 (and Internet Explorer) and caused Netscape to miss
most of the holiday selling season.
- On July 7, 1995, Mike Homer wrote to Daniel Rosen
that "I know that we have discussed this with you
before, but I wanted to reiterate how urgently we
need the following information: 1. RNA (remote
network access) phonebook API. . . . We first
requested these in early June . . . ." Barksdale
added: "Dan - these are big deals, please help us."
GX 240.
- On July 18, 1996, Rick Schell wrote to Paul Maritz
that Netscape "repeatedly asked for the API set that
would allow us to make phone book entries.
Microsoft did not provide those APIs until October
1995, which caused us to miss most of the holiday
selling season. However, those APIs were used prior
to that by Microsoft's Plus pack, available at the time
of the Win95 launch." GX 241.
- Barksdale testified that "we did not receive the APIs
and other technical information we had been seeking
until October 1995 . . . or approximately three months
later, which was well after the launch of Windows 95
and was precisely what Microsoft had threatened at
the June 21 meeting." Barksdale Dir. ¶114. See also
GX 240; GX 241; Barksdale, 10/22/98pm, at 53:14-18.
70.4.2. Microsoft also retaliated against Netscape by refusing to
give Netscape a license to a scripting tool that was necessary for Netscape to enable
its Windows 95 users to access certain Internet service providers.
70.4.2.1. In 1996, Netscape attempted to negotiate a
license for a readily-available scripting tool that Netscape needed to make its browser
compatible with some dial-up ISPs.
- Barksdale Dir. ¶207.
- Barksdale testified that Netscape "believed that the
scripting engine was readily available. John
Freeborg, a Netscape employee, confirmed that fact
by making up the name of an ISP and getting on
Microsoft's ISP mailing list, using his home address.
On June 28, 1996, Freeborg received a packet from
Microsoft explaining that the scripting engine we were
requesting was available to ISPs for redistribution on
a royalty-free basis if the ISP signed a Microsoft
license that, among other things, required the ISP to
use Internet Explorer as the 'preferred web browser.'"
Barksdale Dir. ¶ 208. See also GX 239; GX 243; GX
245.
70.4.2.2. A licensing agreement for the scripting tool was
approved by both Netscape's and Microsoft's legal departments, and Netscape
forwarded a signed copy to Microsoft for signature on July 18, 1996.
- Barksdale testified that: "By mid-July, both
Netscape's legal department and Microsoft's legal
department had approved the licensing agreement
under which we sought the scripting tool. Rick Schell
signed the agreement on Netscape's behalf on July
18, 1996. The agreement was forwarded to Microsoft
for signature." Barksdale Dir. ¶ 209
- Julie Herendeen wrote on July 26, 1996 that "John
sent the contract to Ed Mitchell at Microsoft for
signature last Thursday. Ed gave the contract to Carl
Stork, head of the PC/Win 95 division on Thursday
7/18." GX 243.
70.4.2.3. Microsoft responded by writing a letter to
Netscape refusing to discuss the scripting license unless Netscape cooperated with
Microsoft on other, unrelated matters.
- On August 14, 1996, John Freeborg of Netscape
wrote that Will Poole of Microsoft "advised us that
Paul Maritz has written a letter to Rick listing all open
issues between Microsoft and Netscape (one of
which will be the scripting license) where they feel
both parties could be more cooperative. He wouldn't
elaborate on specifics, but they are unwilling to keep
the Win 95 scripting license as a separate issue." GX
248.
70.4.2.4. Netscape never got a license to the scripting tool
and, as a result, was unable for a time to do business with certain ISPs.
- Barksdale testified that "we were never able to
license the scripting tool" and that "Microsoft's refusal
to license the scripting tool to us -- a tool that was
freely available to others for redistribution on a royalty
free basis -- effectively foreclosed Netscape for a
period of time from doing business with those ISPs,
such a Sprynet, that required scripting." Barksdale
Dir. ¶ 213
70.4.3. Microsoft also retaliated against Netscape by embarking on
a relentless campaign to prevent other companies in the computer industry from aiding
Netscape in any way.
70.4.3.1. Microsoft threatened to retaliate against personal
computer OEMs that did business with Netscape by changing the terms of their
Windows 95 licensing agreements.
- Barksdale testified that "Microsoft made clear through
its words and actions that PC computer
manufacturers should not get too close to Netscape
or there could be negative consequences in their
dealings with Microsoft. This was a serious matter for
those OEMs; without a Windows license, and without
cooperation from Microsoft in general, their P.C.
businesses are worthless." Barksdale Dir. ¶ 30.
- Barksdale testified that "I was told that senior
executives of Microsoft, including Bill Gates, called
the CEOs of certain OEMs to warn them that there
would be negative consequences resulting from doing
business with Netscape." Barksdale Dir. ¶ 163.
- When NCR put a Netscape logo on its home page, a
Microsoft representative told NCR that Netscape was
Microsoft's "#1 competitor" and that NCR's "licensing
relationship" with Microsoft "is going to get a lot
harder." GX 192.
- An employee with Netscape's Japanese affiliate
reported on November 12, 1996 that "I have heard
many times from our partners who have OEM
agreement with us to bundle Navigator in their PCs
that, MS threatens them that if they OEM our
Navigator, MS tries to increase the royalty license for
Windows." GX 199; Barksdale Dir. ¶ 172 (same).
- See also infra Part V.C.2.b; ¶¶ 205-208 (Microsoft
threatened to penalized firms that favored Netscape).
70.4.3.2. Microsoft executives also threatened ISPs and
ISVs that entered into cooperative agreements with Netscape.
- Barksdale testified that, when Pacific Bell announced
a joint venture with Netscape, Microsoft executive
Steve Ballmer called to "express[] displeasure over
the announcement," and to say that Pacific Bell "had
become 'an enemy' of Microsoft by doing business
with Netscape." Barksdale Dir. ¶ 164.
- In a June 14, 1996, e-mail, Autodesk told Netscape
that they were "taking a lot of grief (phone calls, email
threats, etc.)" from Microsoft "for working so closely
with you guys." GX 76.
- When Microsoft learned that Attachmate was
bundling Netscape Navigator with their products and
planning to develop a TCP/IP gateway, Microsoft
threatened to bundle its own 3720 emulation software
(which is directly competitive with another Attachmate
product) into the operating system. Attachmate
abandoned its TCP/IP project and ceased distributing
Navigator. GX 91.
4. Acceptance of Microsoft's market-division proposal would
have resulted in both the maintenance of Microsoft's operating
system monopoly and a Microsoft monopoly in the browser
market
71. Had Netscape accepted Microsoft's naked proposal to divide the browser
market, Microsoft would have succeeded in thwarting the browser threat in its incipiency
and would have eliminated the principal threat Microsoft perceived to its monopoly
position in operating systems.
71.1. Netscape's acceptance of Microsoft's offer would have eliminated
the threat to Microsoft's operating system monopoly that Netscape presented in 1995.
71.1.1. Netscape was the only significant browser supplier, and
thus the only significant potential browser threat to Microsoft's operating system
monopoly, in 1995. Accordingly, had Microsoft convinced Netscape to accept its offer,
Microsoft would quickly have gained a sufficient share of browsers to ensure that no
other browser rival controlled browser-related APIs.
- Professor Fisher testified that Microsoft's market division
proposal "is significant, first, because if Netscape had
agreed, Microsoft would have succeeded in eliminating its
only serious browser competitor and in monopolizing the
market for browsers." Fisher Dir. ¶ 99.
- Fisher testified that, from the introduction of Internet
Explorer 1.0 on, Netscape was always Microsoft's "only
serious browser competitor." Fisher Dir. ¶ 99.
- Barksdale testified that "by the end of 1995, Netscape had
an over 70 percent market share for Internet clients."
Barksdale Dir. ¶ 66.
71.1.2. Had Netscape accepted Microsoft's offer, Netscape would
have been relegated to a small and ever-decreasing share of the overall browser
market.
- Marc Andreessen testified at his deposition that "the
proposal . . . would result in crippling Netscape as a
potential competitor of Microsoft, and, indeed, as an
independent company." Andreessen Dep. (played
10/27/98am), at 25:11 - 18:14.
- Barksdale agreed with Andreessen's assessment that
accepting Microsoft's proposal would have crippled
Netscape as an independent company. Barksdale,
10/27/98am, at 29:17.
- Barksdale testified that: "Most of the matters on which they
would have had Netscape confine its work were not
commercially valuable." Barksdale Dir. ¶ 113.
- Andreessen also testified that he believed "in addition that it
is of very little likelihood that Microsoft would have chosen to
live up to" its side of the bargain, "because once we had
been reduced to such a state of weakness, we would have
had no market power whatsoever, and our desires would
have been irrelevant." Andreessen Dep. (played
10/27/98am), at 26:23 - 27:3.
71.1.3. By quickly gaining a substantial share of the browser
business, Microsoft would have ensured that browser rivals could not gain sufficient
usage to induce applications developers to write to their platform rather than to
Windows and would thereby have maintained its operating system monopoly.
- Warren-Boulton testified that, "by reducing the market share
of competing browsers to low levels, Microsoft could
significantly diminish the possibility that applications
developers will write to those browsers' APIs." Warren-Boulton Dir. ¶ 88.
72. Microsoft would also have quickly monopolized the browser market.
72.1. As will be explained, there is a separate product market for Internet
browsers.
- See infra Part VII.B.1; ¶ 384.
72.2. Had Netscape accepted Microsoft's proposal, Microsoft would
swiftly have gained a dominant share of browsers.
- See supra ¶ 71.1.1.
72.3. Microsoft's large share of the browser market would have been
protected by substantial barriers to entry, including Microsoft's control of the standards
and extensions web site developers employ.
- Dr. Fisher testified that "if IE were the dominant browser and
Microsoft decided to support only Windows-based technology,
developers would have little incentive to create applications that
were not Windows-based." Fisher Dir. ¶ 95.
- See also infra Part VII.D.1; ¶¶ 398-400.
72.4. Microsoft would therefore have gained monopoly power over
browsers.
- See infra Part VII.B.3.e; ¶ 390.
B. Microsoft's proposal of market-division agreements to eliminate
other potentially threatening middleware confirms the
anticompetitive character of its course of conduct against the
browser
2. Microsoft similarly attempted to divide markets with Apple
73. Microsoft made other efforts to divide markets with platform-level
competitors. Those efforts establish a pattern and practice of attempts to eliminate
competition by agreement with rivals.
74. Microsoft sought to divide markets with Apple for the purpose of eliminating
competing platform-level technology.
a. Apple's QuickTime multimedia software, like the
browser, is platform-level software that Microsoft
viewed as a potential threat to its operating system
monopoly
75. Apple Computer's QuickTime is its software architecture for the creation,
editing, publishing, and playback of multimedia content (e.g., audio, video, graphics,
and 3D) on the Macintosh and Windows operating systems. QuickTime is cross-platform; developers using QuickTime technology can create multimedia content that
will run on QuickTime implementations for both Windows and Macintosh.
- Tevanian Dir. ¶¶ 47, 50-51, 54, 57-59, 67-68; Tevanian, 11/5/99am, at
27:1-7; 11/4/99am, at 45:3 - 46:6 (testifying concerning QuickTime's API
and cross platform capabilities).
76. Apple, through QuickTime, competes against Microsoft, among other firms,
in providing multimedia functionality to Windows users.
- Tevanian testified that QuickTime competes with Microsoft's multimedia
technologies, including Microsoft's multimedia APIs (DirectX) and media
player (Windows Media Player). Tevanian Dir. ¶¶ 69-70. See generally
Tevanian ¶¶ 57, 60-65.
- Eric Engstrom, Microsoft's former general manager for multimedia,
acknowledged that Apple competes to varying degrees with Microsoft in
trying to convince developers to target their respective multimedia APIs,
codecs and file formats suitable for their respective players. Engstrom,
2/23/99pm, at 35:24 - 36:10, 79:4 - 84:6 See also Tevanian Dir. ¶¶ 57,
60-65 (explaining APIs, file formats, protocols, codecs). Engstrom also
acknowledged that Apple's QuickTime multimedia player is cross-platform, while Microsoft's Direct X multimedia technology is not.
Engstrom, 2/23/99pm, at 97:18 - 98:2.
77. Because QuickTime is cross-platform middleware, Microsoft perceived
QuickTime as a potential -- albeit somewhat distant -- threat to its control over platform-level interfaces and standards that developers invoke and, therefore, to its monopoly
power.
- Microsoft's Ben Slivka testified that Microsoft considered audio/visual
streaming technologies to be part of a "growing collection of technologies"
that "were a threat to the Windows platform" because they could reduce
the applications advantage that Windows has today. Slivka Dep., 9/3/98,
at 243:20 - 245:8 (DX 2591).
- Dr. Avadis Tevanian, Senior Vice President of Software Engineering at
Apple Computer, also recognized the threat that QuickTime could pose to
Microsoft's operating system monopoly, testifying that: "The widespread
popularity and use of QuickTime pose a significant threat to Microsoft.
The cross-platform capability of QuickTime holds the promise of
weakening the symbiotic relationship between the operating system and
application programs that is the foundation of Microsoft's monopoly
position and that poses such a substantial barrier to competition in the
operating systems market." Tevanian Dir. ¶ 75; see also Tevanian Dir. ¶¶
51, 57-59 (describing QuickTime's cross-platform capabilities).
- In his May 1995 "Internet Tidal Wave" memo to his staff on the threat the
Internet posed to Microsoft's operating system position, Bill Gates
specifically expressed his concern about the popularity of QuickTime
formats on the Internet, which he attributed in part to QuickTime being
cross platform, and the difficulty of dislodging established formats, and
directed his staff to develop a competitive strategy. GX 20 at 4, 6.
- Professor Fisher testified that, with Apple's QuickTime technology,
"Microsoft was confronted with platform-level software to which
applications programs could be written" and that the "platform-level APIs
threatened to erode the applications programming barrier to entry into the
PC operating systems by supporting applications programs that could be
used with multiple operating systems." Fisher Dir. ¶ 118, at (a) and (b).
- Warren-Boulton testified that it is "now more the combination of browsers
with cross-platform technologies" that are "seen as a threat . . . ." Warren-Boulton, 11/19/98am, at 48:13-24.
- See infra ¶ 84.
b. Just as with Netscape, Microsoft sought to divide
markets with Apple in order to eliminate the threat that
QuickTime's platform-level components might pose
78. In order to eliminate the possibility that QuickTime's platform-level
components would become part of a platform that could threaten Microsoft's operating
system monopoly, Microsoft sought to allocate markets with Apple in a manner
strikingly similar to its attempt to divide markets with Netscape.
- Professor Fisher expressly drew the parallel between Microsoft's conduct
with regard to Netscape and Apple: Faced with platform level software
that developers could target and thus reduce barriers to entry in the
operating system market, Microsoft responded "by attempting to get the
supplier of the alternative platform to withdraw from offering it and to
concentrate instead on products that did not offer platform potential," and
"was prepared to act to preclude the supplier . . . from succeeding in
offering the platform, 'even if such actions did not make sense from a
business standpoint.'" Fisher Dir. ¶ 118, at (c) and (d).
78.1. In a series of communications with Apple in 1997 and 1998,
Microsoft pressured Apple to cease competing with Microsoft in multimedia playback on
Windows in exchange for Microsoft's supporting QuickTime as a multimedia authoring
solution. Microsoft's proposal was strikingly similar to its proposal to Netscape: In both
instances, Microsoft offered the firm producing platform-level software Microsoft's
support and a free hand in a complementary product in return for abandoning platform-level client software for Windows; and in both instances, Microsoft threatened to injure
the other firm's business if it did not agree to the proposal.
78.1.1. Microsoft first proposed its market-division scheme to
Apple in 1997 and continued its efforts to secure Apple's acquiescence throughout the
year.
- In April 1997, Microsoft's Eric Engstrom and Christopher
Phillips suggested to Apple that they cede the multimedia
playback market to Microsoft and focus solely on the
"authoring" area of multimedia, i.e., the development of
software tools used to create multimedia content and the
APIs to enable such tools. Tevanian Dir. ¶ 78; Schaaff
Dep., 1/13/99, at 192:8 - 196:1 (at a meeting between Tim
Schaaff of Apple and Eric Engstrom and Chris Phillips of
Microsoft, MS proposed that Apple scale back its efforts in
multimedia playback on Windows in return for MS support
for QuickTime as a multimedia authoring solution). Microsoft
did not at that time offer multimedia authoring capabilities or
APIs. Schaaff Dep., 1/13/99, at 193:13-21.
- Later in 1997, Microsoft repeatedly pressured Apple to cede
the multimedia playback market to Microsoft. In an August
1997 meeting between Apple and Microsoft, Microsoft urged
"Apple to withdraw from the market for multimedia playback
capability." Tevanian Dir. ¶ 80.
- The following month, Mr. Engstrom "again urged Apple to
focus on the authoring segment and to cede the playback
business to Microsoft." Tevanian Dir. ¶¶81-83; Schaaff Dep.,
1/13/99, at 196:3 - 199:4; see infra ¶ 78.2.
- In October 1997, at another meeting between Apple and
Microsoft, Microsoft stated that it "would allow Apple to
continue with QuickTime playback for the Mac operating
system, but would require Apple to relinquish the QuickTime
playback capability in Windows." Tevanian Dir. ¶ 84. At
this meeting, Mr. Phillips again encouraged Apple to back off
QuickTime as a playback technology for Windows. In
return, Phillips offered support for QuickTime as an
authoring technology. Schaaff Dep., 1/13/99, at 200:16 -
206:11. Phillips also offered to support QuickTime in some
way on the Windows CE operating system. Schaaff Dep.,
1/13/99, at 206:15 - 207:14.
- Engstrom admitted that one of his goals in all his
discussions with Apple was to have Apple agree on a single
audio/video playback runtime for Windows based on
Microsoft's Direct X, and that he told Apple that Microsoft's
support for QuickTime as an authoring solution on Windows
was dependent on that agreement. Engstrom, 2/24/99am,
at 25:16 - 28:12; 36:20 - 37:15. Engstrom also admitted that
he hoped to "move the locus of competition upstream," in
other words, away from competition in audio/visual playback
on Windows. Engstrom, 2/24/99am, at 36:6-19. Engstrom
also conceded that it would be pointless for Apple to
continue to offer its own multimedia runtime on Windows if it
accepted Microsoft's proposal. Engstrom, 2/24/99am, at
40:6 - 42:13; 55:6-23.
78.1.2. Microsoft continued to propose dividing the multimedia
business between the companies -- including giving Apple a free hand in authoring --
well into 1998.
- On February 13, 1998, Dr. Tevanian met with Microsoft's Don Bradford to discuss the
technical problems that Windows and Internet Explorer caused with QuickTime and to
discuss threatening comments made by Microsoft employees. Tevanian Dir. ¶¶ 85-87.
At that meeting, Mr. Bradford "conveyed the same proposal that Microsoft had
presented in the past. Specifically, if Apple would abandon the playback segment of the
business, Microsoft would be willing to endorse QuickTime as the solution for the
authoring portion. Mr. Bradford told me that Mr. Gates thought that would be a way to
resolve our dispute." Tevanian Dir. ¶¶ 88-89. Dr. Tevanian testified that Bradford's
response to Tevanian "was very simple, and although he did it in a less threatening
way, he said -- he basically said, 'Well, we want to fix this; we want to be able to work
together, and Bill wonders if a way to solve this is for us to take playback and you to
take authoring.' And I told him simply, 'No, that's not acceptable.'" Tevanian,
11/5/98am, 29:10-25.
- Phillip Schiller, an Apple marketing vice-president, testified
that, in an April 1998 telephone conversation with Eric
Engstrom, Engstrom offered to support Apple in multimedia
authoring but made clear that this support was conditioned
on Apple ceasing competition in multimedia playback on
Windows. Engstrom told Schiller that Apple had to "give up
playback on Windows." Schiller Dep., 1/13/99, at 240:19 -
242:22; see also Tevanian Dir. ¶¶ 90-92 (Engstrom tells
Schiller that Apple would "have to give up multimedia
playback on Windows" in order to work together with
Microsoft on authoring.).
- Mr. Engstrom admitted that he told Apple that, if Apple
agreed to use Microsoft's DirectX as their runtime for
Windows, then Microsoft would support Apple's authoring
technology, and that if Apple did not agree to adopt the
DirectX runtime, then Microsoft would enter the authoring
business. Engstrom, 2/24/99am, at 37:16 - 39:25.
Although Engstrom denied that he had offered to stay out of
authoring, he admitted that he told Apple that if Apple
targeted its authoring solution to Microsoft's DirectX, that he
"would probably not invest as rapidly in that solution as
otherwise." Engstrom, 2/24/99am, at 39:15-25.
- At a June 15, 1998, meeting, attended by Eric Engstrom as
well as Dr. Tevanian and Apple CEO Steve Jobs, Microsoft
proposed that "Microsoft would take over the playback
market for Windows, while allowing Apple to control the
much smaller playback business for the Macintosh."
Tevanian Dir. ¶¶ 93-94. Microsoft's proposal entailed, inter
alia, Apple adopting
-
DirectX as the runtime for Windows,
- Microsoft's proprietary streaming technology, and
-
Microsoft's AAF file format for authoring, all of which Dr.
Tevanian believed were inferior.
Tevanian Dir. ¶ 95.
Dr.
Tevanian testified that "Microsoft's proposal amounted to a
forced abandonment of one of Apple's most successful and
innovative products" and adoption of Microsoft's playback,
streaming, and authoring technology. Tevanian Dir. ¶ 96.
Microsoft's agenda for the meeting suggests that Microsoft's
Direct X will be the only runtime on Windows: "Run-Time is
Direct X on Windows, QuickTime on the Mac." GX 912; see
also GX 908 (July 6, 1998 Waldman email to Gates et al.)
(summarizing Mr. Jobs' view of the proposal: "essentially
'Apple should give up QT and use [Microsoft's] stuff'").
78.2. In order to coerce Apple's acceptance of its market-division
proposal, Microsoft told Apple that, if Apple did not agree to Microsoft's offer of a free
hand in authoring, Microsoft would engage in predatory conduct.
- In September 1997, Microsoft's Engstrom "again urged Apple to
focus on the authoring segment and to cede the playback business
to Microsoft." Tevanian Dir. ¶¶ 81-83. At a meeting at the
Fairmont Hotel in San Jose, California, Mr. Engstrom told Apple's
Schaaff that Microsoft intended to control multimedia playback on
Windows and that Microsoft would devote 100-150 engineers to
authoring if that was what was necessary to control multimedia
playback. Engstrom also told Schaaff that Bill Gates did not think
that authoring was a significant business opportunity for Microsoft,
but that Microsoft would be willing to invest whatever was
necessary to control multimedia playback, "even if it didn't make
sense from a business standpoint." Schaaff, 1/13/99, at 196:3 -
199:24.
- Schaaff took this as a threat that, if Apple did not "back off" from
the Windows playback business, Microsoft would double or
substantially increase the size of their team to compete both in
playback and authoring. Schaaff Dep. 1/13/99, at 196:3 - 199:24;
Tevanian Dir. ¶ 83 ("Mr. Engstrom noted at the meeting that
Microsoft's Bill Gates was not interested in an authoring program
because the market for this product was too small. He assured the
Apple representatives, however, that if Microsoft needed to make
an investment in providing authoring tools to push Apple out of the
playback market, then Microsoft would devote all the necessary
resources to accomplish this goal.").
- Dr. Tevanian's interpretation was the same as Mr. Schaaff's:
"What Mr. Engstrom was saying was that he made us an offer,
which is, if we were to cede the playback market, he would give us
the authoring market. And if we didn't take that offer, he would
immediately deploy engineers to just kill us in that space, too."
Tevanian, 11/5/98am, at 82:19-23. "The threat was that they would
leverage their other advantages in the market and just pound on us
in any way they possibly could." Tevanian, 11/5/98am, at 84:2-4.
Engstrom admitted he had said that developing an authoring
solution was "not the highest return for that particular investment."
Engstrom, 2/24/99am, at 38:22-24.
- Professor Fisher explained Microsoft's anticompetitive purpose in
threatening to enter into the multimedia authoring business:
"Microsoft was going to devote 100 to 150 engineers to competing
against Apple on this, even though, said the Microsoft
representative, it made no business sense. Sounds like a threat to
me. It doesn't sound like the kind of thing that one company says
to another and says, you know, we are going to hang tough on this.
This says we are going to go out of our way to hurt you." Fisher,
1/6/99pm, at 70:14-21. Professor Fisher further noted: "If it doesn't
make sense from a business standpoint, you have to ask what is
the possible motive for it, and here the motive is to get Apple to
cooperate." Fisher, 1/6/99pm, at 73:11-14; see also Fisher Dir.
¶117, 118 (Microsoft's actions with regard to Apple's multimedia
playback technology show that Microsoft was "prepared to act to
preclude the supplier of a potential platform-level software from
succeeding in offering the platform, even if such actions 'did not
make sense from a business standpoint'" and "Microsoft will
respond immediately to prevent any other firm from writing
platform-level software. This is true even though this software
could increase the functionality and performance of, and thus
demand for, Windows-based PCs.").
- Engstrom conceded that he told Schiller that if Microsoft and Apple
did not work together on multimedia software for Windows, then
Microsoft would have to offer authoring solutions that might be
incompatible with Apple's, that he was "mystified by Apple's
insistence on going it alone in developing and marketing a
multimedia runtime for Windows, and that "given Microsoft's
resources and expertise," Microsoft was likely to be successful in
authoring technology. Engstrom Dir. ¶ 68.
c. Microsoft's purpose in proposing a division of markets
to Apple was to ensure Microsoft's continued control
over platform-level interfaces
79. Microsoft's purpose in attempting to allocate markets with Apple, as with its
attempt to allocate markets with Netscape, was to prevent Apple from successfully
establishing platform-level software that might reduce Microsoft's control over interfaces
and standards that developers use and thereby erode the barriers to entry to the
operating system market.
79.1. First, Microsoft's purpose is evident from both its contemporaneous
documents and the testimony of its witnesses.
- Engstrom's supervisor, David Cole, in an email to Gates, Engstrom
and Waldman made clear that Microsoft's primary goal in its talks
with Apple was to "get Apple to give up on having a runtime on
Windows." GX 270 (4/28/98 Cole email).
- Even Engstrom, who claimed that he never explicitly asked Apple
to give up its Windows' runtime, admitted at trial that everyone
knew it would make no sense for Apple to continue to offer its own
multimedia runtime on Windows if it accepted Microsoft's proposal
to use Microsoft's Direct X runtime. Engstrom, 2/24/99am, at 40:6 -
42:13; 55:6-23 ("If they are going to adopt our runtime, they are
going to have to, at some level, give up mentally and emotionally,
on building this duplicative set of services, because it wouldn't
make sense for them to build a service on top of ours that uses our
services while they are still building a service that is -- you know,
that they view as their hope for their future, you know, that sits next
to this piece of Windows.").
79.2. Second, Microsoft expressly communicated its purpose of
controlling platform-level software to Apple.
- Timothy Schaaff testified that Microsoft multimedia chief Eric
Engstrom told him and other executives that "Microsoft wanted to
have control over the user interface . . . and that Microsoft was
determined that the essential APIs that were the foundation of the
operating system should all come from Microsoft and not come
from a third party." Schaaff Dep., 1/13/99, at 194:21 - 195:18;
Schaaff Dep., 8/28/98, at 283:21 - 284:11 ("they stated that it was
Microsoft's opinion, point of view, that they intended to control APIs
for the playback of multimedia content on the Windows platform,
and hence they did not wish to see proliferating or competing with
Microsoft with a separate set of APIs . . . in the playback space")
(DX 2506).
- Dr. Tevanian testified that "Engstrom bluntly warned Mr. Schiller
'We're going to compete fiercely on multimedia playback and we
won't let anyone have playback in Windows. We consider that part
of the operating system, so you're going to have to give up
multimedia playback on Windows.'" Tevanian Dir. ¶ 91; Schiller
Dep., 1/13/99, at 240:19 - 242:22. Dr. Tevanian recounted that,
"what they were proposing at every level -- and sometimes very direct
threats -- was effectively killing QuickTime. . . . And Mr. Hoddie said,
'do you want us to knife the baby'? That was his words. 'Knife the baby'
meaning kill QuickTime. And Mr. Phillips repeated back to him, 'Yes,
we're talking about knifing the baby.'" Tevanian, 11/5/98am, at 28:15 -
30:4.
- Steve Jobs confronted Microsoft about Microsoft's public messages
that Microsoft would use its power in the PC operating systems
market to kill QuickTime. In an e-mail to Bill Gates, Steve Jobs
stated that "There is one thing that threatens to be quite divisive,
and that is the Microsoft NetShow team's recent behavior. They
are really going out of their way to say that they intend to kill
QuickTime, and are being quite threatening and rude about it. . . .
We intend to fight and win with QuickTime, and I hope this honest
and proper effort doesn't meet with down and dirty tactics and
tough rhetoric from the NetShow group -- it could really tarnish our
entire, budding relationship." GX 904; see also GX 897 (In a
January 1998 email to Gates and others re "Steve Jobs Call," Don
Bradford reported on a telephone call he received from Jobs.
"Steve called back to express his concern over NetShow's public
message about killing QuickTime."). Dr. Tevanian testified that
Jobs' was referring to statements made by Microsoft's Netshow
team that Microsoft would kill QuickTime because Microsoft's
technology would be everywhere because it would be bundled with
Windows and with Internet Explorer on the Macintosh, but
Microsoft would never allow QuickTime to survive on Windows.
Tevanian, 11/5/98am, at 94:16 - 95:8.
79.3. Third, Microsoft's purpose is evident from the nature of the
proposal: Microsoft wanted Apple to cease developing complementary software that
runs well on Windows (an activity it usually encourages); Microsoft's proposal, if
accepted, would have reduced demand for Windows and thus makes sense only as an
effort to eliminate potential competition to Microsoft's operating system monopoly.
- See infra ¶ 300.
80. Had Apple accepted Microsoft's proposal, Microsoft's efforts to pressure
Apple to cease competing on the Windows platform would likely have reduced
competition and innovation in multimedia playback, particularly in the development of
cross-platform APIs.
- Dr. Tevanian testified: "Yes, it is true that the Microsoft proposal was that
Apple cede the market for multimedia playback on Windows. But from our
perspective, that was essentially ceding it for everything, because, let's remember,
as we talked about yesterday, one of the goals for QuickTime was to be cross-platform, so you could develop content and run it on either Windows, or
Macintosh, or any other operating system. If we couldn't put that technology on
windows -- if we had to cede that to Microsoft, then it would have undermined
one of the primary goals of the whole product. Having it on the Macintosh would
have been irrelevant." Tevanian, 11/5/98am, at 27:1-11.
- Microsoft's efforts to convince Apple to give up QuickTime multimedia
playback on Windows not only would have required Apple to cease
innovating in multimedia playback on Windows, but would have also
impeded Apple's ability to innovate on the authoring side because they
would be limited to using the Windows playback mechanism. Schaaff,
1/13/99, at 203:3 - 205:3.
- Mr. Engstrom admitted that, if Apple had accepted Microsoft's proposal
and had not continued shipping a multimedia runtime for Windows, Apple
would have been dependent on Microsoft for execution of Apple's
authoring solutions. Engstrom, 2/24/99am, at 48:5 - 49:1.
d. Microsoft retaliated against Apple, just as it did with
Netscape, when Apple refused to accept Microsoft's
proposal
81. When Apple refused to accept Microsoft's proposal to cease competition in
multimedia on Windows, Microsoft retaliated against Apple. The retaliation ranged from
inserting misleading error messages in Windows to offering or withholding assistance to
Apple as it suited Microsoft's strategic goals.
81.1. Microsoft introduced misleading error messages into Windows that
urged users to replace QuickTime with Microsoft technology.
- Dr. Tevanian testified that Microsoft has inserted misleading error
messages in Windows informing users that they might not be able
to play certain multimedia files and asking users if they wanted to
reconfigure their systems to use Microsoft's Active Movie
technology instead of Apple's QuickTime technology. Tevanian Dir.
¶¶ 108-110 and Attachment 5; GX 917; GX 918; Tevanian,
11/4/98am, at 27:12 - 28:22.
- Dr. Tevanian, an experienced software engineer, testified that such
error messages are unlikely to issue accidentally. Tevanian,
11/4/99am, at 61:17 - 62:5.
81.2. When Microsoft made changes to Internet Explorer 4.0 and
Windows that resulted in impaired functioning of QuickTime, Microsoft opportunistically
responded to Apple's requests for assistance by fixing the problem when it suited
Microsoft's strategic interests and refusing to provide meaningful assistance when it did
not.
81.2.1. With the release of Internet Explorer 4.0, Microsoft changed
the interaction between Windows and Internet Explorer so that data in certain media
files were preferentially routed to Internet Explorer for playback. The changes
prevented QuickTime from processing the data and frustrated users' attempts to access
certain content. QuickTime also experienced additional difficulties operating with
Internet Explorer 4.0 and Windows 98. These problems occurred at the very time that
Microsoft tried to convince Apple to give up its multimedia platform-level software.
- Dr. Tevanian testified: "When Microsoft produced its first plug-in
capable browser [Internet Explorer 3.0] and needed to compete in
the Netscape-dominated market by being technologically
compatible, Microsoft used and adhered to Netscape's plug-in
architecture. With the growth of Microsoft's browser market share
through the bundling of Internet Explorer and Microsoft multimedia
software with Windows, Microsoft reduced the compatibility
between its browser and the open Netscape standard, starting with
the introduction of Internet Explorer 4.0." Tevanian Direct ¶ 102.
- Tevanian further testified: "With the successive releases of
Microsoft's Internet Explorer 4.0, Microsoft Windows 98, and
Microsoft multimedia software, Apple has seen a steady
degradation of QuickTime's capability to play back a variety of
QuickTime compatible media file formats while operating with
Microsoft's Internet Explorer running on the Windows operating
system." Tevanian Dir. ¶ 100; see also Tevanian Direct ¶ 101 &
Attachment 4 (chart of test results with various formats); Tevanian,
11/4/98am, at 28:23 - 29:13 (" In that case what would happen is
when a user was browsing the web, looking at web pages, and
would find QuickTime content, files that were based on QuickTime,
instead of playing using QuickTime, even if QuickTime was
installed, Internet Explorer would play -- would try to play it using
Microsoft technology and would often fail. So QuickTime was not
being allowed to actually access the data and play it correctly. The
user wouldn't know it was broken, and often the web page would
show that it required QuickTime, yet QuickTime would not be
invoked after release -- excuse me, after being installed. And we
had no way to solve this that we knew of.").
- Schaaff testified that, when Microsoft introduced Internet Explorer
3.0, it promoted its compatibility with the Netscape browser plug-in
APIs. Since QuickTime already supported the Netscape browser
plug-in API, Apple was able to ensure that QuickTime was
generally compatible and operated properly with both Navigator
and Internet Explorer 3.0. With the release of Internet Explorer 4.0,
certain file types that were previously routed to QuickTime were no
longer routed to QuickTime. Apple's investigation revealed that the
mechanism for routing media types in the Windows operating
system, the Windows registry, which is largely undocumented, was
not routing media types to QuickTime as expected. Depending on
the file type, this can result in the user not being able to access the
content at all or in an impaired manner. Schaaff, 1/13/99, at
211:16 - 222:5; Tevanian Dir. ¶¶ 102-106. Apple's efforts to
reverse engineer the Windows registry software to correct the
problem met with only limited success. Tevanian Dir. ¶ 105.
- Dr. Tevanian testified that the introduction of incompatibilities could
undermine the establishment of Apple's multimedia platform.
Tevanian, 11/4/98am, at 45:3 - 46:12. It is also clear that these
problems were occurring at the same time as Microsoft's public
campaign to convince developers that Microsoft's multimedia
technology would "kill" QuickTime on the Windows platform.
Tevanian, 11/5/98am, at 94:3 - 95:17.
- Microsoft email confirm that "support for the Windows file types are
build (sic) into IE itself," that the Windows registry gives a
preference to Microsoft's ActiveX controls, and that Microsoft
discouraged Apple from writing its own ActiveX controls to route
playback of both of Microsoft formats and industry standards like
MIDI. GX 911; GX 274.
81.2.2. When Apple first requested Microsoft's assistance,
Microsoft corrected one of the problems caused by its redesign because doing so
suited its strategic objective of blunting other platform-level threats.
- In August 1997, Dr. Tevanian sent Bill Gates an e-mail
explaining that Internet Explorer 4.0 disabled QuickTime and
QuickTimeVR on Windows and that IE4 set the default for
".mov" media files to Microsoft's ActiveMovie, rather than
QuickTime. GX 265 (8/8/97 Tevanian email to Gates).
- Unbeknownst to Dr. Tevanian, Mr. Gates forwarded
Tevanian's e-mail to Paul Maritz. Mr. Gates sought to
ensure that the problem was used to Microsoft's advantage;
he instructed Mr. Maritz: "I want to get as much mileage as
possible out of our browser and JAVA relationship here. In
other words a real advantage against SUN and Netscape.
Who should Avie be working with? Do we have a clear plan
on what we want Apple to do to undermine SUN?" GX 265
(8/8/97 Gates email to Maritz).
- Dr. Tevanian testified that Microsoft responded to his
request to Gates by fixing the file associations for the
specific file type he mentioned. Tevanian, 11/4/98am, at
29:14-22, 54:9-22.
- Internal Microsoft email confirm that Microsoft provided a
method that overrode the ActiveX preference for MOV and
QT, (MOV is the format Dr. Tevanian asked Gates about),
but that Microsoft did not wish to do this for any other file
formats. GX 911 (8/5/98 Perry email; 8/6/98 Larkin email);
GX 265 (8/8/97 Tevanian email to Gates).
81.2.3. By contrast, when Apple rejected Microsoft's proposal to
allocate multimedia technology, Microsoft abandoned meaningful efforts to help Apple
solve the compatibility problems.
- Dr. Tevanian testified that he could not understand why MS
could and did correctly fix the .mov problem, but not the
other problems. Tevanian, 11/4/98am, at 31:9 - 32:9, 54:9-22.
- Dr. Tevanian also testified that Microsoft delayed responding
to Apple's complaints. Tevanian, 11/4/98am, at 29:14-22
("We contacted Microsoft. In fact, I recall at that time, the
first time we noticed this, I contacted Bill Gates directly and
asked him to fix it in one specific area, which he did. He got
it fixed somehow. But in many other areas, it never got
fixed. We tried to interact with Microsoft. We were getting
close to shipping QuickTime 3. We weren't getting fast-enough responses. We did try to solve it ourselves. We
couldn't solve it. And that was the end of that story.");
Tevanian, 11/5/98pm, at 77:8-15 ("our engineers questioned
the data they received from Microsoft. Here we are now,
again, in context, a full year -- in fact, it's almost a year to
the date after which I first notified Mr. Gates that we were
having problems, and with that notification, Microsoft was
somehow able to fix one of the file types, and we just could
not understand why if they fixed one of them they didn't fix
all of them.").
- Dr. Tevanian testified that, after trying for months to obtain
information or assistance from Microsoft to correct the
problems and receiving an inadequate response, Apple
received the beta for the Windows Media Player a few days
before the final product was to ship, a grossly insufficient
amount of time to detect and correct any problems that
might exist. Tevanian, 11/4/98am, at 36:22 - 38:10.
- Contemporaneous documents confirm Apple's repeated
attempts to persuade Microsoft to correct the problem with
Windows taking over the QuickTime file associations. For
example, on July 21, 1998, Tim Schaaff sent a lengthy e-mail to Engstrom and Cristiano Pierry at Microsoft detailing
the problems, and noting that the fixes Microsoft claimed to
have provided Apple did not solve the problems. Schaaff
explained in part that: "To the extent that Internet Explorer 4
relies on this undocumented info from the Windows Registry
to determine which software should be invoked to process
different MIME types on the web page, third-party
developers, like Apple, are getting hurt. . . . It's
unacceptable that every time a new version of the Media
Player, or Direct X, or Windows itself is installed that
QuickTime is getting overridden by your software." GX 272.
A week later, having received no response, Schaaff resent
this mail. GX 272 (7/28/98 Schaaff email)
- Another week later, on August 4, 1998, still having received
no response, Apple CEO Steve Jobs again requested
Microsoft's assistance to solve the file association problem.
GX 911 (8/4/98 Jobs email to Maffei).
- On August 5, 1998, Microsoft's Pierry finally responded,
suggesting only that Apple develop an ActiveX control, a
Microsoft proprietary technology, but at the same time
discouraging Apple from doing so. GX 272.
- In an internal Microsoft e-mail to Jim Allchin, Pierry
explained Microsoft's conduct. First, he noted that the
reason that QuickTime is able play .mov files was because
Microsoft had gone out of its way to provide an overwrite
method. GX 911. 8/5/98 Pierry email) Pierry stated, "I really
do not want to provide a similar mechanism to enable them
to" play other file types. GX 911. Pierry then explained that
his "response to Apple right now is sorry, but support for the
Windows file types are build (sic) into Internet Explorer itself.
The only way to take over, and we discourage you from
doing so, is to write your own active x control. It turns out
that they can probably just delete our MIME types from the
registry, then IE would have to use the plug-in. But this
would be a very wrong thing for them to do and it would
cause app compatibility problems for them." GX 911 (8/5/98
Pierry email) (emphasis added).
- None of the email traffic with Apple mentions the
undocumented "enable plug-in flag" which Microsoft wrote to
enable .mov and .qt to play properly Engstrom, 2/24/99am,
at 12:21-25. Nor does e-mail traffic mention the alleged
defects in Apple's plug-in instruction that Microsoft proffered
at trial as the cause of the problem. Engstrom, 2/24/99am,
at 20:25 - 21:20.
81.3. Microsoft also retaliated against Apple by inducing third parties not
to support QuickTime on Windows.
- Tevanian testified that a third party hardware vendor TrueVision
was prohibited by Microsoft from marketing or promoting driver
software for QuickTime for Windows, and from writing driver
software for QuickTime for Windows that would operate with more
than the Final Cut product. Tevanian Dir. ¶¶ 134-138; Schiller Dep.,
1/13/99, at 243:15 - 247:12.
- Engstrom admitted that Microsoft entered into a contract with
TrueVision that prohibited TrueVision from developing or promoting
non-Microsoft interfaces for its driver software for approximately
four months. Engstrom Dir. ¶ 120.
e. Just as with Netscape, Microsoft's proposal was
unrelated to any efficiency-enhancing sharing of
technology
82. Microsoft's effort to force Apple to exit the playback market for Windows
was unrelated to achieving any efficiency or proconsumer benefit.
82.1. First, Microsoft asserts that it was simply seeking a way to increase
consumer satisfaction by providing uniform standards for multimedia. Engstrom Dir. ¶
46-47. But forcing Apple to exit the playback market on Windows was not reasonably
necessary to achieve workable standards, which could have been achieved through
cross-licensing codecs (and other software) and/or cooperation on standards and
protocols for data creation, storage and transfer, while maintaining the consumer
benefits and innovation that competition provides.
- Tevanian testified that, through cross-licensing, Microsoft and
Apple could "establish a level playing field where everyone can
compete. We viewed it as important to have open standards where
customers could buy technology and vendors could have different
implementations of the technology." Tevanian, 11/5/99am, at 60:4-13.
- Tevanian explained that a "single approach" has benefits but
ending competition was not necessary to achieve them: "we have a
different view of how to achieve that than Microsoft does. In
particular, we view the way to achieve that is to establish open
standards where everyone can compete with different
implementations, and they could compete based on the quality of
the implementations or other metrics that would be important to
consumers. In the Microsoft model, the goal was to control it, so
not only would they control the interfaces, but they would control
the implementations. . . . So while it may have appeared to have
benefitted consumers, the way they were proposing to achieve it
we did not agree with." Tevanian, 11/5/99pm, at 31:8 - 32:11.
- Timothy Schaaff testified that Apple personnel
- redacted -
DX 2586; Schaaff Dep., 8/28/98, at 508:7 - 512:10 (DX
2586A) (sealed). Mr. Schaaff also testified about discussions
internally and with Microsoft about other arrangements with
potential benefits for consumers that did not depend on eliminating
competition in the playback market. DXs 2586; Schaaff Dep.,
8/28/99, at 337:19 - 338:15 (DX 2586) (licencing codecs) (sealed),
353:15 - 354:3
- redacted -
361:18 - 365:2 (same).
- In its June 15 written proposal, Microsoft listed a number of items,
such as cross licensing codecs, that would have improved
compatibility and interoperability issues but do not inherently
require that the two firms agree to cease competition. GX 912.
82.2. Second, the contemporaneous evidence demonstrates that, far
from seeking to benefit consumers by improving the ensuring compatibility, Microsoft's
overriding objective was to control the APIs to which developers write.
- See supra ¶ 78.
- Engstrom wrote that it was important to convince Intel not to assist
Sun in writing Java multimedia APIs, "esp. those that run well, ie
native implementations, on Windows." GX 235.
- Microsoft told Intel that it sought to eliminate platform-level threats
through a strategy of "embracing" the platform-level standards,
"extending" them in Microsoft-dependent ways, and thereby
"extinguishing" the threat to Microsoft's control over standards.
See supra Part V.A.3; ¶ 91
- See infra ¶ 84.
83. Engstrom's testimony (Engstrom Dir. ¶ 49) that he never told Apple that it
would have to give up its runtime on Windows is not credible.
83.1. Engstrom's testimony is contrary to the more reliable testimony of
Schaaff and Schiller, as well as inconsistent with the contemporaneous documents.
- See supra ¶¶ 78.2, 79.2
- Microsoft's David Cole clearly expressed to Mr. Gates and Mr.
Engstrom that eliminating Apple's QuickTime runtime was
Microsoft's ultimate goal: "If we can get Apple to give up on having
a runtime on Windows . . . ." GX 270.
83.2. Engstrom ultimately conceded that Apple would have little incentive
to develop a runtime if it accepted Microsoft's offer.
- Engstrom testified that "none of the presentations . . . were
predicated on the fact that they would have to stop" offering a
QuickTime runtime on Windows (Engstrom, 2/24/99am, at 51:22-24) but later conceded that if Apple adopted the Microsoft runtime,
"they are going to have to, at some level, give up, mentally and
emotionally, on building this duplicative set of services because it
wouldn't make sense . . . ." Engstrom, 2/24/99am, at 55:6-23.
What Mr. Engstrom appears to mean by duplicative is competitive.
Engstrom, 2/24/99am, at 51:14-21; 35:24 - 36:13.
2. Microsoft also attempted to divide markets with RealNetworks,
using the same carrot and stick approach it used with other
potential platform rivals
84. Microsoft engaged in a similar attempt to divide markets with RealNetworks as part
of its pattern and practice of seeking anticompetitive agreements to eliminate potential threats to
the applications barrier to entry.
84.1. Microsoft perceived that RealNetworks (then known as Progressive
Networks) multimedia streaming software had the potential to develop into a platform threat, at
least in the multimedia area.
- On June 5, 1997, Microsoft's Jim Durkin reported on an internal
Microsoft strategy meeting attended by Messrs. Gates, Maritz and Muglia.
GX 1576 (6/5/97 Durkin email). Durkin quoted Microsoft Vice-President
Muglia as saying: RealNetworks "is like Netscape. The only difference is
we have a chance to start this battle earlier in the game." GX 1576.
Durkin also reported that Gates and Maritz had made the decision that
"Winning the streaming battle means three things - winning the file format
war, winning the client architecture war, and winning the server wars."
GX 1576.
- Mr. Maritz testified that, as of June 1997, Gates believed that streaming
was a strategic area that Microsoft needed to win. Maritz, 1/27/99am,
56:25 - 57:10; GX 1576.
- Mr. Maritz also admitted that, although he believed in June 1997 that
RealNetworks did not pose the same sort of threat as Netscape, it "had
the potential to grow, over time, into a software platform.@ Maritz,
1/27/99am, 57:15 - 58:4.
- Mr. Engstrom testified that RealNetworks presented some set of APIs that
compete with Microsoft's APIs for developer attention (Engstrom,
2/23/99pm, at 35:24 - 36:10; 83:21 - 84:6) and that RealNetworks
technology operates cross platform (Engstrom, 2/23/99pm, 98:3-25).
84.2. Microsoft told RealNetworks that it viewed the "core" multimedia
streaming functionality on the client as part of the operating system and requested that
RealNetworks cease competing with Microsoft in offering that functionality.
- Bruce Jacobsen, Chief Operating Officer of RealNetworks and a former
Microsoft employee, testified that he met with Microsoft Vice-President
Robert Muglia in the summer of 1997, to discuss, among other things,
Microsoft's distribution of RealNetworks software with Windows and
Internet Explorer. Jacobsen Dep., 1/13/99, at 153:2 - 158:25; cf. GX 1369
(sealed) (7/18/97 Agreement between Progressive Networks and
Microsoft); GX 884 (sealed) (6/17/97 agreement between Progressive
Networks and Microsoft).
- Mr. Jacobsen recorded a summary of the meeting shortly after the
meeting. GX 1368. Mr. Jacobsen summarized the meeting as follows:
AWas cordial but pointed. His basic message was the [sic] wanted us out
of core AV. He said that MSFT had concluded that fundamental datatypes
like words and numbers were in essence a core part of the operating
system . . . He said that he thought video was one of the most exciting
datatypes -- since monitors were visual things, video had to be though
[sic] like 'words'. and microsoft had to control this franchise. He said
that anyone who competed against MSFT in the operating system 'lost' --
that there were only two people left in town who still competed against
msft as a potential OS vendor -- Sun and Oracle -- and the rest had been
obliterated, and MSFT was targeting these last two. He referenced their
scalability day as part of killing Sun. So the message was that if we
wanted to do value add on top of their video, fine; if not, we were an OS
contender and msft would target us for obliteration. He cited PeopleSoft
as ok -- he said adobe had pretensions of OS, but had basically backed
off.@ GX 1368.
84.3. In order to induce RealNetworks to cease competing in core streaming,
Microsoft proposed that, if RealNetworks stopped competing in base level streaming, Microsoft
would give its full support to RealNetworks as a value-added software provider; but if
RealNetworks continued to compete, Microsoft would use its resources to injure it.
- Mr. Jacobsen testified that Muglia explained that Microsoft would seek to
injure RealNetworks' business if RealNetworks continued to compete in
the fundamentals of audio/visual streaming. Jacobsen Dep., 1/13/99, at
155:4 - 158:25. Mr. Jacobsen quotes Muglia as saying that Microsoft had
won most of the operating system wars and the only remaining threats
were Oracle and Sun. Jacobsen Dep., 1/13/99, at 156:22 - 157:4. Muglia
said Microsoft was trying "to reduce the economic viability of those
companies so they wouldn't have the wherewithal to invest and position
themselves as operating system competitors of Microsoft." Jacobsen Dep.,
1/13/99, at 156:22 - 157:4. Muglia told Jacobsen that a company like
Adobe had at one point "operating system pretenses" or "pretensions" but
had been chased out of that space. Jacobsen Dep., 1/13/99, at 157:5-10.
Muglia told Jacobsen that Microsoft wanted RealNetworks to be like
PeopleSoft, a value-added provider that builds applications on top of
operating systems but does not threaten any core part of Microsoft's
environment. Jacobsen Dep., 1/13/99, at 157:11 - 158:8. Muglia
continued: "On the other hand, if you try to do the fundamentals of
streaming audio and video, then we would view you as a core competitor
and use all our resources to hurt you in your core businesses. " Jacobsen
Dep., 1/13/99, at 157:18-22; 158:9-25 (Jacobsen asked Muglia
whether Microsoft was asking RealNetworks to abandon core
streaming audio and video and Muglia replied affirmatively).
- Mr. Muglia warned RealNetworks, Jacobsen testified, that
"Microsoft would aggressively target us as a company, using all of
Microsoft's resources, if we stayed in the audio and video space . .
. . Bob also said, and I agreed with him, that Microsoft had been
successful prior in targeting companies and having severe
economic effects on them. Bob did not use Borland as an
example, but Borland certainly popped to my mind . . . . The
phrase that runs through the industry is that Microsoft performed a
cashectomy on Borland, that it lowered the prices of its product,
which caused severe disruptions in Borland's cash flow and also in
the stock price, which caused Borland to take a series of significant
steps, including disposing of some products which historically had
been significant competitors to Microsoft products . . . . The
example he did use of Adobe . . . where Microsoft had had a very
significant effort and success in changing the destiny of a company
. . . . So, there was very clear message that they wanted us to
leave the space, and that there would be consequences if we
didn't. Jacobsen Dep., 1/13/99, at 161:20 - 163:1.
- GX 1368 (quoted above).
- Muglia Supp. Dir. ¶ 26 (Muglia denies mentioning PeopleSoft, but
admits citing SAP, another software company that builds on top of,
but does not compete with, Windows, as a model for what
Microsoft expected from RealNetworks).
84.4. Microsoft induced RealNetworks to enter into a contract that
restricted its ability to work with other potential platform-level competitors to Microsoft,
Sun and Netscape.
- See infra Part V.F.2; ¶ 286.
84.5. As with Netscape and Apple, RealNetworks's product experienced new
technical problems working with Windows when RealNetworks declined to abandon the core
streaming business.
- Mr. Jacobsen testified that Microsoft's Windows Media Player took over
MIME types without giving users a choice, overwrote Real Networks
software without giving users a choice, essentially depriving the user of
the use of the $29.95 player that had previously been installed. Some but
not all of these problems were patched following Mr. Glaser's testimony
before the United States Senate. Jacobsen Dep.,1/13/99, at 163:3 - 167:21,
173:8 - 174:8.
V. Microsoft Engaged In A Predatory Campaign To Crush The Browser Threat
To Its Operating System Monopoly
85. With the browser threat to its operating system monopoly still robust after its
failure to divide markets with Netscape, Microsoft embarked on a calculated campaign
to protect its monopoly by thwarting the widespread adoption of rival browser products.
That campaign had as its object increasing Microsoft's share of the browser market and
sufficiently weakening Netscape and other rivals to ensure that non-Microsoft browsers
(or other middleware) did not become an important platform to which developers wrote
applications that ran on PCs.
A. After Netscape refused Microsoft's offer to divide the browser
market, Microsoft embarked on a predatory campaign to eliminate
the browser threat
3. Microsoft made obtaining browser share a central corporate
objective
86. Failure to divide the browser market with Netscape frustrated Microsoft's
objective of eliminating the threat that widely-used non-Microsoft browsers, in particular
the Netscape browser, posed to Microsoft's operating system monopoly.
87. Microsoft nonetheless recognized that it could blunt the browser threat by
weakening rivals and gaining browser market share.
- In an April 6, 1995 internal memorandum entitled "Netscape as Netware,"
Paul Maritz explained the threat posed by Netscape if Netscape enjoyed
high market share. Maritz stated that if Netscape Navigator gained
"significant market share," then "content providers see more to be gained
in exploiting unique features of Netscape clients than in trying to be
'generic' across all clients." Maritz explained: "This feedback loop drives
Netscape market share higher (as content providers encourage its use) to
the point where Netscape can go 'proprietary'. . . Eventually they become
a real 'platform,' and they are eating 'per PC' revenue that would
otherwise go to the OS or to the Apps." GX 498, at MS98 0168614.
- In an April 4, 1996 internal Microsoft memorandum, entitled "FY97
Planning Memo 'Winning the Internet platform battle'," Brad Chase wrote,
"Go for maximum browser share. Why should you care? This is a no
revenue product, but you should worry about your browser share, as
much as BillG because: we will loose [sic] the Internet platform battle if we
do not have a significant user installed base. The industry would simply
ignore our standards. Few would write Windows apps without the
Windows user base. -- at your level, if you let your customers deploy
Netscape Navigator, you loose [sic] the leadership on the desktop." GX
39, at MS6 5005720 (emphasis in original).
88. Microsoft understood the importance of increasing browser market share
and made gaining browser market share a central corporate goal. This "very important"
and "#1" goal to increase browser market share -- articulated among others by Bill
Gates and his senior executives -- was a central focus of Microsoft's corporate strategy
from 1995 to the present day.
- In his "Netscape as Netware" memorandum, Paul Maritz wrote: "I think the
most important thing we can do is to 'not lose control' of the Web client.
By controlling the client, you also control the servers. We should not allow
any one Web client to get to high volume. This means (i) not letting a
vacuum open up, and (ii) ensuring that we get broad distribution for our
Web client." GX 498, at MS98 0168614.
- Bill Gates wrote, in January 1996, that "Winning Internet browser share is
a very, very important goal for us." GX 295.
- Paul Maritz, in June 1996, repeated that "Without browser share,
everything is hard. So job # 1 is browser share." GX 42, at MS6
6010346.
- Microsoft executive Carl Stork wrote in September 1996 that "Browser
share is Job 1 at this company, and OSR2 is the vehicle to get IE3 on
these machines." GX 44.
- In an internal March 25, 1997 Microsoft e-mail exchange between Stork,
Megan Bliss, and others, Bliss wrote: "I thought our #1 strategic
imperative was to get IE share (they've been stalled and their best hope is
tying tight to Windows, esp. on OEM machines). That is, unless I've
woken up in an alternative state and now work for Netscape." GX 56, at
TXAG 0009634.
- An April 1998 marketing plan for IE5 lists as an
Specifically, the document read:
- redacted - GX 432 (sealed).
4. Microsoft embarked on a predatory and anticompetitive
course of conduct designed to gain browser share
89. To achieve its goal of weakening browser rivals and protecting its operating
system monopoly, Microsoft set out on a campaign to gain browser market share
through predatory and anticompetitive means. Microsoft's practices included giving its
browser away "forever free," coercing third parties not to deal with or support
competitive browsers, and to carry its browser in ways that disadvantaged rivals, and
paying other browser suppliers' customers and distributors not to carry other browsers
or to do so only on disadvantageous terms. Among other things:
89.1. Microsoft, without legitimate justification and for the purpose of
blunting the browser threat, tied its Internet Explorer browser to Windows, refusing to
offer an unbundled option despite the plain existence of separate demand for browsers
and operating systems. See infra Part V.B.
89.2. Microsoft imposed anticompetitive restrictions on OEMs' ability to
modify the Windows desktop and start-up screens, even though doing so reduced the
value of Windows. See infra Part V.C.1.
89.3. Microsoft bestowed favors on OEMs that assisted Microsoft in its
exclusionary strategy, punished OEMs that did not, and contractually restricted OEMs
from removing the browser. See infra Part V.C.2.
89.4. Microsoft entered into exclusionary contracts with ISPs and OLSs
accounting for the majority of consumer Internet access in the United States for the
purpose and with the effect of limiting competitor browser usage and raising rivals'
costs. See infra Part V.D.
89.5. Microsoft entered into exclusionary agreements with ICPs that
included restrictive provisions that cannot be explained except as components of a
predatory campaign designed to exclude browser rivals and protect Microsoft's
operating system monopoly. See infra Part V.E.
89.6. Microsoft entered into an exclusionary agreement with Apple
designed to limit competitive browser usage, raise rivals' costs, and increase usage of
Internet Explorer on the MacIntosh operating system. See infra Part V.F.
89.7. Microsoft, after studying Netscape's sources of revenues and for
the purpose of cutting off Netscape's "air supply," invested hundreds of millions of
dollars in developing, promoting, and distributing its Internet Explorer browser, even
though it planned to and did make the browser "forever free" and did not collect
browser-related ancillary revenues. See infra Part V.G.
90. Microsoft had no plan or expectation that these acts would be profitable or
make business sense, except by preventing rival browsers from developing into a rival
development platform and thereby preserving Microsoft's operating system monopoly.
- In an internal memorandum Bill Gates sent on May 19, 1996, to his senior
executives, he outlined a strategy against Netscape in the "browser war."
Gates indicated that Microsoft would price many of Microsoft's Internet
products "free." After outlining his strategy, Gates concluded that: "At
some point financial minded analysts will begin to consider how much of a
revenue stream Netscape will be able to generate." GX 41, at MS6
6012954-56.
- Paul Maritz emphasized that Microsoft's goal of increasing browser share
was more important than earning revenue from the browser. In a July 11,
1997 internal Microsoft email, Maritz wrote: "There is talk about how we
get more $'s from the 1000+ people we have working on browser related
stuff, but I have not lost sight of the fact that Browser Share is still an
overwhelming objective. You may notice that I have kept IE marketing
spend at very high level through FY'98. and resisted pressure to reduce
this or switch it to other products. I also said 'no' on the proposal to
charge separately for the Shell." GX 112; Maritz, 1/26/99pm, at 18:25 -
20:8, 21:22 - 22:22 (Maritz testified that he rejected a proposal about
splitting IE4 and charging a price for one of the pieces because "it would
interfere with the objective" he "had which was to get more people using
Internet Explorer.").
- See also infra Part V.G.
3. Microsoft's efforts to pressure Intel to stop developing or
supporting platform-level software illustrate Microsoft's
predatory intent and tactics
91. In various meetings in 1995, Microsoft (i) forced Intel not to support rival
platform-level software and (ii) candidly articulated its predatory plan to use its
monopoly power and other predatory means to thwart the browser threat. Microsoft's
use of its monopoly power to pressure Intel not to support Netscape or offer Intel's own
platform-level technologies well illustrates both Microsoft's predatory intent and the
anticompetitive practices it employed to blunt threats to its operating system.
a. In an August 1995 meeting, Microsoft pressured Intel
into not resuming platform-level software and not
supporting Netscape and Java
91.1. As will be detailed below (Part VI.B.2.), Intel had developed
software that Microsoft viewed as a platform-level software that might someday
compete with Windows. In response to this potential threat, Microsoft engaged in a
campaign to force Intel not to ship its software, then known as NSP. This effort
culminated in an August 2, 1995, meeting, in which Bill Gates -- in a blunt use of
Microsoft's monopoly power -- threatened to withdraw support for Intel's
microprocessors unless Intel dropped support for platform-level software efforts and
cooperated in Microsoft's Internet strategies.
- In May 1995, Microsoft vice-presidents Paul Maritz, Brad
Silverberg, and Microsoft's Carl Stork, met with Intel executives to
discuss Intel's NSP program. The Microsoft executives
complained that Intel was shifting the software boundary with its
NSP project by writing software that Microsoft considered to be part
of its operating system space. GX 275; McGeady, 11/9/98pm, at
23:3 - 26:23. In Microsoft's view, NSP made Intel a competitor in
Microsoft's operating system space. GX 275; McGeady,
11/9/98pm, at 26:25 - 27:11.
- Bill Gates explained to Intel's Andy Grove that Intel's attempts to
compete with, rather than follow, Microsoft in software were
unacceptable: "The problem we have is that we have to sort of
choose in software related issues which company will lead and
which will follow. In chips its very clear. In software you have a
group that won't allow us to lead and has all the prestide (sic) and
profits of Intel to drive them forward." GX 277.
- In an internal July 7, 1995 Microsoft e-mail, Gates reported that he
had tried to convince Grove "to basically not ship NSP." GX 278;
Maritz Dir. ¶ 320. Gates predicted that Intel would exert less
pressure to ship NSP in 1995 but that it "will take a major effort for
us to convince them to back off from this." Gates further elaborated
that Microsoft is the "software company here and we will not have
any kind of equal relationship with Intel on software." GX 278.
- In an August 28, 1995 memorandum summarizing the meeting,
Steven McGeady wrote: "On August 2 1995, in a meeting of Intel
and MS executives, Bill Gates told Intel CEO Andy Grove to shut
down the Intel architecture labs." GX 280. According to McGeady,
Gates was upset that Intel was "making investments in software of
any sort" because "he felt that anything" Intel "did in software was
competitive." McGeady 11/9/98pm, at 10:10 - 14:3.
- McGeady explained the reason Intel abandoned its NSP
development: "Intel did fail to introduce NSP into the marketplace
because, as a primary cause, Microsoft in particular Bill Gates, told
Andy Grove that MS did not want NSP in the marketplace," and
because "Microsoft helped . . . in our business interests by
threatening to withhold support for other microprocessors in the
meantime." McGeady, 11/10/98pm, at 81:6-23.
- See also infra Part VI.B.2; ¶¶ 347-350.
91.2. During this meeting on August 2, 1995, Microsoft told Intel not only
that it wanted Intel to stop developing platform-level software, but also that it wanted
Intel not to support other platform-level software that ran on Windows, especially
Netscape's browser and Sun's Java technologies, in any way that could contribute to
their development as a rival platform.
- Gates made clear to Intel executives on August 2, 1995, that
"Microsoft would not support" Intel's "next processor offerings if we
did not get alignment between Intel and Microsoft on platform
issues" and on communications issues, i.e., Internet issues.
McGeady, 11/9/98pm, at 14:14 - 15:4; GX 279 ("Gates would not
agree to let processors/OSs programs to progress unencumbered
by platform communications program issues.").
- In addition to setting limits on Intel's software efforts, Gates raised
"Internet issues." GX 279, at MS CID 00078. Gates cautioned that
Microsoft was "very sensitive to what Intel might do on the client
side. Example: JAVA, a show stopper." Id. (By "client," Gates
meant "browser." Maritz, 1/27/99am, at 27:12-21).
- McGeady explained that Gates also told Intel that it should
"concentrate 70% of" its "resources on working with Microsoft's
technology and 3- percent on any third parties'" technologies such
as Netscape. McGeady, 11/12/98pm, 19:5 - 20:9. Gates further
explained, with regard to this "30/70 use of 3rd party technologies,"
that "Intel using Netscape in Windows environment is not a
problem (provided we [Intel] do not set up the 'positive feedback
loop' for Netscape that allows it to grow to de facto std.)." GX 279,
at MS CID 00078.
- As McGeady testified, Gates permitted Intel's internal use of
Netscape browser as a stand-alone application on Windows, but
"he didn't want" Intel "to do anything that would encourage
developers to begin to move to Netscape, thereby increasing the
value of that platform to create this positive feedback loop, this
increasing-returns kind of situation with Netscape." McGeady,
11/12/98pm, 19:5 - 20:9. McGeady also testified that "it was very
clear that Bill did not want us doing any development or technology
work with Netscape that would improve . . . the viability of Netscape
Navigator in the marketplace. They wanted us to, if we absolutely
had to, use it just as a standalone product. He would acquiesce to
that, but he didn't want us doing any technical work with them."
McGeady 11/12/98pm, 20:10-20; GX 279 (Whittier's minutes) ("BG:
Supporting certain third party deals will be problem . . . we need to
consider in the context of their (pervasive) internet program to
assure we are not unknowingly stepping on one of their key
strategies!").
b. In subsequent meetings in the Fall of 1995, Microsoft
explained to Intel that its strategy would be to kill
Netscape and control Internet standards
91.3. After Microsoft used its monopoly power to prevent Intel from
developing its own platform-level software, Microsoft continued to pressure Intel not to
support Netscape's browser and bluntly described to Intel its predatory scheme and
objective. At a meeting held on November 9, 1995, Microsoft executive Paul Maritz met
with Intel executives and explained to Intel, in explicit terms, that Microsoft's strategy
was to kill Netscape and control Internet standards.
91.3.1. During this meeting, Paul Maritz told Intel that Microsoft
would "cut off Netscape's air supply" -- that, in other words, by "giving away free
browsers," Microsoft would choke off Netscape's sources of revenue and retard its
ability to invest in developing its technology.
- McGeady testified that Maritz told Intel that Microsoft
planned to "cut off Netscape's air supply," or in other words,
"by giving away free browsers, Microsoft was going to keep
Netscape from getting off the ground." McGeady,
11/9/98pm, at 53:6-23; GX 1640.
- McGeady testified that Maritz explained Microsoft's strategy
"was, first of all, to give the browser away from (sic) free,
keep Netscape from getting any revenue from it, and that
was their specific cut off of air supply. In other words, don't
allow them to have any revenue to continue paying their
engineers to build new products." McGeady, 11/9/98pm, at
54:21 - 55:16.
- Maritz's inconsistent testimony to the contrary lacks
credibility. When asked whether he told Intel that Microsoft's
plan was to "cut off Netscape's air supply," Maritz testified at
his deposition "it's possible, but I just don't recall it," Maritz,
1/25/99 pm, 74:18-75:16, but at trial, Maritz testified
unequivocally and inconsistently that he "never said in the
presence of Intel personnel or otherwise, that Microsoft
would cut off Netscape's air supply or words to that effect."
Maritz, 1/25/99pm, at 72:21 - 73:16; Maritz Dir. ¶ 333. And
despite claiming that he did not say "words to the effect" of
Microsoft planned to "cut off Netscape's air supply," he
testified that he "may have pointed out on occasion that the
base Internet technologies, the browser and the Internet
server, we were going to incorporate into Windows and not
charge for it separately." Maritz, 1/26/99am, 7:22 - 8:1.
Maritz also conceded that he told Intel "on many occasions
that it was" Microsoft's "strategy to integrate Internet support
into" its "operating system and not charge for it separately."
Maritz, 1/26/99am, 7:22 - 8:8.
91.3.2. Paul Maritz also explained to Intel representatives that
Microsoft's response to the browser threat was to "embrace, extend, extinguish"; in
other words, Microsoft planned to "embrace" existing Internet standards, "extend" them
in incompatible ways, and thereby "extinguish" competitors.
- McGeady testified that Maritz told Intel that Microsoft's
strategy was to "embrace, extend, extinguish." McGeady,
11/9/98pm, at 53:17 - 54:8; McGeady, 11/10/98 am, at
21:22 - 23:19; GX 564.
- McGeady testified that Microsoft was going to take Internet
standards, like HTML, "and extend it to the point where it
was incompatible with the Netscape browser and encourage
people to develop to their version of HTML so that pages
couldn't be read with Netscape's browser." McGeady,
11/9/98pm, at 55:7-14.
- Russell Barck, an Intel executive, testified at his deposition
that "in relation to Netscape, . . . Maritz . . . said the term
'embrace and smother' with respect to a strategy with
respect to Netscape." Maritz, 1/26/99 am, 55:19 - 57:1.
- Rob Sullivan testified at his deposition that Maritz said the
phrase "embrace and smother." Maritz, 1/26/9am, 57:2-11.
When asked about his understanding of the meaning of the
embrace and smother concept, Sullivan testified that he
"understood that concept to mean that Microsoft intended to
deprive Netscape of revenue and viability." Microsoft would
achieve this "by giving away their products, by embracing
the Internet standards and extending them in a way that
favored the Windows platform." Maritz, 1/26/99am, 58:16 -
59:8.
91.3.3. Paul Maritz also explained to Intel representatives that
another aspect of Microsoft's strategy to combat Netscape was to create dependencies
between the operating system and the browser.
- McGeady testified that, based on the meeting with Maritz,
Microsoft planned "to create some various levels of
dependencies between the operating system and the
browser that would differentially advantage their browser."
McGeady, 11/9/98pm, at 54:15 - 55:14; GX 564.
91.4. Microsoft continued to monitor to ensure that Intel did not voice
public support for Netscape, even to the point of regulating Intel's internal browser use.
- In an internal June 6, 1996 email, Bill Gates reported to other
executives that he spoke with Intel's CEO, Andy Grove, and Bill
Gates reported: "I said it was important that Intel NOT ever publicly
say they are standardizing on Netscape browsers." McGeady,
11/9/98pm, 49:3 - 50:2; GX 289.
92. Microsoft's efforts to ensure that Intel not support Netscape, and its blunt
warning that it would "cut off" Netscape's "air supply," were neither isolated events nor
normal competitive banter or "locker room talk." Rather, they were part of a calculated,
multifaceted predatory scheme, the details of which are set forth below.
B. Microsoft tied its Internet Explorer browser to Windows 95 and
Windows 98 in order to impede browser rivals such as Netscape,
and for no legitimate purpose
93. A central part of Microsoft's predatory campaign to prevent Netscape's
browser from developing into a platform that could erode the applications barrier to
entry was Microsoft's tying of its Internet Explorer browser to Windows 95 and Windows
98 and its refusal to offer, or to permit OEMs to offer, an unbundled option.
93.1. Internet browsers and personal computer operating systems are
separate products. Consumers view browsers and operating systems as separate
products and demand one without the other. In response to that separate demand,
Microsoft and other software firms have found it efficient to promote and distribute
browsers and operating systems separately. See infra Part V.B.1; ¶¶ 96-119.
93.2. Despite the existence of this separate demand for browsers and
operating systems, Microsoft tied its browser to its Windows operating system, and
refused to offer an unbundled option, for the purpose of hindering the development of
Netscape and other browsers. See infra Part V.B.2; ¶¶ 120-149.
93.2.1. Microsoft tied Internet Explorer 1 and 2 to Windows 95 by
requiring OEMs to obtain Internet Explorer in order to obtain Windows 95 and
prohibited OEMs from removing Internet Explorer.
93.2.2. Subsequently, fearing that its merely contractual tie was
not sufficient to eliminate the threat that Netscape's browser posed to its operating
system monopoly, Microsoft changed its product design in Internet Explorer 3 and 4 to
commingle browser and operating system code. Still, recognizing the desire of users to
have the Windows 95 operating system without Internet Explorer, Microsoft designed
and advertised an easy means for users to remove the browser. Microsoft, however,
refused to provide a version of Windows 95 from which the browser had been removed
or to permit OEMs to remove the browser from the PCs they sold.
93.2.3. Microsoft designed Windows 98 to further implement the
tying arrangement by eliminating the end user's ready ability to "uninstall" Internet
Explorer and by interfering with his ability to choose a different default browser.
93.3. There is no sound justification for Microsoft's tying Internet Explorer
to Windows. See infra Part V.B.3; ¶¶ 150-167.
93.4. Microsoft's tying arrangement and contractual prohibition on
unbundling inflicted significant harm on competition and consumers. See infra Part
V.B.4; ¶¶ 168-176.
1. Internet Explorer and Windows operating systems are
separate products
94. Internet browsers and operating systems, including Internet Explorer and
Windows, are separate products that are sold in separate product markets. There is
separate demand for both browsers and operating systems that is efficient for suppliers
to meet.
a. Browsers and operating systems are universally
recognized by industry participants to be separate
products
(1) An Internet browser supplies web browsing
95. An Internet web browser ("Internet browser") is a software program that
enables its user to view, retrieve, and manipulate content located on the Internet's
World Wide Web and other networks (hereinafter "web browsing").
- Microsoft's own dictionary defines a "web browser" as a "client application
that enables a user to view HTML documents on the World Wide Web,
another network, or the user's computer; follow the hyperlinks among
them; and transfer files." Microsoft Press, Computer Dictionary (3d
ed. 1997), at 505 (GX 1050).
- Professor Franklin Fisher defined a browser as "the application that
permits users to access and browse the world wide web or, for that
matter, other networks." Fisher, 1/6/99am, at 5:3-5.
- Dr. Warren-Boulton defined a browser as "software that enables computer
users to navigate and view content on the World Wide Web." Warren-Boulton Dir. ¶ 68.
(2) Industry participants view a browser as an
application, and not as part of an operating
system
96. Industry participants -- including consumers, other operating system
vendors, ISVs, corporate information technology officers, academic computer scientists,
and the industry press (including Microsoft's own computing dictionary) -- universally
regard web browsers as application programs separate from the underlying operating
system.
96.1. Other operating system vendors, even those that bundle a browser
or multiple browsers with their operating system products, have always considered the
browser to be a separate application.
- Apple Computer's Avadis Tevanian testified: "The fact that Internet
Explorer and Navigator are bundled with the Mac OS does not
make them part of the operating system. The Mac OS operating
system will continue to function if either or both of these browsers
are removed . . . [and] we permit value added resellers the
flexibility . . . to remove browsers or other applications . . . ."
Tevanian Dir. ¶ 26; see also Tevanian Dir. ¶¶ 8-9 (explaining the
difference between operating systems and applications).
- John Soyring from IBM testified that "IBM has not found it
necessary technically to integrate the browser with the operating
system -- the browser worked well running on the operating system
like any application." Soyring Dir. ¶ 18.
- Sun officials consistently describe Sun's "HotJava" browser as an
"application that performs web-browsing functionality." Sasaki Dep.
(played 12/16/98pm), at 22:5-18.
- Brian Croll testified that the browser that Sun bundles with the
Solaris operating system environment is "an application that runs
on the environment. That's basically on top of the CDE." Croll
Dep. (played 12/15/98pm), at 38:12-14. Croll later defined an
"application" as "a piece of software that sits on top of the operating
system and that people use and performs a function that they are
looking for." Id. at 66:11-16.
- Ron Rasmussen from The Santa Cruz Operation testified that SCO
"bundles" Netscape Navigator with its OpenServer and Unixware
products (Rasmussen Dep. (played 12/15/98am), at 54:10 - 56:25),
but that "our view is that the browser is an application."
Rasmussen Dep. (played 12/15/98am), at 64:20. Rasmussen also
testified that "when SCO says 'we bundle a feature,' it means its a
feature which is not part of the core base operating system
functionality. It means that it's something that the user can choose
to install or remove, and the operating system, whose primary
function it is to serve applications, will still function properly."
Rasmussen Dep. (played 12/15/98am), at 55:14-19.
96.2. Consumers also regard browsers as applications rather than as
parts of any operating system product.
- Jon Kies, the Senior Product Manager at Packard Bell/NEC,
testified that "browsers are considered by most of our
customers as a third party application." Kies Dep. (played
12/16/98am), at 7:19-20.
- Glenn Weadock concluded from his research and interviews
that corporate information managers "typically consider
browser software as application software, like email or word
processing, not as an operating system or as part of a
particular operating system." Weadock Dir. ¶ 22 (collecting
illustrative statements by corporate managers). Weadock
further testified: "No corporate PC manager, in fact no one
outside of the Microsoft organization, has ever described a
Web browser to me as operating system software or as part
of Windows 95 or any other operating system." Weadock
Dir. ¶ 22 (emphasis in original).
- Boeing's Scott Vesey testified: "From my perspective, I
would view them as software applications because they are
tools that are used to interpret data rather than what I would
normally view as the operating system, which is the
components of software that are used to directly manipulate
the hardware that forms that PC. The applications are used
to interpret or parse data." Vesey Dep., 1/13/99, at 284:15 -
285:9.
- Netscape's Jim Barksdale testified: "Consumers have had
no problem appreciating that browsers are separate
products," and "still demand Netscape Navigator and
Netscape Communicator separately from any operating
system products." Barksdale Dir. ¶ 90.
96.3. When the industry press or prospective customers evaluate the
features and quality of Internet Explorer, they invariably compare it to Netscape's Navigator browser
application, and not to any operating system.
- Barksdale testified that "the industry as a whole recognizes
browsers as separate products from operating systems.
Browser market share is tracked (separately from operating
system market share) by many third party organizations,
such as IDC and DataQuest. The 'browser wars,' referring
to the commercial battle between Netscape Navigator and
Microsoft Internet Explorer, are frequently reported on in the
press. I have seen many product reviews comparing
Navigator to Internet Explorer; I have never seen a product
review comparing Navigator to any Windows operating
system." Barksdale Dir. ¶ 90.
- An internal Gateway presentation from March 1997 includes
a detailed "Basic Feature Comparison" between "Netscape
and Microsoft Browser Products." GX 357 (sealed).
- Many press reviews of browsers directly and explicitly
compare Internet Explorer to Netscape Navigator and
Communicator and talk about them as applications
independent of any particular operating system. See, e.g.,
GX 1262 (1996 ZDNet review); GX 1272 (1997 CMPnet
review); GX 1274 (1997 PC Week Online review); GX 1285
(1997 ComputerShopper.com review); GX 1287 (1998 PC
Magazine Online review); GX 1288 (1998 ZDNet News
review).
96.4. Experts in software design describe browsers as applications, and
not as parts of any operating system.
- James Gosling of Sun Microsystems testified that "the
browser is best understood as a software application, not as
a part of a computer's operating system. This is true both as
a matter of function and as a matter of software design. As
a matter of function, browsers perform tasks for the end user
that relate to obtaining and displaying content on the
Internet or other networks. Users may wish to choose a
particular Internet browser that best fits their needs, or if
they have no need to 'browse the Web,' perhaps no browser
at all. Technically, browsers are treated by the computer
like any other application. In virtually every operating
system with which I am familiar, the particular files that
enable browsing are loaded into memory and used in
exactly the same way as other software applications. Even
in Windows 98, where Microsoft apparently loads some
browser-related files into memory even when the user may
never need that functionality, these files are loaded in the
same way as other software applications. In essence,
Microsoft simply shifts the time required to load the browser
code from when it is first needed by the user to every time
the computer boots up." Gosling Dir. ¶¶ 38-39.
- Gosling also testified: "A browser is an application that, like
a JVM, runs on the operating system installed on a user's
computer. It permits the user to access information encoded
in hypertext markup language, or HTML, and other types of
content found on the Internet or other networks, and to
navigate around these networks." Gosling Dir. ¶ 34;
Gosling, 12/9/98pm, at 41:20-23.
- Professor Felten testified that "Internet Explorer is part of the
distribution which Microsoft sells under the name Windows
98. However, their Internet Explorer is an application which
can be separated from Windows 98." Felten, 12/14/98am,
at 30:21-24.
- Marc Andreessen testified that "I can't say that I ever
thought that a browser was necessarily separate from
everything. But it would certainly be fair to say that I think
that the browser has been separate from an operating
system, for example." Andreessen Dep., 7/15/98, at 122:20
- 123:7 (DX 2555).
- Even Dr. Michael Dertouzos, Director of the Laboratory for
Computer Science at M.I.T. and formerly on Microsoft's
witness list, agreed: "Historically and today, it is the case
that browsers are treated as applications." Dertouzos Dep.,
1/13/99, at 414:2-4.
(3) In its ordinary commercial conduct, Microsoft
treats Internet Explorer as a separate product
97. Microsoft similarly treats Internet Explorer as a product separate from its
Windows line of operating system products.
(a) Microsoft promotes Internet Explorer as a
product, positions it in competition with
other Internet browsers, and tracks its
market share relative to those of other
browsers
97.1. Microsoft distributes Internet Explorer separately from Windows in a
variety of different channels, including retail sales, service kits for ISVs, free downloads
over the Internet, and with other products produced both by Microsoft and third-party
ISVs.
- On cross-examination, Microsoft's Cameron Myhrvold conceded
that Internet Explorer is distributed separately from Windows in
"many, many ways." Myhrvold, 2/9/99pm, at 37:7 - 38:7.
- An internal Microsoft "Timeline Summary"
- redacted -
GX 669 (sealed).
- When asked whether Microsoft released "something called Internet
Explorer 3 separately from OSR2 around the time that OSR2 was
released," Carl Stork answered that Microsoft "released it on the
Web and I believe we released it in some kind of a retail Internet
starter kit type of product as well." Stork Dep., 8/11/98, at 38:18-23
(DX 2595).
97.2. From the introduction of Internet Explorer 1.0 in mid-1995 to the
present day, Microsoft has always promoted and marketed Internet Explorer as a
product separate from Windows.
- Soyring testified that "Microsoft itself has at certain times treated
Internet Explorer as separate from Windows. In the fall of 1997,
Microsoft held a major public relations event to introduce Internet
Explorer 4, independent of Microsoft's promotion of Windows."
Soyring Dir. ¶ 19.
- In describing Microsoft's marketing plans for Internet Explorer in
August of 1995, Yusuf Mehdi wrote that Microsoft would "treat it as
a distinct product in the sense of setting up clear news, reviews,
and feature coverage objectives." GX 153.
97.3. Microsoft's internal strategy documents dealing with Internet
Explorer consistently described Netscape Navigator (and not any of Microsoft's
traditional operating system competitors) as Internet Explorer's "primary competitor"
and identified gaining "browser share" vis-a-vis Netscape as the primary objective for
Internet Explorer marketing efforts.
- An "Internet Product Management Strategy" in November 1995
identifies Netscape as the "primary competitor" and lists as its
objective to "Make the IE the people's choice of Web browsers via
aggressive distribution and promotion." GX 673, at MS6 6005881.
- In notes from an offsite meeting among the Internet Explorer
project team in November 1997, Microsoft's Chris Jones describes
the role of the Internet Explorer team as "gain browser share." GX
364, at MS7 004722.
- In December 1996, Microsoft's David Cole wrote: "There is still the
message here that Internet Explorer is still a browser, where Nav is
groupware. No credit for Netmeeting, mail, news, etc. We need to
change that perception." Microsoft's Yusuf Mehdi responded that
"it is probably a good example though of the need to have a single
group taking on communicator else we wil never get the full
message across. I have thought more about our conversation and
more firmly believe that you need a single group and product that
you market against communicator. It makes sense to me that this
use the IE brand and team because of equity, experience, and
relevancy in product, team, and marketing. The group would
market IE4 which includes: Active Desktop, Browser, Mail, News,
Netmeeting, FrontPad, Admin Kit, etc." GX 658, at MS6 6010327.
- In June 1997, Chris Jones sent a memo to Bill Gates entitled "How
to get to 30% share in 12 months." The memo contains a lengthy
discussion of how Microsoft should design and market Internet
Explorer to take market share away from Netscape. GX 334, at
MS98 0104679.
97.4. Internal Microsoft assessments of Internet Explorer's success
invariably compared its features, performance, and market penetration to those of
Netscape Navigator.
- A March 1997 Microsoft "Competitive Guide" compared the
features of Internet Explorer 4.0 against those of Netscape
Communicator. GX 477, at MS7 004179.
- Chris Jones' notes from a November 1997 Internet Explorer team
meeting claims that "[w]e have won every head to head review
against Netscape." GX 364, at MS7 004719.
97.5. In fact, the contemporaneous documents show that Microsoft
regularly tracked Internet Explorer's market share relative to that of Netscape
Navigator.
- A January 1998 "IE International Business Review" slide
presentation breaks down 1997 browser shares in both domestic
and international markets. GX 815, at MS98 0202889.
- An October 1996 e-mail from Yusef Mehdi to Paul Maritz and
others reports current browser share as measured by weekly call
downs, share at random web sites, and Internet Explorer
downloads. GX 344.
- See also e.g., GX 713 (April 1998 Mehdi email comparing Internet
Explorer and Navigator share and noting that "48 is a big number
and implies that we have caught Netscape"); GX 495 (comparison
of Internet Explorer and Navigator share); GX 700 (same); GX 708
(same); GX 713 (same); GX 714 (same); GX 714A (same); GX 716
(same).
(b) Microsoft treated Internet Explorer and
Windows separately until the issue arose in
litigation
(1) Before litigation, Microsoft called
Internet Explorer a browser in its
ordinary commercial conduct
98. In the ordinary course of its business, Microsoft has frequently described
Internet Explorer as a browser application rather than a part of the operating system.
- In a July 1995 memo to the OEMs, Microsoft described Internet Explorer
as a "32-bit Windows 95 World Wide Web browser and graphical FTP
utility." GX 36.
- In December 1995, Brad Silverberg wrote to Bill Gates and Paul Maritz
that Internet Explorer 3.0 "is a standalone web browser that runs on
Win95." GX 37.
- See also GX 141 (Windows 95 would contain "[a]ll the necessary
plumbing" to access the Internet, including a TCP/IP stack and support for
the PPP and SLIP protocols, and that it would "[s]upport[] popular third
party Internet applications, such as Mosaic").
99. Microsoft also entered into extensive agreements with PC OEMs, ISVs,
ISPs, and ICPs regarding the placement and promotion of Internet Explorer that were
separate from any agreements regarding licensing terms for Windows and that
invariably referred to Internet Explorer as a "browser," not as a part of the operating
system.
- A September 1996 amendment to a May 1996 licensing agreement with
Compaq required Compaq to "Offer the Microsoft Internet Explorer as the
preferred worldwide web browser for users of any COMPAQ Internet
Product(s) listed in Exhibit B [Support Software CD for Compaq Desktop,
Portable and Workstation Products and Compaq Resource Kit for
Microsoft Windows NT]." GX 1130, at MSV 0005706 (Ex. D, Amd. 1).
- A July 1996 license and distribution agreement with Compaq required
Compaq to "Offer the Microsoft Internet Explorer as the preferred
worldwide web browser for users of the Support Software CD for Compaq
Desktop Products." GX 1137, at MSV 005747.
- The Internet Sign Up Wizard Referral and Microsoft Internet Explorer
License and Distribution Agreement with AT&T, dated July 23, 1996,
- redacted -
GX 1212, at MS6 5000435 (Ex. B, §6) (sealed).
- The August 1995 Internet-Sign Up Wizard Referral Agreement with
CompuServe
- redacted - GX 1144, at MS6 5001138 (Exhibit B,
Section 5) (sealed).
- The December 1995 Internet Explorer Source License & Distribution
Agreement with CompuServe required that Compuserve to "ship the
Internet Explorer as its primary World Wide Web browser software client
for Windows 95 . . . ." GX 1125, at MS6 5000091.
- An August 1996 Internet-Sign Up Wizard Referral and Microsoft Internet
Explorer License and Distribution Agreement with Earthlink
- redacted - GX 1141, at MS6 5000015 (Exhibit
C, § 6) (sealed).
- A May 1996 Internet Explorer Addendum to Strategic Relationship
Framework Agreement with MCI
- redacted - GX 1132, at MS6 6008292 (sealed).
- A September 1996 Promotion & Distribution Agreement with Prodigy
- redacted - GX 1148, at MS6 50010000 (Section 3.1) (sealed).
- Numerous Memoranda of Understanding that Microsoft entered with
major OEMs in July and August of 1997 provided significant inducements
for those OEMs to promote and distribute Microsoft's upcoming Internet
Explorer 4 browser, which initially was offered and distributed wholly
separate from any operating system release. See, e.g., GX 163 (under
seal) (8/29/97 "Memorandum of Understanding ("MOU") re: Internet
Explorer 4.0,"- redacted -
GX 1166 (under seal) (7/21/97 MOU with DEC,
similar language); GX 1168 (under seal) (8/8/97 MOU with Packard-Bell,
similar language); GX 1171 (under seal) (8/20/97 MOU with Dell, similar
language).
- See also e.g., GX 856, at MS98 0100300 (Section 2.3(d)) (July 1997
Disney Active Desktop agreement); GX 1159, at TM 000057 (June 1997
Hollywood Online Active Desktop agreement); GX 1157, at MS98
0100570 (Section 2.2) (June 1997 Intuit agreement); GX 1153, at MS98
0100811 (Section 2.1(a)) (December 1996 Pointcast agreement); GX 855,
at WD 0004 (Section 2.3) (July 1997 Wired Digital Active Desktop
agreement); GX 1166 (July 1997 IE4 launch event agreement).
100. Similar references to Internet Explorer as a "browser" appear in Microsoft's
internal and external correspondence right up to the present day.
- Microsoft describes Internet Explorer 5.0 as a "smaller, faster, more stable
browser." GX 688.
- An Internet Explorer 5 OEM Marketing Review from May 1998 asserts
that "IE has around 50% browser share," and that end users "view both
browsers as parity products." GX 233, at MS98 0125654.
(2) Since litigation began, however,
Microsoft has made a concerted
effort to change its language in order
to aid its legal position
101. Recently, however, in order to support its litigation position that Internet
Explorer and Windows 98 are the same product, Microsoft officials have made a
concerted effort to reposition Internet Explorer and change the terminology used by
Microsoft personnel.
- When Bill Gates was preparing to testify before the Senate in March 1998,
he sent an e-mail to top Microsoft executives suggesting the need for a
"survey . . . where ISVs declare whether they think having the browser in
the operating system the way we are planning to do it makes sense and is
good." GX 377, at MS98 0122148. Nathan Myrhvold responded that the
survey was "a GREAT idea," but that "it is CRUCIAL to make the
statement . . . worded properly. Saying 'put the browser in the OS' is
already a statement that is prejudicial to us. The name 'Browser'
suggests a separate thing. I would NOT phrase the survey or other things
in terms of 'put the browser in the OS.' Instead you need to ask a more
neutral question about how internet technology needs to merge with local
computing. I have been pretty successful in trying this on various
journalists and industry people." GX 377, at MS98 0122146 (emphasis in
original).
- That same month, James Allchin wrote to Yusuf Mehdi that he was "very
concerned over how IE is presented in win98 (and NT5). Even the simple
things like the About Box makes it appear separate. Furthermore, our IE
web site needs a sweep . . . where we ensure it is clear th[at] IE is just a
capability of Windows . . . ." GX 378. Mehdi responded that they were
"making good progress reviewing the language of ie as a feature of
windows with the web team. (we don't refer to it as a product or even
browser, it is browsing software)." GX 378.
b. The recognition that browsers and operating systems
are separate products reflects the marketplace reality
that consumers, for a wide variety of reasons, demand
operating systems and Internet browsers separately
102. Consistent with the universal recognition that browsers and operating
systems are separate products and that different browsers have different characteristics, many
consumers desire to separate their choice of operating system from their choice of
browser.
- Professor Fisher testified: "There is a market for Internet browsers.
Before Microsoft gave away its browser for free, a price for browsers was
determined in the market and the market could have continued to perform
this function. There is substantial demand for browsers that is separate
from the demand for operating systems. Browsers are distributed
separately from the operating system by ISPs and by retailers. There is
demand for operating systems without browsers and for operating
systems with a choice of browsers." Fisher Dir. ¶ 80.
- A survey conducted by Compaq in February 1998 of 283 PC decision
makers at U.S. companies found that "About 80% of companies wipe or
reformat the hard drives of new desktops. . . . The operating system re-installed most often are OSR2 and the retail version of Windows 95.
Large businesses lean more toward the retail version of Windows 95,"
which does not include a browser. GX 1242, at 7.
- Dell's Joseph Kanicki testified: "'Some businesses and government
customers prefer not to have Internet Explorer pre-installed on their
computers because, one, the customer may have its own software or
software standards which do not include the latest version of Internet
Explorer; two, the customer may wish to install a competitive browser
instead of Internet Explorer; or three, the customer may wish to prevent its
employees from accessing or attempting to access the Internet or World
Wide Web.'" Kanicki Dep., 1/13/99, at 332:12 - 333:22 (quoting Kanicki
Decl. ¶ 2).
(1) Some consumers demand browsers and
operating systems separately because different
browsers have different features and they prefer
to obtain a PC containing only the desired
browser
103. Although Netscape Navigator and Internet Explorer deliver roughly
comparable functionality to the end-user, they are not identical. Each program has unique
attributes that may appeal to different audiences, and there is considerable dispute as
to which product's implementation of even their shared features is superior.
- See infra Part V.B.3.c.(1)(a); Part V.B.4.c.
(2) Some consumers, particularly corporate
customers, demand browsers and operating
systems separately because they prefer to
standardize on the same browser across many
PCs and across different operating systems
104. Many corporations use a variety of different hardware and operating
system platforms in various departments throughout their organization.
- Scott Vesey of Boeing testified that "Boeing is a multi-platform company
and that it supports computers that operate with a number of different
operating systems," including Unix, Macintosh, and a variety of Windows
platforms. Vesey Dep., 1/13/99, at 269:13 - 270:24.
- Weadock testified that in his interviews, "some managers (including those
at Informix, Ford, Federal Express, Boeing, and Morgan Stanley/Dean
Witter) have stated that their organizations deploy a variety of operating
systems and hardware platforms, and therefore prefer a browser having
greater cross-platform availability and compatibility." Weadock Dir. ¶ 24a.
105. Such organizations experience significant benefits in the form of increased
productivity and lower training and support costs from standardizing on one browser
across all of their various hardware and operating system platforms.
- Vesey testified that the "various browser standardization or browser
acquisition decisions that Boeing has made" were "made separately from
decisions about acquiring an operating system," and that he would
"prefer" to have "the option of continuing to be able to choose what Web
browser Boeing uses independently from any decisions Boeing might
make about what operating system to use." Vesey Dep. (played
11/17/98am), at 52:12 - 53:14.
- An internal Boeing presentation entitled "ARR 525 Recommendation:
Windows Browser Evaluation" by Scott Vesey in October 1997 identified
"[p]latform support" as a key issue and noted that "Solaris, HP-UX . . . and
AIX are standard UNIX variants within Boeing, and that IE 4.0 for
UNIX/Solaris would not be in production until Q1 98. In contrast,
Communicator 4.0 was available on all platforms." GX 634, at TBC
000537.
- Vesey testified that the "Netscape browser was a product that we could
run across all of the platforms that we had currently installed in the Boeing
company both Windows, Macintosh, and Unix workstations using a
common software product with common user interface." Vesey Dep.,
1/13/99, at 271:6-24.
- Microsoft's Joe Belfiore testified that Microsoft makes the user interface of
the cross-platform versions of Internet Explorer consistent with the
Windows version to decrease training costs. Belfiore Dep., 1/13/99, at
369:13 - 370:21.
- In discussing the benefits to organizations from having a standard word
processor, a standard spreadsheet, or a standard web browser, Weadock
testified that "[t]here are many benefits, many cost savings, and
configuration savings. You have benefits to the user in terms of
productivity. They don't get distracted. They -- they can learn one
application and use that to do word processing or to do web browsing.
There are also advantages in terms of technical support. You don't have
to teach your technical support staff all about how to support two
browsers. You can teach them how to support one browser because
that's the standard in the company." Weadock, 11/17/98am, at 70:11-15;
see also Weadock Dir. ¶¶ 38-39.
- Joe Kanicki of Dell Computer explained that, in Dell's experience,
corporations often want to standardize on a single browser for "stability
and for support. The total cost of ownership for the corporation stabilizes.
The more frequently products are revised, the more expensive it is or
potentially could be for a corporation to stay up with those revisions."
Kanicki Dep., 1/13/99, at 331:3 - 332:10
106. Standardizing on one browser also permits an organization to develop
specialized internal applications or viewable content more cheaply and with confidence
that those resources will be compatible with all its internal systems that focus on the
Internet.
- Vesey testified that "the single defining quality" that makes the web
valuable to Boeing is the ability "to put an electronic document in one
place and have it be accessible by virtually anybody, irrespective of
platform." Vesey Dep. (played 11/17/98am), at 23:13-19.
- Vesey testified that, if Boeing had to deploy both Navigator and Internet
Explorer, its support costs would "'be higher due to a couple of things.
Probably first and foremost would be that the potential for a web
application developer to develop an application that depended specifically
on a particular site, from the end user perspective that would possibly be
the biggest impact. They would have to know . . . when I'm assessing this
particular web site, I have to use this particular browser. And then if they
tried to go to that site with an alternative browser, they wouldn't be able to
render whatever content was available there. The other reason, the other
essential reason, would have to do with . . . the local use of the software.
On the Windows 95 desktop there is a default browser setting. And the
default browser behavior, generally speaking, when you have IE 4 and
Netscape 4 installed, you can alternate between having either set as the
default browser. . . . In some cases, those default browser settings do
become confused and can make it difficult for the user to get a particular
browser configured as the default browser. So that can become
confusing for end users." Vesey Dep., 1/13/99, at 288:2 - 289:11.
- Glenn Weadock testified that "companies often develop intranets
designed to work with a -- with a particular browser." He also testified that
"if something works and looks right in Navigator, it may not work and look
right to employees who are running Internet Explorer." Weadock,
11/17/98am, at 73:15-19.
- Weadock testified that users sometimes "develop their own applications
that (if useful and well designed) may spread throughout an organization.
The development of intranets internal company networks based on
Internet technologies has accelerated this trend. The greater the degree
of software standardization, the greater the likelihood that such user-developed applications can work properly throughout the organization."
Weadock Dir. ¶ 38.
- Weadock testified that "some organizations develop their own custom
software that only works with a particular browser, and that compatibility
with that custom software may provide an ongoing motivation to use that
particular browser." Weadock Dir. ¶ 24c.
107. For these reasons, a company that desires to standardize on a single
browser across several different hardware and operating system platforms will want to
make its browser choice independent of the decision to purchase any one operating
system.
- Weadock testified: "If a company is deciding, in part at least, on which
browser it wants to standardize on, based on a variety of hardware
platforms in the organization running different operating systems, then it's
a very short logical jump to state that companies are making this browser
decision independent from the decision that they make about any one
operating system." Weadock, 11/17/98am, at 16:8-15.
- Based on his research and interviews with corporate information
managers, Glenn Weadock testified that organizations generally want to
make browser decisions and operating system decisions separately.
Weadock Dir. ¶ 21; see also Vesey Dep. (played 11/17/98), at 52:12 -
53:14 (Boeing). Weadock testified that there is substantial demand for
the original (retail) version of Windows 95 among corporations, "[b]ecause
they have the greatest control over what applications they can install onto
it, because it is the cleanest version of Windows 95. It doesn't contain
software that they don't want. And, in particular, it doesn't contain Internet
Explorer, which they may not want." Weadock, 11/17/98am, at 62:12-20;
Weadock, 11/16/98pm, at 24:23 - 25:4 (testifying that some users may
elect to forego the technological advances of later versions of Windows
and use the retail version of Windows 95 because it did not come with a
web browser).
108. Microsoft recognized this separate demand for browser standardization
across, and independent of, demand for operating systems.
- David Cole urged his Win32 Internet Explorer 4 team to assist the teams
working on the Win16, Unix and Mac versions of Internet Explorer since
"[g]etting the cross platform versions done is key to market share on all
platforms, including Win32." GX 60, at MS7 004624.
- A January 1998 draft of a Transition Plan for Internet Explorer 5 for
Macintosh included the following: "Microsoft has now put out several
versions of Internet Explorer on several platforms. While the win32
version of IE has continued to make serious strides in terms of
functionality, and major inroads in terms of market share, the cross-platform versions have not made the same market share gains. While the
lack of cross-platform market share is troubling, the negative impact on
win32 IE market share is unacceptable . . . . As we talk to more and more
customers, it is becoming increasingly apparent that the cross-platform
browsers directly affect overall IE market share exponentially." GX 370, at
MS98 0121263.
- In November 1997, Brian Hall reported on a focus group study of Internet
Explorer 4 and Navigator 4 users, listing as a "key takeaway" that "The
desire is for one 'core browser' with similar UI and same content and
feature support across platforms." GX 219, at MS7 006361.
- Paul Maritz wrote in a June 1996 e-mail: "We have no desire to sell
anything on UNIX. However, owing to customer demand, we are going to
have to provide an IE solution on UNIX." GX 653, at MS98 0156372.
- According to Microsoft's own data, corporations "want our offerings to be
as consistent as possible" "to avoid confusion among their users and
support staff . . . [t]hey want uniformity on authoring, deployment,
management, and general browser user interface." GX 217, at MS98
0109147.
(3) Some consumers demand browsers and
operating systems separately because they may
wish to upgrade one without upgrading the other
109. Many consumers and OEMs demand browsers and operating systems,
including Internet Explorer and Windows, separately in order to have the ability to
upgrade the operating system without changing browsers.
- Microsoft's Bill Veghte testified that Microsoft considered shipping
Windows 98 with Internet Explorer 3 instead of Internet Explorer 4,
because there was OEM demand for hardware-related improvements like
USB support that were ready for inclusion before IE4 was completed.
Veghte Dep., 1/13/99, at 783:2 - 786:8.
- Gateway's James Von Holle testified that Gateway asked Microsoft to
release support for new hardware devices, including "AGP graphics, DVD
disks, and dual displays," for Windows 95, instead of holding those
features for Windows 98. Von Holle Dep., 1/13/99, at 302:6 - 303:12.
- An internal Gateway list
- redacted -
GX 357, at GW 026522 (sealed).
- Dell's Joseph Kanicki testified that when customers who do not have the
current version of Internet Explorer are updating their operating systems,
they may not want to upgrade to the new version of Internet Explorer.
Kanicki Dep., 1/13/99, at 335:17 - 336:2.
110. Conversely, consumers may want to obtain upgrades to their browser
application software without altering their operating system.
- As Glenn Weadock testified, "changing operating system software has a
greater potential for creating problems than changing a single application
does, inasmuch as all applications rely on the operating system" and a
change to the operating system "can cause unwanted problems with other
applications still residing on the system, or confusion among users now
confronted with changes to the operating system." Weadock Dir. ¶ 32g.
- Microsoft's Chris Jones acknowledged that customers may want to get
"the latest browsing technology" but have their "start menu and task bar . .
. remain the same." Jones Dep., 1/13/99, at 552:22-24.
- Veghte testified that it remains important for Microsoft to ship Internet
Explorer 5 as a separate product "because there will be a class of
customers that may want to get those capabilities without upgrading their
operating system." Veghte Dep., 1/13/99, at 787:5-13 (emphasis added).
(4) Some customers demand browsers and operating
systems separately because they want no web
browsing capability at all
111. Some consumers demand browsers and operating systems, including
Windows and Internet Explorer, separately because they want no web browsing
capability at all. Not all PC users want browsers, but all need operating systems.
- Microsoft's David Cole acknowledged that Microsoft "had feedback from
corporate customers that wanted to prevent access to the Internet, so
when . . . they buy a new machine from a PC manufacturer, they want the
ability to remove easy access to the Internet so their employees, you
know, aren't spending their time out on the Web doing whatever." Cole
Dep., 1/13/99, at 395:1-20.
- Joseph Kanicki from Dell testified that he believed some of Dell's
customers did not want Internet Explorer because "the customer may wish
to prevent its employees from accessing or attempting to access the
Internet or World Wide Web." Kanicki Dep., 1/13/99, at 333:11-22.
- Weadock testified that some organizations "may wish to make it difficult
for certain employees to access the public Internet, in order to reduce the
amount of unproductive time employees spend 'surfing the Net' on
subjects unrelated to their jobs. Without a browser, accessing the
Internet's World Wide Web is impractical." Weadock Dir. ¶ 23a.
- Sun's Curtis Sasaki testified that "many corporate customers . . . want to
restrict their user's access to the web" and that Sun has been told by
various customers, including the Florist Trade Bureau and several
universities, that "many of them did not want their employees to have
access to web browsing." Sasaki Dep. (played 12/16/98pm), at 26:25 -
28:22.
- Soyring testified that "[some] enterprise customers want to control the
applications which can be used by employees in the enterprise, and do
not want employees to spend time 'surfing the internet.'" Soyring Dir. ¶
17.
- When asked whether he was "aware of any customers who did not want
to install the Web browser because they didn't want their customers
surfing the web," IBM's Jeffrey Howard answered that "I am aware that we
did have requests coming in from our field personnel that talked to those
large customers who reported back to use that they wanted to restrain,
you know, what applications customers could get to, and specifically
having their employees sitting and surfing the Web on work time was a
fear that was sometimes voiced." Howard Dep., 8/31/98, at 115:20 -
116:6 (DX 2572).
- Packard Bell/NEC's Mal Ransom testified that "Typically, our corporate
customers don't want or don't necessarily want access to the Internet or
browser loaded on their employees' machines, so they've got the choice
of what they do." Ransom Dep. (played 12/16/98pm), at 74:4-8.
- Compaq's John Rose estimated that only 70% of businesses are running
a web browser on their desktops. Rose, 2/18/99pm, at 53:22 - 54:3.
- Sun's James Gosling testified that for systems without a display, like a
server, a customer would have no use for a browser. Gosling,
12/10/98pm, at 60:17 - 61:2.
111.1. Even in organizations that use the Internet regularly, there will
usually be at least some employees who do not need browsing functionality.
- IBM's John Soyring testified, for example, that some of IBM's
customers wanted OS/2 without a browser for "systems used by
baggage handlers or bank tellers." Soyring Dir. ¶ 17.
- Weadock testified that "even when we look at a company that is
investing heavily in intranet technology, such as Federal Express, .
. . they don't necessarily have browsing software or browsers on all
of their PCs. There are just some categories of users who may
have no need to access an intranet or the Internet." Weadock,
11/16/98pm, at 15:15-25.
111.2. Although a corporation might restrict an employee's access to the
Internet in other ways, such as by removing the modem or ethernet connector from
certain PCs or by limiting Internet access to a proxy server, such alternatives are often
less efficient than simply using a PC without a browser.
- Weadock testified that an employee might need a modem installed
for dial-up telecommuting, even if the employer wished to restrict
his or her access to the Web. Weadock, 11/17/98am, at 41:17 -
42:16.
- Weadock also testified that "those methods don't address the
resource use issues of having browsers on the . . . PCs. They also
don't address the issues of user confusion that might arise from
attempting to run software that is there and perhaps accessible,
even though I've tried to remove it and couldn't, and then pick up
the phone and call the Help desk and say, 'Hey, What's this?' So
there are lots of reasons, other than just resource use, that
companies may want no browser software on a PC. It's generally
accepted practice among IT managers in businesses large and
small to put the least amount of software on a computer that will do
what their users need to do. You just save all kinds of costs that
way, all the way from resource use to support and training."
Weadock, 11/17/98am, at 42:3-16.
- Jeffrey Howard testified that there are other ways to prevent users
from browsing the Web from OS/2, but that "most customers,
particularly in the Warp Version 3 and Warp Connect time frame,
usually found hard disk space to be at a premium and tried to have
the minimal amount of code that they could installed on the
desktop machine itself just from a management standpoint and
from a support standpoint, because you needed the space
available for swap files and paging, et cetera." Howard Dep.,
8/31/98, at 118:21 - 119:4 (DX 2572).
(5) OEMs are surrogates for end users; and thus, for
the above reasons, they too demand browsers
and operating systems separately
112. Because the personal computer OEM industry is extremely competitive and
OEMs must satisfy consumer demand to stay in business, OEMs also demand
browsers and operating systems, including Internet Explorer and Windows, separately.
- Gateway repeatedly asked Microsoft for a version of Windows 98
with web browsing uninstalled, in part because they were
"concerned that the installation of the full MS product (including
channels) results in a much slower system performance if the
customer chooses an alternate browser after full installation on
IE4." GX 1073, at MS 98 0204593.
- Jon Kies testified that Packard Bell/NEC took advantage of the
January 1998 stipulated remedy to offer some of its PC models
without Internet Explorer. Kies Dep. (played 12/16/98am), at 6:11-19.
- Kanicki testified that because Dell's customers "may wish to install
a competitive browser instead of Internet Explorer," Dell's license
agreement with Microsoft permits it "to install a competitive browser
on a machine that is shipped with Windows 95 or Windows 98."
Kanicki Dep., 1/13/99, at 336:4-19.
- Mal Ransom testified that because many of Packard-Bell/NEC's
"commercial customers don't want access to the Internet or
browser loaded on their employees' machines," for the Versa line
of notebook computers those customers "get the choice of which
browser to pre-install," if any. Ransom Dep. (played 12/16/98pm),
at 73:13 - 74:11.
- Compaq also made efforts to satisfy its customers' demand for
browsers other than Internet Explorer. See infra Part V.B.2.c; ¶
130.1.
c. To satisfy this separate demand, firms -- including
Microsoft -- have found it efficient to supply browsers
and operating systems separately
113. To satisfy this separate demand, both operating system vendors and
browser vendors supply browsers and operating systems separately.
(1) Internet Explorer and other browsers have been,
and continue to be, supplied separately from
operating systems
113.1. Browser suppliers have found it efficient to supply browsers
separately from operating systems.
- Netscape's Barksdale testified: "Indeed, Netscape does not sell
any operating system products, and was able to sell millions of
browser licenses to consumers and enterprises separately from
any operating system." Barksdale Dir. ¶ 90.
- James Gosling testified: "The HotJava browser is a software
application that was released by Sun in 1995. At the time the HotJava
browser was developed, Sun contemplated undertaking the revisions and
improvements necessary to maintain it as a competitive product for
desktop computers such as Windows PCs. However, after Microsoft
announced that its Internet Explorer browser would always be given away
for free, Sun concluded that it made little business sense at that time to
compete vigorously to sell a consumer browser application to compete
against a product that was being given away for free." Gosling Dir. ¶ 37;
Gosling, 12/3/98pm, at 80:17 - 81:3 (testifying that Sun never sold
HotJava "as a commercial browser" because, "given that the
market price for browsers, those days, seemed to be zero, it hardly
seemed like a sensible thing to do").
- Dr. Warren-Boulton noted that "Opera, which has limited presence
in some distribution channels, is distributed independently of an
operating system product." Warren-Boulton Dir. ¶ 76.
113.2. Microsoft also found, and continues to find, it efficient to supply its
browser separate from any of its operating system products in numerous channels.
113.2.1. Microsoft has consistently offered its Internet Explorer
browser on a standalone basis at retail, by downloading, and through ISPs, OLSs and
ISVs.
- See supra Part V.B.1.a.(3); ¶ 99.
113.2.2. In response to competition from other browsers, and in
order to satisfy demand for a standard browser product to run on multiple operating
systems and thereby increase Internet Explorer's market share, Microsoft also created
standalone versions of Internet Explorer that run on other operating systems and earlier
versions of Windows.
- A Microsoft focus group study in November 1997 shows that
"Win32 browser qualities are reflected on to other platform
version in users' minds" and users' desire "is for one 'core
browser' with similar UI and same content and feature
support across platforms." GX 218, at MS7 006353.
- Chris Jones wrote in November 1995: "To compete with
netscape, we need to have cross platform
(Win3.1,Win32,Mac) clients which support the NT server
(log-on, security, etc.)." GX 334, at MS98 0104685.
- Barksdale testified: "To compete with Netscape, Microsoft
began offering cross-platform versions of Internet Explorer."
Barksdale Dir. ¶ 91.
- James Allchin stated that Internet Explorer for the Macintosh
is an application, and not a part of any operating system.
Allchin, 2/2/99pm, at 13:8-12.
- Dean Schmalensee testified that Internet Explorer for the
Macintosh and for Windows 3.x are applications, and not
part of the operating system. Schmalensee, 1/20/99am, at
20:14 - 21:9.
113.2.3. Except for minor differences to comply with user interface
guidelines for those other systems, and evidencing that they respond to the same
separate demand for Internet browsers, the non-Windows versions of Internet Explorer
supply the same browsing functionality and "look and feel" to the end user as Internet
Explorer for Windows 95/98.
- Professor Edward Felten testified: "The Windows 98 and
Solaris versions of IE Web browsing offer nearly identical
user interfaces, and the MacOS version offers the same
user interface modified to meet the user interface guidelines
specified by Apple for Macintosh software." Felten Dir. ¶ 75;
Felten Dir. ¶ 82 (testifying that a user's web browsing
experience with the versions of Internet Explorer running on
the Sun Solaris, Apple Macintosh, and Windows 98 is
substantially similar).
- Joe Belfiore testified that Microsoft makes cross-platform
Internet Explorer to appeal to companies with non-Windows
operating systems and makes the user interface of the
cross-platform versions consistent with the Windows version
to decrease training costs. Belfiore Dep., 1/13/99, at 369:13
- 370:21.
- Microsoft created the cross-platform versions of Internet
Explorer specifically to appeal to organizations that wanted
to use the same browser across multiple platforms. See
supra Part V.B.1.b.(2); ¶ 110.
(2) Operating system vendors -- at least those
which, unlike Microsoft, lack market power --
supply operating systems separately from
browsers
114. Operating systems are efficiently provided separate from browsers, and
every operating system vendor other than Microsoft supplies operating systems
separately.
(a) Some operating system vendors offer
consumers the choice of licensing the
operating system without a browser
115. A number of operating system vendors offer consumers the choice of
licensing the operating system without a browser.
115.1. Sun does not bundle any browser with its JavaOS operating
system.
- Sasaki testified that Sun licenses its JavaOS product separately
from its HotJava browser; "the product Java OS ships to our
licensees, our licensees can also license the browser technology
[called HotJava], and its up to them to decide whether or not they
include it in their product or not." Sasaki Dep. (played
12/16/98pm), at 21:25 - 23:6. Sasaki also testified that the price of
JavaOS does not include a browser (Sasaki Dep. (played
12/16/98pm), at 26:8-16) -- and that only 21 out of Sun's 36
JavaOS licensees also licensed the HotJava browser. Sasaki
Dep. (played 12/16/98pm), at 26:17-24.
115.2. Lucent offers an unbundled option.
- James Frasca testified that Lucent's "view is that the web browser
is part of the application suite, not the operating system" and that
Lucent has licensed versions of Inferno without a web browser.
Frasca Dep., 1/13/99, at 137:15-19; Frasca Dep., 1/13/99, at
141:19-22; Frasca Dep., 1/13/99, at 143:8-9; Frasca Dep., 1/13/99,
at 144:16 - 145:9 (Lucent would license to a hardware OEM a
version of the Inferno product without the browser if the OEM
wanted to distribute a third-party browser).
115.3. Santa Cruz Operation offers an unbundled version of its operating
system.
- With Unixware 7, a multi-user product, SCO bundles only "a single-user license for the administrator to read Online Doc and to
manage the web server." Additional browser licenses for additional
users must either be purchased from SCO "as an optional product"
or acquired elsewhere. Rasmussen Dep. (played 12/15/98am), at
59:23 - 60:15.
115.4. Caldera offers an unbundled version of its operating system.
- Bryan Sparks testified that Caldera allows OEMs to offer an
unbundled version of its operating system. He explained: "It
doesn't make sense for us to" require OEMs to include the
browser. He continued: "The reseller knows what the customer
needs better than we do. He is closer to the customer. We let him
decide that. He is buying the boxed product and has the browser,
but we don't mandate that he install it or configure it, if he doesn't
wish to." Sparks Dep. (played 12/16/98am), at 50:12-23.
(b) Operating system vendors other than
Microsoft sometimes bundle one or more
browsers with their systems but allow
VARs, OEMs, or end users to remove them
or not to install them
116. Operating system vendors that lack Microsoft's monopoly power, and
hence its incentive to engage in anticompetitive conduct that thwarts consumer demand, do not
impose contractual or technical restrictions on OEMs' or end users' ability to remove a
browser.
116.1. No operating system vendor other than Microsoft places
restrictions on its customers' ability to remove an unwanted browser.
- Allchin testified, "as I sit here today, I don't know of any" operating
system vendor other than Microsoft that bars its customers from
removing any browser. Allchin, 2/3/99am, at 45:11-19.
116.2. Even when other vendors offer a browser with their operating
system, they allow OEMs and end users to remove it or not to install it.
116.2.1. Although IBM includes a browser as an application in its
OS/2 Warp version 4 operating system package, the installation process allows the user to
choose whether or not to install it. IBM also permits any other OEM or value-added
reseller (VAR) selling computers with OS/2 to remove the browser before the sale.
116.2.1.1. IBM does not consider the browser to be part of
the operating system.
- John Soyring testified on cross-examination that
IBM's Web Explorer "is not part of the OS/2 operating
system itself. . . . We did develop it separately as a
separate program. It is included in the OS/2 Warp
product package. And we set it up as a selectively
installable and selectively removable application
program that can be either used with or not with
OS/2." Soyring, 11/18/98am, at 21:12 - 22:2; see
also Soyring Dir. ¶¶ 14-18.
- Soyring also testified that OS/2 performs properly as
an operating system whether or not any web browser
is installed. Soyring, 11/18/98pm, at 78:5-7.
116.2.1.2. IBM permits users to not install or to remove their
browsers.
- Soyring testified that users of IBM's OS/2 operating
system have always been free not to install "Web
Explorer," to remove that browser after installation,
and to install a competing browser if desired.
Soyring, 11/18/98pm, at 77:12-17.
- IBM's Dana O'Neal testified, "
- redacted -
O'Neal Dep., 8/31/98, at 72:3-7 (DX 2578A)
(sealed).
116.2.1.3. IBM makes the browser removable from its
operating system because it recognizes that there is a separate demand for browsers
and operating systems.
- Jeffrey Howard testified that he was "aware that we
did have requests coming in from our field personnel
that talked to those large customers who reported
back to us that they wanted to restrain, you know,
what applications customers could get to, and
specifically having their employees sitting and surfing
the Web on work time was a fear that was sometimes
voiced." Howard Dep., 8/31/98, at 115:20 - 116:6
(DX 2572).
116.2.1.4. IBM includes a browser in its packaging for OS/2
for the same reason it bundles other applications like a word processor: because it
helps convince customers that key applications exist for OS/2, which is necessary in
order to overcome the applications barrier to entry.
- Soyring testified that IBM chose to bundle Netscape
Navigator in particular with some later versions of
OS/2 "because, at that time, its brand was the most
popular brand recognition in the industry. And, again,
it goes back to the problemg we were facing before in
that not--popular applications just hadn't been buil[t]
for OS/2. So we thought by, one, delivering
customers earlier, and, secondly, getting a major
brand to recognize and adopt the OS/2 operating
system by offering a product would be an additional
spur for--or stimulant to sell additional OS/2 copies.
So we entered into a licensing agreement. We spent
millions of dollars with Netscape to be able to make
that happen and we packaged it as part of the next
generation of OS/2 Warp, which is OS/2 Warp 4 in
the shrink-wrapped product." Soyring, 11/18/98am,
at 39:3-19; Soyring, 11/18/98am, at 44:9 - 45:1
(explaining that the same reasoning drove IBM to
bundle word processing, spreadsheet, database, and
personal information management applets in OS/2);
Soyring, 11/18/98pm, at 75:10-21 (discussing users'
perceptions that they would have "a difficult time
finding applications" for OS/2).
116.2.2. Apple bundles both Internet Explorer and Netscape
Navigator with the MacOS but permits both end users and value added resellers to
delete either or both.
116.2.2.1. Apple does not consider a browser to be part of
the operating system.
- Tevanian's definition of an operating system says nothing
about browsing capability. He defined an operating system
as "the primary software that controls a computer. The
operating system provides various basic services for a
computer such as process management, user interaction,
data management for the hard disk, network interfaces and
control of peripheral devices such as printers and
keyboards." Tevanian Dir. ¶ 8.
116.2.2.2. Apple allows users and resellers to remove either
Navigator or Internet Explorer or both if they wish and does not "hard-code" anything in
its operating system to require the use of a particular browser.
- Apple's Tevanian testified: "The Mac OS operating
system will continue to function if either or both of
these browsers are removed. As noted above, we
permit value-added resellers ('VARs') the flexibility to
reconfigure our systems to meet their direct
customers' needs. We provide VARs the flexibility to
remove browsers or other applications, and to
reconfigure the Macintosh desktop to address what
they perceive to be their customers' desires."
Tevanian Dir. ¶ 26.
- After the Court asked Tevanian whether it is "possible
for you to extricate your browser from operating
system without otherwise impairing the operation of
the system," Tevanian replied, "Yes, other than you
can't browse the web." Tevanian, 11/5/98pm, at
67:10-15; Tevanian, 11/5/98pm, at 70:9-17 (testifying
that the operating system would remain intact).
116.2.2.3. Apple allows users to remove the bundled browsers
because it understands that there is separate demand for browsers and operating systems.
- Upon being asked whether he felt that there was a
"separate market" for Internet browsers, Tevanian
stated that he thought "it's fair to say there is a
market. There are some people who, first, they would
select the operating system; then they might select
the browser, and not want to make the decision
together. So in that sense, it's separate from the
desktop computer market in general." Tevanian,
11/4/98pm, at 18:3-22.
116.2.3. Sun bundles its "HotJava" browser with its Solaris
operating system but permits end users to remove that application.
116.2.3.1. Sun does not consider a browser to be part of
the operating system.
- Curtis Sasaki's definition of an operating system says
nothing about providing browsing capability. He says that
an operating system includes "a kernel which controls how
things are managed in terms of memory. It also controls
the I/O functionality, such as talking to a network, talking
to your keyboard, displaying things on the screen. So,
that's called device drivers. So all of that is what I would
consider an operating system, as well as a set of APIs
which are on top, which application developers write to."
Sasaki Dep. (played 12/16/98pm), at 17:16-25; Sasaki Dep.
(played 12/16/98pm), at 26:8-10 ("Q:: Is the browser part
of the operating system? A: No. It's separate. Q: When--are--does the price of the Java OS product include a
browser? A: No. It does not. Q: Are there separate prices
for browsers? A: That's correct.").
- Sun's James Gosling similarly excludes browsing
capability from his defintion of an operating system. He
says that an "operating system has two primary functions:
(1) to interact with and control the computer's processor
and other hardware (monitors, keyboards, disk drives, etc.);
and (2) to interact with, and execute instructions from,
software applications, generally through a series of
applications programming interfaces known as 'APIs.'"
Gosling Dir. ¶ 8. Based on this definition, Gosling
concludes that "the browser is best understood as a
software application, not as part of a computer's operating
system." Gosling Dir. ¶ 38; Gosling, 12/9/98pm, at 30:23 -31:9.
116.2.3.2. Sun permits and makes it easy for end-users,
VARs, and OEMs to remove bundled browsers from the operating system if they so
desire and does not "hard-code" anything in its operating system to require the use of a
particular browser.
- Gosling testified that Sun included the "HotJava"
browser on the CD-ROM with its Solaris operating
system, but "it was absolutely a replaceable,
repluggable application. We didn't tell anybody that it
was nonremovable, like any of the other pieces that
happened to be there. Customers can and do use
replacements for just about everything." Gosling,
12/9/98am, at 38:16-25.
- Brian Croll testified that when an OEM or a VAR
licenses Solaris 2.6, it is not required to ship the
HotJava browser to its end user customers. Sun
provides two means for OEMs and VARs to offer an
unbundled version of the operating system: "At one
level you can choose not to add that package which
has the Java browser as well as other things, if that's
the first choice. Then the second choice is, once you
have loaded it, you can go through the de-install
process to take away." Croll Dep. (played
12/15/98pm), at 66:22 - 68:4.
- Moreover, Croll testified that OEMs and VARs are
permitted to supply their end user customers with
additional browsers or a different browser if they
choose to do so because "there is no reason for us to
keep them from doing that." Croll Dep. (played
12/15/98pm), at 68:5-13.
- Curtis Sasaki stated that when the Java OS is in use on a
network, "the system administrator . . .can remove the
browser and not affect Java OS." Sasaki Dep. (played
12/16/98pm), at 29:3 - 30:7.
116.2.3.3. Sun allows users, VARs, and OEMs to remove
bundled browsers because it understands that there is separate demand for browsers and
operating systems.
- James Gosling testified that he could not think of "any
plausible technical reason to design Windows 98 in a way
that makes it difficult to remove Internet Explorer." By
contrast, Gosling proposed several reasons "why it would
be desirable to design the operating system so that the
browser could be removed," including the facts that users
might want to deploy their operating systems without
displays (e.g. as a server), they might want to replace their
browsers with superior products, or they might want to
utilize specialized browsers, such as a browser designed for
persons with visual impairments. Gosling, 12/10/98pm, at
60:10 - 62:1.
- Croll testified that the Web Start and Answer Book 2
features of the Solaris operating system do not
require the use of the HotJava browser that is
bundled with the OS in order to function "because we
assume that after the operating system is loaded for
the first time that customers are going to want to have
other browsers." Croll Dep. (played 12/15/98pm), at
64:23 - 65:12.
116.2.4. The Santa Cruz Operation (SCO) bundles Netscape
Navigator with its products but permits customers to choose whether to install or
remove it.
116.2.4.1. SCO does not consider a browser to be part of
the operating system.
- When asked whether any browser product is part of
the core of any SCO operating system product, Ron
Rasmussen answered no. "Our view is that the
browser is an application." Rasmussen Dep. (played
12/15/98am), at 64:13-20.
- Rasmussen testified that SCO "bundles" Netscape
Navigator with its OpenServer and Unixware
products, (Rasmussen Dep. (played 12/15/98am), at
54:10 - 56:18) but that "Our view is that the browser is
an application." Rasmussen Dep., (played
12/15/98am), at 64:20. Rasmussen also testified that
"When SCO says 'we bundle a feature,' it means it's
a feature which is not part of the core base operating
system functionality. It means that it's something that
the user can choose to install or remove, and the
operating system, whose primary function it is to
serve applications, will still function properly."
Rasmussen Dep. (played 12/15/98am), at 55:14-19.
116.2.4.2. SCO permits end users, VARs, and OEMs to
remove bundled browsers from the operating system if they so desire and does not
"hard-code" anything in its operating system to require the use of a particular browser.
- Rasmussen testified that there are a number of ways
to remove the bundled browser from Unixware 2.1.3:
"The first way is during installation of the operating
system, the browser can be deselected so it never
does get installed on the system -- on the hard disk of
the computer." Moreover, "[i]f they chose to install
Navigator as part of the operating system installation,
they can go back in with a utility to do software
removal and they can, again -- they get a point-and-pick list and they can select it for removal."
Rasmussen Dep. (played 12/15/98am), at 60:21 -
62:7.
- Rasmussen further testified that Navigator is
uninstallable and/or removable from Unixware 7 "in
the same fashion" as from Unixware 2.1.3.
Rasmussen Dep. (played 12/15/98am), at 62:15-24.
- Similarly, Rasmussen testified that Navigator is
optional on Open Server Release 5 as well: The
browser is "removable both at initial system load time
so it never gets onto the system, and it's also
removable afterwards if you chose during initial
system load to install it." Rasmussen Dep. (played
12/15/98am), at 62:25 - 63:6.
116.2.4.3. SCO allows users, VARs, and OEMs to remove
bundled browsers because it understands that there is separate demand for browsers and
operating systems.
- Rasmussen testified that the reason SCO gives its
users those options with regard to the web browser is
that: "Not everybody wants the functionality in the
operating system, so we provide them that option to
remove it or install it at a later time." Rasmussen
Dep. (played 12/15/98am), at 62:25 - 63:18.
Rasmussen further explained that "people are buying
the operating systems to run applications, and if their
application does not require a web browser, then
perhaps they don't want the web browser there. So if
you're running an accounting application and you
don't need the web browser, perhaps you don't want
to install it to save space on the disk, or, in some
instances, we've had resellers tell us they consider
the web browser an unproductivity tool as people surf
the web rather than doing their work." Rasmussen
Dep. (played 12/15/98am), at 63:19 - 64:6.
116.2.5. Operating system vendor Be, Inc., bundles the only
browser currently available with its BeOS but permits users to remove it.
116.2.5.1. Be does not consider a browser to be part of the
operating system.
- Although both Be and Microsoft's James Allchin
describe the Net Positive browser as an "integrated
browser," Be refers to Net Positive as an application
and lists it in the "apps" directory on the computer.
Allchin, 2/2/99am, at 14:3-11.
116.2.5.2. Be permits end-users, VARs, and OEMs to
remove bundled browsers from the operating system if they so desire and does not
"hard-code" anything in its operating system to require the use of a particular browser.
- The Net Positive browser can be removed from the
BeOS. Although Be's help system will not function
fully in the absence of a browser, the help system will
work if another browser or HTML renderer is installed
after Net Positive is removed. Allchin, 2/2/99am, at
10:4 - 21:4.
- Allchin acknowledged that removing the Net Positive
browser from the applications directory frees up 1.3
megabytes of RAM on the BeOS applications
directory. Allchin, 2/2/99am, at 13:5 - 19:20; GX
1771.
116.2.6. Novell bundles a browser with Netware but permits the
user to remove it and use a third-party browser.
116.2.6.1. Novell does not consider a browser to be part of
the operating system.
- Novell's Sean Sanders defined "a desktop operating
system" in a way that says nothing about browsing
capability. He defined it as "a special set computer
programs that allows for the management of . . .
computing resources that a specific end user would
use on their desktop PC. So it allows them to kind of--behind the scenes it does some management of the
physical computer such as managing the memory,
the disk drive and some of the other technical aspects
that are included within the box. But it also provides
a--kind of a--generally a friendly front end to the
system that the user can manipulate to better use
their software programs and the resources that are
generally there specific to the desktop." Sanders
Dep. (played 1/13/99), at 185:13 - 186:3.
116.2.6.2. Novell permits end-users, VARs, and OEMs to
remove bundled browsers from the operating system if they so desire and does not
"hard-code" anything in its operating system to require the use of a particular browser.
- David Wright testified that the Netware 5.0 operating
system will function without the browser. Wright Dep.
9/18/98, at 16:4-12 (DX 2601). Similarly, if the
browser is initially installed and then uninstalled, the
operating system will still function. Wright Dep.,
9/18/98, at 16:13-18 (DX 2601).
- When asked how the relationship between browser
products and Novell's Netware differed from the
relationship between Internet Explorer and Windows
98, Weadock testified: "It differs in several
fundamental ways. For example, Netware as an
operating system does not depend on any particular
browser. The browser that comes on the Netware CD
is used, for example, to access the help and
documentation for the Netware product. Users --
customers that choose not to use that browser with
Netware can remove it. Users that choose to use
another browser, a different browser, can install a
different browser. They can install Internet Explorer if
they want. So in those key areas we see differences
between how Novell, quote unquote, bundles
Netware and a browser -- I mean, it's a very loose
bundling and it offers the customer significant choices
-- with what Microsoft is doing with Windows 98, in
which the browser is nonremovable, in which the files
associated with the browser do, in fact, disable the
operating system if you go out and delete them one
by one, and in which case the customer is not
completely free to install an alternative browser
because of the hardwired methods within the user
interface of Windows 98 that still invoke Internet
Explorer regardless of the actions that the customer
may have taken to reverse or disable that choice."
Weadock, 11/17/98am, at 47:1 - 48:1; see also
Weadock, 11/16/98am, at 84:13-23.
116.2.6.3. Novell allows users, VARs, and OEMs to remove
bundled browsers because it understands that there is separate demand for browsers and
operating systems.
- Sanders testified that, when a user installs the Intra-Netware product, they have a choice of whether or
not to install Netscape Navigator. Sanders Dep.,
1/13/99, at 186:10-16. Sanders explained that
Novell's rationale for giving users this choice is that:
"Some users do not have desire to use all of the
functionality that comes in the entire bundle that is
Intra-Netware, and as such, we provide them with the
option to make those decisions as to what they would
choose to use and not to use." Sanders Dep.,
1/13/99, at 190:18 - 191:2.
116.2.7. Caldera gives users a choice of multiple browsers with its
OpenLinux product but makes them easily and fully removable.
- Caldera bundles the KDE browser on its OpenLinux
operating system; in addition, "Netscape is preloaded by
default onto those systems." Felten, 6/10/99am, at 26:1-20.
116.2.7.1. Caldera does not consider a browser to be part
of the operating system.
- Sparks testified that he does not consider any
browser to be "part of the Linux operating system."
Sparks Dep. (played 12/16/98am), at 50:8-11.
116.2.7.2. Caldera permits end-users, VARs, and OEMs to
remove bundled browsers from the operating system if they so desire and does not
"hard-code" anything in its operating system to require the use of a particular browser.
- Professor Edward Felten testified that Caldera's KDE
browser "is separable and removable and
replaceable." Felten, 6/10/99am, at 25:12-17. When
asked what his basis was for saying that the browser
is separate from the operating system, Professor
Felten answered: "Well, there are several reasons for
saying that. First of all, the KDE browser is
developed by a different organization than the one
that develops the Linux operating system. . . . In
addition, the Linux operating system works with other
browsers. In fact, the OpenLinux works fine with
Netscape, and Netscape is preloaded by default onto
those systems. And, in addition, the KDE browser
runs on other operating systems, such as Solaris,
HP-UX and IRIX." Felten, 6/10/99am, at 26:1-18.
- Allchin conceded that the browser bundled by
Caldera comes from a different organization, KDE
(Allchin, 2/1/99pm, at 73:5-15) and that the operating
system will continue to work if it is removed. Allchin,
2/1/99pm, at 73:25 - 74:13.
116.2.7.3. Caldera allows users, VARs, and OEMs to
remove bundled browsers because it understands that there is separate demand for browsers and
operating systems.
- Bryan Sparks testified that Caldera allows OEMs to
sell a version of OpenLinux without a browser
because: "Why wouldn't we? As long as we had a
contract. I'm not in a position where I can be picky on
customers that I can get. So if they wanted to
customize it, they'd be happy to." Sparks Dep.
(played 12/16/98am), at 50:24 - 51:8.
(c) Until recently, Microsoft likewise
accommodated this separate demand by
enabling users to remove Internet Explorer
from Windows
117. Although Microsoft required OEMs and users to obtain Internet Explorer in
order to obtain Windows, it nonetheless continued, until recently, to recognize separate
demand for an operating system without a browser by supplying end users (although
not OEMs) with a means of removing or "uninstalling" the browser.
- See infra Part V.B.2.d.(4).(a); ¶ 139.
2. Microsoft tied Internet Explorer to Windows in order to impede
browser rivals and protect its operating system monopoly
118. By contrast to other operating system vendors, Microsoft both refused to
license its Windows operating system without a browser and imposed restrictions-- first
contractual and later technical -- on OEMs' and end users' ability to remove its browser
from Windows. As its internal contemporaneous documents and licensing practices
reveal, Microsoft decided to tie Internet Explorer and Windows together in order to
prevent Netscape from developing into a significant threat to Microsoft's operating
system monopoly, and not for any pro-competitive purpose.
a. Before it decided to blunt the browser threat, Microsoft
did not plan to tie its browser to Windows
119. Microsoft argues that it made the decision to build its own browser and
bundle it with Windows 95 at an April 1994 retreat dedicated to Internet issues (Allchin Dir. ¶¶
225-227). That argument is inconsistent with the evidence, which shows instead that
Microsoft had no firm plans at that time to bundle its browser with the operating system.
119.1. Microsoft's internal correspondence and external communications
from early to mid-1994 show that Microsoft was planning, at most, to bundle low-level
Internet "plumbing" such as a TCP/IP stack, but not applications such as a browser,
with Windows 95.
- In response to a question about how to handle press and OEM
inquiries concerning Microsoft's internet plans, Alec Saunders
wrote in April 1994: "It's getting very confusing and at the moment
a lot of external people are asking if we will be shipping internet
apps. The position we have taken so far is that Chicago [Windows
95] contains all the plumbing you need to hook up to the net -- but
cool apps like Mosaic are stuff you need to obtain from 3rd parties."
GX 124.
- A February 1994 email from David Cole to Bill Gates and other
senior executives reporting on "Chicago beta 1 content" provides a
detailed description of features of Windows 95 but does not
mention integration or bundling of web browsing functionality. Its
only mention of Internet support is a reference under the title
"Networking" to "Complete TCP/IP support. A fast, protmode stack
with no low memory requirements. A full set of TCP utilities.
Windows Sockets support for 16 and 32-bit apps. DHCP support
for 'plug and play' TCP/IP networking." GX 597, at MS98 0010791.
- Steven Sinofsky wrote in June 1994: "We do not currently plan on
any other client software [in the upcoming release of Windows 95],
especially something like Mosaic or Cello." In fact, Microsoft's goal
at the time was to "get[] as many third parties writing as many
internet things on top of WinSock as posible, including as many
WWW, Gopher, TN3270, etc clients as they can afford to do." GX
125.
- A November 1994 draft of a "communications story" for marketing
Windows 95 made no mention of inclusion of an Internet browser,
claiming only that Windows 95 "supports popular Internet
applications, such as Mosaic, WinWAIS or WinGopher." Brad
Chase responded that "i don't think we deliver what you say. I
think integration is impt but we don't really integrate. You still use a
phone for example. You have to get some third party program to
actually have a UI into the Internet." GX 601.
119.2. Microsoft publicly stated throughout 1994 that Windows 95 would
not include a browser.
- A November 1994 marketing brochure entitled "Microsoft Windows
95 Questions and Answers" responds to the question "Can
Windows 95 connect to the Internet?" as follows: "Yes. Windows
95 includes the networking support you need to connect to the
Internet. It includes a fast, robust, 32-bit TCP/IP stack . . . as well
as PPP or 'dial-in' support. Windows 95 supports the large number
of tools used to connect to the Internet, such as Mosaic, WinWAIS,
and WinGopher, through the Windows Sockets programming
interface. Windows 95 also includes standard Internet support,
such as telnet and ftp." GX 398, at MS98 0107100.
119.3. The testimony of Phillip Barrett, a former Microsoft employee who
was responsible for the early development of Internet Explorer, confirms that Microsoft
had no genuine plans to bundle a browser with Windows 95 in 1994.
- Barrett testified that he attended the April 1994 "offsite" at which
the subject matter was "what was Microsoft going to do about the
internet." (Barrett Dep., 1/13/99, at 100:3-14) and participated in a
"breakout" session led by Bill Gates and also including Brad
Silverberg and John Ludwig (Barrett Dep., 1/13/99, at 101:14-18)
which focused on "the internet service providers and the necessary
plumbing--plumbing being the infrastructure--to allow large
numbers of people to get online and use the internet." Barrett
Dep., 1/13/99, at 101:21-25.
- Barrett testified: "The plans were to put a TCP/IP pack . . . and
then dial up modem support" into either Windows 95 or a
subsequent service pack. Barrett Dep., 1/13/99, at 106:9-14.
Barrett does not "recall any discussions taking place" about
building a web browser into Windows 95, (Barrett Dep., 1/13/99, at
107:2-4) and testified that, to the best of his knowledge, no such
plans had been made by the conclusion of the retreat. Barrett
Dep., 1/13/99, at 107:24. Barrett testified that Bill Gates' assertion
that Microsoft decided to integrate a web browser into Windows 95
at the April 1994 retreat "is not consistent with my memory of the
retreat." Barrett Dep., 1/13/99, at 111:9-10.
- Barrett testified that, after the retreat, he moved into the Windows
group "to focus on Internet technology." Barrett Dep., 1/13/99, at
108:2-7. Barrett's new job was to "figure out a strategy" with
respect to development of a web browser. Barrett Dep., 1/13/99, at
108:13-16. Barrett testified that he and his group did not develop a
web browser and that at the time he left Microsoft in October 1994
he was not aware of any plans to develop a browser for inclusion in
Windows 95. Barrett Dep., 1/13/99, at 108:21 - 109:9. Barrett also
testified that Microsoft could not have had such formal plans
without his knowledge, because they "would have fallen into [his]
area of responsibility." Barrett Dep., 1/13/99, at 109:11-21.
120. Even after it became aware of the threat posed by Netscape's browser,
and as late as June 1995, Microsoft had no firm plans to bundle its browser with
Windows 95. Instead, Microsoft planned to ship its browser in a separate "frosting"
package (eventually called Microsoft "Plus"), for which it planned to charge.
- A January 1995 draft press release announcing the purchase of the
Mosaic code stated: "At the present time there are no plans to ship. . . the
Mosaic software in the Windows 95 box when it ships in August of this
year. . . . Our plan is to deliver this capability shortly after Windows 95
ships." GX 138, at MS6 600545.
- A document entitled "Top 20 Features Microsoft Windows 95" lists
"Plumbing for the Internet" as number 13, and states that "Windows 95 . .
. has all the necessary 'plumbing' built into the operating system, and
provides you with all the necessary 32-bit drivers such as TCP/IP to
access the net, dial up protocols such as [PPP and SLIP, and] built in
tools and utilities to make the basic connections, such as FTP and telnet.
Now you can access the internet directly, through the Microsoft Network,
or add WEB browsing capabilities by adding Microsoft Plus! to Windows
95." GX 152.
- In an email exchange with Ben Slivka on June 15, 1995, Brad Chase
observed that "there is still an effort to throw this [Internet
Explorer/O'Hare] into windows 95." Slivka responded that "[the u]pgrade
schedule is pretty tight," and that "[i]f we're not in the upgrade, it makes
our life easier, and we get more Plus revenue. . . :-)" GX 149.
- Ben Slivka suggested to Brad Silverberg in April 1995 that Microsoft might
not want to put Internet Explorer in the Windows 95 box because of size
constraints. "Putting in the Web browser is possible, but it's 475k
(compressed 170k), and it's not useful unless you're already online, and
you're already struggling to fit on 12 disks." GX 146.
- A June 28, 1995, update for Microsoft executives on the testing process
for adding O'Hare to the OEM version of Windows 95 states that "we still
don't have a firm go-ahead. Each 'Meeting to decide to do/not do this'
becomes 'let's keep going and meet in two more days.'" GX 151.
- Based on Microsoft's internal contemporaneous documents, and other
evidence, Professor Fisher concluded that Microsoft made the decision to
bundle Internet Explorer with Windows "no earlier than the middle of
1995." Fisher, 1/6/99pm, at 26:7-8.
b. Microsoft changed its plans, and decided to tie its
browser to Windows, in order to impede Netscape
121. In late 1994 and early 1995, however, executives within Microsoft began to
realize
that the popularity of Netscape's browser posed a serious threat to Microsoft's
operating system monopoly.
- See supra Part III.B.2; ¶ 56.
122. As a result, Microsoft decided that gaining a large share of the Windows 95
browser usage market was the best way to prevent Navigator from developing into a genuinely
competitive rival platform.
- See supra Part V.A.
123. Microsoft considered a variety of strategies for rapidly gaining a large share
of the browser market and, as explained, tried to coerce Netscape into abandoning its
Windows 95 browser business altogether.
- See supra Part IV.A.
124. But in early 1995 Microsoft executives also began to consider bundling
Internet Explorer with the upcoming release of Windows 95 and forcing OEMs to take it.
- In an April 1995 email, Brad Silverberg told Ben Slivka and John Ludwig
that "I have spoken to Paulma and he is in agreement that we should get
our Internet client distributed as broadly as possible as soon as possible.
What this means is that I want Ohare in Win95." GX 608.
- A June 1995 summary prepared by John Gray of a meeting discussing
the issue reports that they "[t]enatively decided to procede on path of
putting Ohare and Rome into initial OEM products but NOT in retail." GX
612, at MS98 0122185.
125. Microsoft's motive in tying Internet Explorer to Windows 95 (that is, in
refusing to offer OEMs the option of licensing Windows without the browser) was to
thwart the platform threat posed by Netscape's browser.
- On June 23, 1995, John Ludwig wrote to Paul Maritz and others that
"obviously netscape does see us as a client competitor. i'm glad you
didn't tell them many specifics. we have to work extra hard to get ohare
on the oem disks." GX 623.
- In April 1995, Rick Rashid of Microsoft wrote to Paul Maritz that "[j]ust as
they [Netscape] are a threat to us, we are a threat to them. Our best
interest is served by effectively eliminating the special browser and
special server model altogether and making the Windows desktop the
'browser' and Windows NT the 'server.'" GX 521.
- Allchin conceded that he believed that including Internet Explorer in the
OEM version of Windows 95 would be a way to help increase Internet
Explorer's market share. Allchin, 2/3/99am, at 56:7-11. "The sooner we
got it to everybody, the better off we would be. That was absolutely
believed. And we were going to distribute it through every vehicle we
could." Allchin Dep. (played 2/3/99am), at 58:2-5; Allchin, 2/3/99am, at
58:9-22.
- Dr. Warren-Boulton testified that "if, indeed, Windows 98 were provided
separately and distinctly without browser functionality, that given what I
understand to be the costs, incremental costs, of providing it separately
and given what I understand to be the potential demand for the product to
be provided separately, that it would be profitable to provide that product
separately." However, "because of Microsoft's incentive to control the
browser market, what would otherwise be profitable to sell as a separate
product is not being sold as a separate product." Warren-Boulton,
11/24/98pm, at 37:3 - 38:2. "A monopolist of an operating system has a
particular incentive not to allow the market to have, if you like, a level
playing field choice." Warren-Boulton, 11/24/98am, at 59:10-12.
- Based on the contemporaneous internal Microsoft documents, and other
evidence, Professor Fisher concluded that "Microsoft made its decision to
combine its browser and operating system not to achieve efficiencies but
to foreclose competition." Fisher Dir. ¶ 143; Fisher, 1/12/99pm, at 10:21 -
11:3 (testifying that "it's all over the Microsoft documents. They did this in
order to thwart the platform threat, in order to prevent the possibility that
Netscape and Java would lead to a situation in which the applications
barrier to entry into operating systems would be eroded.").
c. Microsoft used its operating system monopoly to
compel OEMs licensing Windows 95 also to license Internet
Explorer 1 and 2
126. Reflecting Microsoft's very late decision to tie Internet Explorer to Windows
to combat the Netscape platform threat, the first version of Windows 95 for the retail
channel did not include Internet Explorer. Microsoft offered Internet Explorer only in a
separate "plus pack" CD that it distributed entirely separately from Windows 95.
- James Allchin testified that "Microsoft included Internet Explorer 1.0 in the
OEM version of Windows 95, but not in the initial retail version. Rather, in
the retail channel Internet Explorer 1.0 was included in the Plus! package,
a set of software enhancements that Microsoft offered to customers
upgrading to Windows 95. Allchin Dir. ¶ 247.
- David Cole acknowledged that subsequent retail versions of Windows 95
came with Internet Explorer 1.0 and 2.0 on a separate disk; the primary
disk contained the original version of Windows 95 that was released at
retail in July 1995. Cole Dep., 1/13/99, at 401:10 - 402:25.
127. By contrast, Microsoft required OEMs to license a version of Windows 95
that included Internet Explorer 1 and, later, Internet Explorer 2. Microsoft required
OEMs to install Internet Explorer on all PCs on which Windows 95 was installed and
contractually prohibited those OEMs from removing the browser.
- Amendment No. 2 to Microsoft OEM License Agreement for Operating
Systems wth Dell Computer Corporation,
- redacted -
GX 1121 (sealed).
- Compaq's John Rose conceded that Microsoft's contractual provisions
required Compaq to include Internet Explorer with the PCs it shipped.
Rose, 2/18/99pm, at 8:25 - 10:1.
- Microsoft's operating system license agreement with Gateway,
- redacted - GX 458, at MS98 0009146 (sealed); GX 652 (Gateway
response to a CID) (sealed); GX 1129 (Amnd. No. 1 to Packard-Bell's
operating system license with nearly identical provisions in the "Additional
Terms," sections (a)(1) and (a)(1)(C)); GX 1183, at MS98 0009095-0009096 (Hewlett-Packard's license with nearly identical provisions in the
"Additional Provisions Key," (sections (q)(a) and (q)(a)(iii)) (sealed).
128. Because the OEMs had no commercially viable alternative to Windows 95,
Microsoft succeeded in forcing them to agree to its tying arrangement, despite clear
demand from OEMs for Windows without Internet Explorer.
128.1. For instance, Compaq removed the Internet Explorer icon in part
to feature Netscape; but, when Microsoft threatened to terminate Compaq's Windows
license, Compaq quickly capitulated to Microsoft's demands that it restore the icon.
128.1.1. In late 1995, Compaq removed the Internet Explorer (and
MSN) icons from the Windows 95 desktop on its Presario line of personal computers in
order to feature Netscape.
- John Rose acknowledged: "I understand that, in early 1996,
Compaq did remove, on some consumer products, the
Internet Explorer icon (as opposed to Internet Explorer
software) from the Windows 95 default desktop on its
Presario line of personal computers." Rose Dir. ¶ 25.
- Rose also acknowledged that Compaq had a strategy to
feature Netscape along with AOL. Rose, 2/19/99am, at
64:14-23.
- See also Part V.C.2.a.(1); ¶ 202.2.
128.1.2. Microsoft responded to the removal of the Internet
Explorer (and MSN) icons by threatening to terminate Compaq's Windows license.
- Microsoft's Don Hardwick and Microsoft in-house counsel
Peter Miller both sent letters to Compaq stating Microsoft's
intent to terminate the Windows 95 license agreement
between the companies if Compaq did not restore the icons
to their original status. GX 649; GX 650.
- See also Part V.2.a.(1); ¶ 202.3.
128.1.3. In response to Microsoft's threat, Compaq restored the
icons to the desktop.
- On June 21, 1996, Compaq gave in to Microsoft's demands.
Celeste Dunn sent a letter to Hardwick stating that Compaq
has "made the changes you requested to the Windows 95
desktop of the current release of the Compaq Presario
systems. We have replaced the Microsoft Network and
Internet Explorer icons on the Windows 95 desktop as
executable icons so they look and function exactly the same
as how we originally received them from Microsoft and have
placed Microsoft Network, Internet Explorer icons and
Internet Setup Wizard icons in their original locations under
the Start button on the Windows 95 desktop." The letter
also pointed out that icons for AOL and for Netscape were
on the Windows 95 desktop for Presario systems. GX 645.
- On June 25, 1996, Microsoft sent Compaq a letter
withdrawing the Notice of Intent to Terminate Compaq's
Windows operating system license agreement based on
Compaq's representations. GX 301.
128.2. Other OEMs recognized that they had no choice but to license
Internet Explorer along with Windows.
- See supra ¶ 129.
d. Microsoft next tied Internet Explorer 3 and 4 to Windows
95
(1) Microsoft concluded that merely tying Internet
Explorer to Windows was not sufficient to defeat
Netscape and that, to win the browser war, it must
make Windows and Internet Explorer difficult to
separate
129. Microsoft eventually concluded that its purely contractual tie between
Windows 95 and Internet Explorer would not be sufficient to prevent Netscape from
developing into a serious threat to the applications barrier to entry. It decided,
therefore, to make Windows and Internet Explorer difficult to separate.
- In a 1996 marketing plan entitled "How to get to 30% Share in 12
Months," Brad Chase wrote, "Shell Integration. The Internet is a part of
Windows. We will bind the shell to the Internet Explorer, so that running
any other browser is a jolting experience. Shell/Browser user model
becomes the same." GX 684, at MS6 6007119.
- A review of marketing plans for Internet Explorer 3 states: "What we can
do that Netscape can't -- Building on our Windows assets -- Integration, a
customer win, we can do this better -- Other technology assets (direct,
active etc.) -- Incentives." GX 488, at MS6 5005758.
- A January 1996 Microsoft presentation describes as a Response
Summary to cross-platform Java: Increased Internet Explorer share,
Integrate with Windows. GX 52, at MS7 003270.
(2) In furtherance of this objective, Microsoft tied
Internet Explorer 3 to Windows by commingling
the code that supplies web browsing with the
code that supplies operating system functions,
forcing OEMs to license that product, and
refusing to supply an unbundled option
130. In order to aid its effort to win the browser war, Microsoft offered its
operating system only as part of a software package (which it calls "Windows 95" or
"Windows 98") that also contained Internet Explorer 3 (and eventually Internet Explorer
4) and in which much of the underlying software code that supplies web-browsing and
operating system functions is contained in the same files. Microsoft thus not only used
its monopoly power to force OEMs (and end users) to take the browser with the
operating system, but also made the browser and operating system difficult to separate.
Microsoft did so despite the fact that it had, as a matter of software design, significant
discretion as to how to package its browser and operating system products. Microsoft
made a strategic design decision, rather than a design decision driven by
considerations of demand and cost.
(a) Software routines and files need not be
developed or distributed together to
achieve seamless integration of their
functions
131. Microsoft had significant discretion because whether different software
products are delivered by one or multiple groups of code is a matter of packaging rather
than engineering.
131.1. Software consists of a series of detailed instructions to the various
components of a computer. It is usually written in one of many specialized artificial
languages designed to be comprehensible to human programmers and then "compiled"
into a form that interacts directly with the hardware.
- Professor David Farber testified that a "compiler translates
instructions (written in a language efficient for programmers) into
the language understood by the computer hardware." Farber Dir. ¶
19.
131.1.1. The software code necessary to supply the functionality
of a modern application or operating system can be extremely lengthy and complex.
- Professor Farber testified: "Applications may be large, often
involving a very large number of routines." Farber Dir. ¶ 17.
- Microsoft estimates that the set of instructions that it calls
Windows 98 consists of approximately 18 million individual
lines of code. Allchin, 2/2/99am, at 41:12-17.
131.1.2. To make that complexity manageable, modern software is
usually written as a series of individual routines, ranging from a few tens to a few
hundreds of lines of code apiece, that perform specific functions. Large programs are
created by "knitting together" many such simple routines with higher level routines.
- Professor Farber testified that routines "typically contain a
few tens to a few hundreds of lines of code each." Farber
Dir. ¶ 13.
- Professor Farber testified that "each software product is built
up from simple low level routines that are then called by
routines at a higher level of composition. Routines at each
level are called by yet higher level routines until the desired
functionality of the end product is achieved. In this manner,
all software is built up layer by layer through the use of often
large numbers of routines, but each with limited complexity."
Farber Dir. ¶ 14.
- Joe Belfiore testified that "when we do good software
architecture, what we're able to do is to break what is a
complex and very full set of functionality into meaningful
components, each of which sort of can be self-contained
and can implement the job that it's supposed to do in a very
efficient way. And if you do a really good job of this, then
each of those separate components are very useful to other
parties that want to take advantage of them." Belfiore Dep.,
1/13/99, at 377:2-11.
131.2. As a result of its modular structure, modern software is extremely
malleable. Underlying routines can be packaged together in essentially any way that
the designer chooses.
- "As a result of this layering," Professor Farber testified, "software
has an inherently malleable and modular structure which gives
software developers broad freedom in combining (i.e., bundling)
different functions into software products." Farber Dir. ¶ 15.
- Glenn Weadock testified that software designers have great
flexibility "in how to combine the atomic units of code, called
subroutines or functions, to make up files (or "libraries") on disk . . .
. They can create a so-called 'monolithic' program that consists of a
single, large file; they can create a highly modular program that
uses a hundred different library files (called DLLs, for Dynamic
Linked Library) to contain a thousand different subroutines; or they
can choose any intermediate degree between these two extremes."
Weadock Dir. ¶ 29.
- Belfiore conceded that the organization of files into various DLLs
can be changed or designed with specific goals in mind. Belfiore
Dep., 1/13/99, at 153:23 - 154:4.
- Hadi Partovi testified that Microsoft has moved functions in one
DLL into different DLLs in succeeding versions of the product.
Partovi Dep., 1/13/99, at 659:7-23.
- Weadock testified: "A software designer with source code access
may choose to place an application subroutine into a file that
contains operating system subroutines . . . . Microsoft, for example,
has chosen to design Internet Explorer so that some of the code
that it uses co-resides in the same library files as other code
needed for Windows 98 or even Windows 95 to run." Weadock Dir.
¶ 30.
- Professor Edward Felten testified that SHDOCVW.DLL "is a great
example of the point I'm trying to make about packaging of
functions into files. This SHDOCVW file is really a bundle of
separate functions. It contains some functions having to do with
displaying the Start menu. It contains some functions that have to
do specifically with Web browsing, and it contains some general
user interface functions as well. And to talk about this file as doing
one thing or being part of one product is really incorrect." Felten,
12/14/98am, at 60:18 - 61:2.
- Professor Felten testified: "Due to the malleable nature of software,
functions may be moved from one DLL to another, or a single DLL
may be disaggregated into two DLLs." Felten Dir. ¶ 60.
131.3. Software routines need not reside in the same file to function
together in a perfectly seamless fashion. Except at the extremes, therefore, how a
software engineer chooses to organize routines into particular files is a matter of
packaging as opposed to engineering necessity.
- Professor Farber testified that "a software developer is free (subject
to minimal limitations of no relevance here) to change the
partitioning of routines among files at any time without changing
their function or correct operation when the files are combined
during execution in an end user's computer. Thus, there is
generally no technical reason why a particular routine must be
included in the same file with another routine so long as the
routines are appropriately compiled and linked in the end user's
computer." Farber Dir. ¶ 18.
- Professor Felten testified that as part of the transition from Internet
Explorer 4 to Internet Explorer 5, Microsoft split SHDOCVW.DLL
into two parts, SHDOCVW.DLL and a new file called
BROWSEUI.DLL. He also testified that Microsoft moved some
code from SHDOCVW.DLL into SHELL32.DLL. Felten, 6/10/99am,
at 49:4-15.
- William Harris testified that software development "is inherently
flexible. There are numerous ways in which to design a program to
achieve the same functionality or effect. Similarly, a developer can
combine, or separate, any two or more software products or
components. It is typically the goal to combine such products or
components in such a way as to appear 'seamless' to the user, in
other words to make the two products appear like one. A good
example of this is what Intuit has done with Quicken and Internet
Explorer." Harris Dir. ¶ 82.
131.4. Files of software routines need not be shipped, or even designed,
together to achieve seamless integration of their functions. Whether the necessary files
are shipped together with the operating system, installed by an OEM prior to selling a
computer in the retail channel, or accumulated by the end user through the purchase of
separate products from multiple companies, the functionality ultimately delivered to the
consumer can be exactly the same.
- Professor Farber testified: "All the routines that are called directly
or indirectly by a program should be available when the program is
being used. But whether those routines originate from one
particular software program or another is irrelevant to the
performance of the functions, so long as the software is written and
installed such that the programs work together." Farber Dir. ¶ 18.
- Professor Farber also testified that "software has an inherently
malleable and modular structure which gives software developers
broad freedom in combining (i.e. bundling) different functions into
software products. This malleability also gives a software
developer two related types of design freedom: (1) to integrate two
separate cd-roms because the functions on one particular cd-rom
can be integrated by an OEM or retail end user with functions on
another cd-rom and (2) to determine which functions to include
within software sold as one product and which to separate and sell
as a different product, whether produced by the same or a different
software developer, for installation and use together by the a retail
end user." Farber Dir. ¶ 15.
- Professor Felten testified: "The mere fact that two functions are
implemented in the same file, or that two products are 'integrated'
into a single product, does not imply that they must be
implemented in this fashion; because of the nature of software,
functions can be separated into distinct files, or 'integrated'
products can be separated into distinct products without any loss of
capability." Felten Dir. ¶ 31.
- Professor Felten testified: "In some cases in Windows 95 and IE 4,
functions used in Web browsing and functions unrelated to Web
browsing are implemented in the same program file. That these
functions are implemented in the same file does not mean that
these functions are inevitably intertwined." Felten Dir. ¶ 31.
- Professor Felten testified that Windows Explorer is configured to
allow other, entirely separate applications to display information in
its embedded subwindows. "This 'Active Documents' specification
that Microsoft has released allows anyone to write a piece of
software that can display anything in an embedded subwindow like
this. And so, one of the points to make about this is that the fact
that a completely separate application like Microsoft Word or like
some ISV application can display something in that embedded
subwindow, does not imply that Microsoft Word or that ISV
application is part of Windows Explorer. It just says that it can
display something inside that window frame that Windows Explorer
puts up." Felten, 12/14/98pm, 50:4-14.
- Weadock testified that Novell designed "an HTML-based help
system that is -- that works with multiple browsers. It works with
Navigator. It works with Internet Explorer." Weadock, 11/17/98am,
at 48:14 - 49:5.
(b) Although recognizing it could have chosen
a different approach, Microsoft made
Internet Explorer 3 and Windows difficult to
separate and offered only a bundled
version to OEMs and end users
132. Although it recognized its ability to choose other approaches, beginning
with Internet Explorer 3 Microsoft placed in the same files the routines that supply both
operating system functionality and web browsing functionality, and Microsoft refused to
give OEMs the option of combining the two products themselves.
132.1. The OSR 2.0 release of Windows 95, released in August 1996,
updated DLLs that supplied both web browsing and other functions.
- Microsoft Vice-President David Cole testified that "Internet Explorer
3.0 is an integral part of the OEM Service Release 2.0 (or 'OSR
2.0'). . . . OSR 2.0, which was first made available to computer
manufacturers in August 1996, includes a wide range of product
enhancements in addition to Internet Explorer 3.0, such as support
for larger hard drives, improved multimedia support, a variety of
networking enhancements, new power management features, and
many others." Cole Decl. ¶¶ 41-42 (DX 2220).
- Weadock testified that a software developer "may choose to place
an application subroutine into a file that contains operating system
subroutines . . . . Microsoft, for example, has chosen to design
Internet Explorer so that some of the code that it uses co-resides in
the same library files as other code needed for Windows 98 or
even Windows 95 to run." Weadock Dir. ¶ 30.
- Professor Felten testified that there "is other software code specific
to IE web browsing that could be deleted from these shared
program libraries." Felten Dir. ¶ 58.
- Felten testified that SHDOCVW.DLL "is really a bundle of separate
functions. It contains some functions having to do with displaying
the Start menu. It contains some functions that have to do
specifically with Web browsing, and it contains some general user
interface functions as well. And to talk about this file as doing one
thing or being part of one product is really incorrect." Felten,
12/14/98am, at 60:15 - 61:2.
132.2. Microsoft did not offer OEMs a version of Windows without web
browsing. Microsoft refused to do so despite the fact that it offered Internet Explorer 3
separately to end users in a way that, when combined with an earlier version of
Windows 95, supplies precisely the same functions.
- When asked whether he considered the retail version of Internet
Explorer 3 to be "integrated" with Windows 95, once installed by
the customer, Carl Stork answered: "Once it's installed, I consider it
to be integrated. . . . It was developed much as we develop our
operating system upgrades that the end result would be an
integrated whole. And it's just a question of how it's delivered."
Stork Dep., 8/11/98, at 53:18 - 54:14 (DX 2594).
133. In addition to offering OEMs Windows only with Internet Explorer already
installed, Microsoft prohibited OEMs by contract from removing any aspect of the
browser from the Windows software package.
- See supra Part V.B.2.c.; ¶ 127.
134. Because OEMs have no viable commercial alternative to Windows,
Microsoft's refusal to offer an unbundled option coerced OEMs into licensing the
browser as a condition of licensing Windows.
- See supra Part II.A; ¶ 15; Part V.B.2.c; ¶ 130.
(3) Microsoft similarly tied Internet Explorer 4 to
Windows
135. Microsoft also used its monopoly power to force OEMs to license and
distribute Internet Explorer 4 as a condition of licensing Windows.
135.1. Microsoft initially offered Internet Explorer 4 to OEMs in
September 1997 on a separate disk from Windows 95 and gave OEMs the option of
licensing Windows 95 without it. Microsoft recognized that Internet Explorer 4 could be
distributed separately from Windows and that, once added to the system by OEMs or
end users, it would have the same functionality as if it had been bundled with Windows
in the first place.
- A December 11, 1997 letter from Microsoft to OEMs notes that
Microsoft had initially shipped IE 4.0 to the OEMs as part of a
"supplemental release of Updated Windows Features" in late
September. GX 1064, at MS6 6013683.
- When asked whether there were any ways in which installing the
version of Internet Explorer 4 available on the web would result in a
different experience for the user in any way, as compared with
receiving Windows 95 and Internet Explorer 4 "integrated" on a
new computer, Stork answered "I'm not aware of any." Stork,
8/11/98, at 48:9 - 52:24 (DX 2594).
135.2. By December 1997, however, Microsoft retracted that option and
instead required OEMs to license and install Internet Explorer 4 as a condition of
licensing Windows 95.
- GX 418 (Microsoft business terms with Toshiba) (sealed); GX 410
(Microsoft business terms with Digital Equipment Corp.) (sealed);
GX 538 (Microsoft business terms with Packard Bell NEC, Inc.)
(sealed); GX 625 (Microsoft business terms with Micron
Electronics, Inc.) (sealed); GX 588 (Microsoft business terms with
Gateway 2000 Inc.) (sealed); GX 697 (Microsoft business terms
with Sony Corporation) (sealed); GX 1059 (Microsoft business
terms with Hitachi LTD) (sealed).
- In May 1997, David Cole wrote to Paul Maritz and Moshe Dunie
that "The overriding priority is getting market share up. Getting IE4
into memphis is part of that." GX 626.
(4) Microsoft also tied the browser to the operating
system by refusing to license OEMs, and refusing
to permit OEMs to offer their customers, Windows
with Internet Explorer "uninstalled"
136. Although Microsoft designed Internet Explorer and Windows to be difficult
to separate and forced OEMs to license the combined product, it nonetheless provided
a ready means for users to remove or "uninstall" the browser. But Microsoft denied
OEMs the option of licensing Windows 95 with Internet Explorer uninstalled and
prohibited OEMs from offering such a version of Windows to their customers.
(a) Microsoft configured Internet Explorer to
"uninstall" in response to demand for
Windows without Internet Explorer
137. Microsoft configured both Internet Explorer 3 and 4 to "uninstall" from
Windows 95 through the "Add/Remove" control panel.
137.1. The "uninstall" feature removes the end-user's ability to browse
the web with Internet Explorer but does not adversely affect other software installed on
the computer.
- Professor Felten testified that Internet Explorer 3 can be removed
from Windows 95 through Microsoft's "Add/Remove" control panel.
Felten Dir. ¶¶ 23-24 (explaining the process); see also GX 1202
(videotaped demonstration of that process). Professor Felten also
testified that that process "has no apparent effect on the non Web
browsing functions" of Windows 95. Felten Dir. ¶ 27.
- A Microsoft technical support article entitled "Cannot Uninstall
Microsoft Internet Products in OSR 2" describes a two-step process
for removing IE 3 from OSR 2 using the Add/Remove Programs
Control Panel. The article does not describe any adverse effect of
the removal of Internet Explorer 3 on any non browsing
functionality provided by Windows 95. GX 1367.
- Professor Felten testified that Internet Explorer 4 can also be
removed from Windows 95 via the "Add/Remove" control panel.
That process causes the system to revert to the previous version of
Internet Explorer installed on the system (or, in the case of an OEM
version of Windows 95 originally shipped with Internet Explorer 4,
to Internet Explorer 3). Internet Explorer web browsing can then be
entirely removed from the system by following the "remove"
procedure for that earlier version. Felten Dir. ¶ 32; GXs 165, 166,
and 172 (Microsoft Knowledgebase articles describing that
process).
- After performing experiments on versions of Windows 95 and IE,
Glenn Weadock concluded: "Two practical methods exist of
removing Internet Explorer 3.02 from a Windows 95 machine. One
can run the Microsoft-supplied deinstallation program to effectively
disable the user's ability to run the web browser program, while
leaving enhancements to operating system files in place. (Note
that this option is apparently not available to OSR2 users, but its
effects can be simulated by reinstalling Internet Explorer 3.02 using
the downloadable version from Microsoft's Web site, and then
deinstalling as one would do on an earlier version of Windows 95.)
Alternatively, one could revert the system so that it contains the
original versions of the DLL files that the Internet Explorer 3.02
installation enhances. Either method results in a fully functioning
Windows 95 system, although the second method may result in the
removal of enhancements that some application software vendors
may use for their convenience in providing Internet-related features
to users." DX 1715, at ATR 22876.
- Microsoft's Allchin conceded that there were "a variety of ways" to
remove Internet Explorer from Windows 95. Allchin, 2/2/99pm, at
4:21-24.
137.2. Microsoft configured Internet Explorer to "uninstall" in response to
demand for an operating system without Internet Explorer. Indeed, Microsoft
advertised to end users that the "Add/Remove" control panel could be used to remove
Internet Explorer from Windows 95.
- A Web page from Microsoft's website entitled "The IE Challenge"
encourages customers to install and use Internet Explorer 3.0, and
notes "IE Uninstalls Easily if you want to use a newer version, or
simply get rid of it (and so does Navigator!)." GX 352.
- Microsoft's Web site describes "How to Uninstall Internet Explorer
4.0," and lists situations in which a user might want to take that
action. GX 164; see also GX 165 (describing a different method);
GX 166 (describing how to "manually" uninstall Internet Explorer
4.0); GX 170 (Microsoft technical article describing how to uninstall
Internet Explorer 4.0 in Win95 and WinNT using Add/Remove); GX
172 (describing how to remove Internet Explorer 4.0 from Win95
using IEREMOVE.EXE).
- David Cole testified that Microsoft designed Internet Explorer 3 to
Add/Remove from Windows 95 because "users have given us
feedback that they would like choices about what they see on their
desktop and they don't see on their desktop, and in that particular
case we had feedback from corporate customers that wanted to
prevent access to the Internet, so when they -- they buy a new
machine from a PC manufacturer, they want the ability to remove
easy access to the Internet so their employees, you know, aren't
spending their time out on the Web doing whatever." Cole then
testified that having Add/Remove capability addressed that concern
by removing "the obvious user-accessible means to run -- to run
Internet Explorer." Cole Dep., 1/13/99, at 395:7 - 396:6.
(b) "Uninstalling" Internet Explorer removes
the Internet browser product
138. Software products routinely share code. A single file -- in the case of
Windows 95 and 98 called a dynamic linked library -- may be used by many different
programs, regardless how the file originally came to be installed on the system.
138.1. An operating system like Windows makes shared code available
for use by all of the applications on the system. Microsoft has designed Windows so
that many of the files that perform basic functions, like drawing a window on the
monitor, can be used by third-party software applications.
- Professor Farber testified that software developers "write their
programs with the expectation that certain functions can and will be
performed by the operating system of the computer on which the
software will be used. The application invokes the operating
system by calling routines supplied as part of the operating system.
The interconnection is referred to as an application-programming
interface (API)." Farber Dir. ¶ 20.
- Professor Felten testified that "IE Web browsing uses some code
that is specific to IE Web browsing, and some code that is shared
(that is, it supports other functions in addition to IE Web browsing).
There is nothing unique about IE Web browsing in this regard:
virtually all PC applications make use of some application-specific
code and some shared code that ships with Windows." Felten Dir.
¶ 61.
- Professor Felten testified that "it's a mistake to say that because
code is invoked in some case, that code is specifics [sic] to what's
happening in that case. The code that detects key presses, for
example, is used by almost every application. And so if one
wanted to conclude that that code was part of the Web browser, I
suppose you could, but you would also have to conclude that it's
part of the personal finance package, it's part of the multimedia
player, it's part of the word processor and everything else. You
have to look a little bit more carefully than does this code get used
in executing this function or not." Felten, 12/14/98am, at 59:6-18;
Felten, 12/14/98pm, at 6:22 - 7:14 ("To use an example different
than the one I used this morning, another example, you would look
at the code which is able to draw windows in general, draw window
on the screen. That code is used by virtually every application.
And although it's used by every application, it's not really right to
say it's part of every application.").
138.2. Shared code is not, however, the exclusive province of operating
systems. Applications programs can, and routinely do, also share code with other
applications programs.
- Professor Felten testified: "When I use the word 'platform,' at least
all I mean is that this is software that offers API's -- software that
offers services to other software. And whether something is a
platform or not says nothing about whether it's part of the operating
system, whether it has to be shipped with the operating system, or
anything like that. I described before all the third-party products
would serve as platforms on Windows." Professor Felten also
testified that the availability of platform-level services saves work
for other software developers, "regardless of who offers that
platform service and regardless of whether it's packaged with the
operating system." Felten, 12/14/98am, at 52:13 - 53:5.
- Professor Felten testified that "many or most application programs
offer APIs these days and, of course, they are not part of any
operating system." Felten, 6/10/99am, at 53:25 - 54:2.
- Microsoft's David Cole conceded that "system services," defined as
"modules of code that provide function for other modules of code,"
can be found in any software package, not just operating systems."
Cole Dep., 1/13/99, at 390:2-14.
- Michael Devlin conceded that Rational's products call upon APIs
that are provided by Microsoft applications, such as Microsoft
Access, and even by third-party applications, in addition to APIs
provided by Windows. Devlin, 2/4/99am, at 41:6 - 42:3.
- Richard Schell testified that the fact that Internet Explorer contains
DLLs did not change his conclusion that it was an application.
"Well, there are two issues here. One is: Does the fact that it's
made up of DLLs make it not an application? And the answer to
that is no. Lots of applications consist of DLLs along with a main
program. I mean, you can pick up any application that's out there.
There are lots and lots of DLLs that come along with them. You
just go down the -- pick up Windows 95, go through the -- you
know, using the Explorer, you can find DLLs for every application,
so the fact that it has -- that there are DLLs that constitute the
application doesn't make it not an application. The fact that they
happen to be distributed with the operating system also doesn't
make it not an application. Microsoft can, does, has distributed
DLLs with the operating system that are helpers for other things,
and it's their choice to distribute those -- those things. The
unfortunate fact of the matter is that Microsoft as a monopolist
chooses what they distribute with the operating system whenever it
pleases them, and then they say, well, it's part of the operating
system." Schell Dep., 9/15/98, at 252:5 - 254:3 (DX 2587).
138.2.1. The various applications that comprise Microsoft's suite of
office productivity software, for example, share a great deal of code.
- Robert Muglia testified that Microsoft Office is "an integrated
package" including distinct applications known as Word and
Excel, which were "designed to be integrated" together into
Office but that Microsoft nonetheless distributes Word and
Excel separately. "The way I might say that is that Office is
an integrated package overall. It was designed to be
integrated. We produced, because our customers would like
us to produce it, a separate word-processing program that
we derived from the overall integrated Office package and a
separate spreadsheet program." Muglia, 2/26/99pm, at
67:17 - 70:3.
138.2.2. Java virtual machines are shipped in the Java runtime
environment with Java "class libraries" that are freely available for the use of anyone
programming in Java.
- James Gosling testified that Java virtual machines include a
collection of code called the Java classes, which provide
basic building blocks (or "APIs") that Java developers can
use in their programs. Gosling, 12/2/98am, at 47:14 - 49:10.
- Gosling testified that Sun and others also make additional
Java class libraries freely available to developers, who must
then ensure that they are present on the end user's
machine. Gosling, 12/2/98am, at 56:23 - 57:16.
139. Because applications may share code with each other and with the
operating system, when an application is added to Windows, the pertinent shared code
may or may not be loaded, and when the application is removed, shared code generally
is left behind.
139.1. When an application is added to Windows, it routinely checks to
make sure that all of the shared program libraries, or DLLs, that the application invokes
are present. Typically, if any of those DLLs are missing, or present in an outdated
version, the application will install them.
- Weadock testified that applications that change Windows DLLs are
common. Weadock, 11/17/98am, at 25:15 - 26:10.
- Boeing's Scott Vesey testified that "many applications do make
changes in the Windows system subdirectory." Vesey Dep.,
1/13/99, at 153:21-23 (DX 2596).
- Carl Bass, Chief Technical Officer and Vice President of
Engineering at Autodesk, testified that Autodesk's principal
product, AutoCAD, uses the WININET file included with Internet
Explorer 3.0 and 4.0 and that, "if the necessary version is not
present, or if the version of WININET on the user's PC is older than
the version included with AutoCAD, the program will install the
version of WININET that is bundled with AutoCAD." Fisher ¶ 165.
139.2. Conversely, it is well-recognized -- including by Microsoft -- that
shared DLLs should be left behind when removing software products from a
multiproduct system.
- Professor Felten testified that "leaving in place shared files that
perform other functions conforms to the ordinary way in which
software application programs are removed." Felten Dir. ¶ 57.
- Page 29 of Microsoft's Handbook for Applications suggests: "User
data files including the following should remain on the hard disk:
Resources that other programs might use, such as sharable DLLs,
sharable fonts, and sharable Registry entries. It is better to err on
the side of safety regarding other applications. If you are not sure
whether removing a DLL might harm other applications, it is better
to leave it behind." GX 431, at 29.
140. Accordingly, it is commonplace to describe a product as having been
"removed" from a multi-product system even when the shared code that is used and
distributed by that product and others remains behind.
- William Harris testified that "removing an application does not mean
removing all components of the application, in other words it does not
require deleting components shared by other applications. Quicken, like
most applications, utilize shared components of software code, or 'DLLs'.
Any time multiple applications share a DLL, and you remove one of the
applications and the shared DLL along with it, the other application will no
longer work properly. So for example, if Quicken called on a DLL that was
used by another application, like Internet Explorer, and an end user
removed Quicken and the shared DLL, the other application would not
work properly. This is easily avoided, though, by retaining the shared DLL
when removing or uninstalling an application." Harris Dir. ¶ 86.
- Professor Felten testified: "The code that detects key presses, for
example, is used by almost every application. And so if one wanted to
conclude that that code was part of the Web browser, I suppose you
could, but you would also have to conclude that it's part of the personal
finance package, it's part of the multimedia player, it's part of the word
processor and everything else." Felten, 12/14/98am, at 59:6-18; Felten,
12/14/98pm, at 6:22 - 7:14 ("To use an example different than the one I
used this morning, an example, you would look at the code which is able
to draw windows in general, draw windows on the screen. That code is
used by virtually every application. And although it's used by every
application, it's not really right to say it's part of every application.").
141. Because they share code with both the operating system and with each
other, software products commonly are defined -- including by Microsoft -- according to
the functionality they supply to the consumer, rather than by the code they distribute.
- Glenn Weadock testified that "both industry professionals and computer
customers think of a software product more as that which enables a set of
related features than as a collection of specific files. For example, when a
reviewer evaluates a software product in a computer magazine, the
reviewer typically focuses on the product's feature set . . . . The list of files
that come in the box, or the list of code units that those files contain, is
rarely if ever provided." Weadock Dir. ¶ 15.
- Weadock also testified that defining software products "as a particular
collection of files is ultimately impossible if code units within the same file
are shared, either by multiple applications or by a single application and
an operating system. . . . Attempting to define software strictly as a
collection of files is a fruitless exercise when some of those files perform
double duty in different contexts." Weadock Dir. ¶ 14.
- John Rose testified that "consumers want to purchase a personal
computer that allows them to view, communicate, or manipulate visual
graphics displayed on the personal computer's screen regardless of
whether the data or software code that responds to the manipulation
resides on the personal computer's hard-disk drive, a CD-ROM, or on a
computer that may be continents away. For basic features of the
computing experience, it is irrelevant to users whether the feature is
incorporated in application or operating system software." Rose Dir. ¶ 22.
- Jones described Internet Explorer for the Macintosh as "the thing that will
let [our customers] go and deploy and take advantage of the services on
the Internet and computing on the Internet." Jones also testified that the
Internet Explorer 5 package "contains a set of features that people can
use to browse the Web, that ISVs can target and ICPs can target." Jones
Dep., 1/13/99, at 555:18 - 556:7.
- Carl Stork testified that: "If you were to try to say the browser is just
viewing web pages, it really wouldn't be very interesting for end users
because the Internet is so much more than that" including
"communications plumbing, things like TCP/IP stacks, dial-up networking,
PPP. Proxy Server perhaps. Things like URL resolution, HTML
rendering, playing with various formats, whether it's things like active
server pages or ActiveX controls. Java outputs. Media streams.
Supportive protocols to send and receive e-mail. The ability -- possibly the
ability to transfer through things like FTP. I don't know if I mentioned the
ability to have Java applets. I mean for an Internet experience -- for
things to be attractive, things need to work seamlessly, which means you
need a broad stream of capabilities." Stork Dep., 1/13/99, at 759:10 -
760:8.
- Professor Felten testified that because there is a long code path
necessary to perform almost any function in a modern computer, "it would
be a mistake to say that because something is on that code path, it's
necessarily part of the application that the user is using." Felten,
12/14/98am, at 57:20 - 58:19.
- Weadock pointed out that "Microsoft's word processing software product,
Word for Windows, ships with the file COMCTL32.DLL, but that file is also
used by Windows 95." Weadock Dir. ¶ 14.
- Weadock testified that applications that change shared program libraries,
or DLLs, that are shipped with Windows are common. Such applications
include Norton Utilities and Microsoft Word. "I don't know anybody that
thinks that Microsoft Word, or Norton Utilities, or Microsoft Golf, or any of
these other various products that may include updated DLL's are part of
Windows. They are separate applications. The fact that an application
includes Windows DLL's or DLL updates does not make it therefore part
of the operating system." Weadock, 11/17/98am, at 25:15 - 26:10.
- Professor Felten testified that Windows Explorer is configured to allow
other, entirely separate applications to display information in its embedded
subwindows. But "the fact that a completely separate application like
Microsoft Word or like some ISV application can display something in that
embedded subwindow, does not imply that Microsoft Word or that ISV
application is part of Windows Explorer. It just says that it can display
something inside that window frame that Windows Explorer puts up."
Felten, 12/14/98pm, at 49:25 - 50:14; Felten, 12/14/98pm, at 50:15-25
("Q: And does the fact that other applications like Microsoft Word or,
perhaps, third-party ISV applications can use the embedded window as a
viewer to display things say anything about whether or not that application
is part of the operating system? A: No. Certainly, if it did, one would have
to conclude that from this picture that Microsoft Word is part of the
operating system, and we know that's not the case.").
142. As a result, whether a product, including Internet Explorer, is present on a
PC from the perspective of end users depends on whether its functionality can readily
be accessed, not on whether some of the code that is necessary to supply those
functions may be present.
142.1. It is common in the computer industry for the underlying code
necessary to employ a software product to be installed on a computer, but in a disabled
and unusable form. When the end user actually purchases the product, he or she then
receives an "activation key," or password, which enables the dormant functionality.
- Weadock testified that "it is possible, and sometimes a matter of
commercial practice, to have software that exists on a disk or PC in
the sense that its code modules are physically present, but does
not exist in any practical way from the user's standpoint because
the software is hidden, protected, or otherwise disabled." Weadock
Dir. ¶ 19 (collecting examples).
- Weadock testified that, as a variation on the same theme, software
is often promoted by making trial versions freely available for
download from the Internet. That software functions for a trial
period, but then disables itself unless the user purchases an
activation key. "Expired trialware or shareware physically exists on
the PC in terms of bits and bytes, but once expired, the program is
effectively absent until the user pays for it." Weadock Dir. ¶ 19.
- Phillip Barrett testified that the way Real Networks' products
"Player and PlayerPlus are related is basically there's one player.
PlayerPlus features are activated by a license key that one gets by
coming to our web site and going through a secure form and
purchasing that license key." Barrett Dep., 1/13/99, at 112:17 -
113:2.
142.2. Accordingly, it is commonly accepted in the industry that a
software product is not present on a particular machine unless the end user has access
to the functionality it supplies.
- Weadock testified: "The existence of a software product on any
particular PC -- that is, whether it is effectively present or absent
from the customer standpoint-depends on both the presence of the
software enabling the product's feature set, and the means to use
that feature set." Weadock Dir. ¶ 18.
- Weadock testified that corporate technical support managers
consider "inhibiting the user-accessible means of access to a
software product (e.g., an icon on the 'desktop' screen of the user
interface, or entries in menus of program options) has the same
effect, from the support cost standpoint, of removing a program in
its entirety . . . . Because removing the user-accessible means of
using a browser product makes the product disappear from the
perspective of the user, support costs are significantly reduced . . .
." Weadock Dir. ¶ 28b.
- Barrett testified that, although a user may have the bits of code that
implement the Player Plus functionality installed on their machine,
"From the user's perspective, what they have is the standard
player" until they pay for an upgrade key, because they are unable
to access the Player Plus features. Barrett Dep., 1/13/99, at
113:11 - 114:4 (GX 1450).
142.3. Thus, removing the ability to browse the Web using Internet
Explorer effectively removes the Internet Explorer product.
- Jon Kies testified that "if we provide" Windows 95 "without Internet
Explorer in the menu item, the customers feel like there's no
browser installed whether or not the actual code may exist below
the surface or the user interface." Kies Dep. (played 12/16/98am),
at 27:1-16.
- Professor Felten testified: "If you have removed the ability to
browse the Web, as far as the user is concerned, Web browsing --
the Web browser is gone." Felten, 12/14/98am, at 33:5-14.
- Professor Fisher testified that "a browser consists of the ability to
do the things I described. Now, to the extent that removing the
visible means of access removes that ability, I suppose one could
say that without the visible means of access, there isn't a browser."
Fisher, 1/6/99am, at 8:19-24. Fisher further testified that
Microsoft's tie of Internet Explorer and Windows 98 would "from an
economic perspective" "disappear" if "Microsoft removed all means
of accessing Internet Explorer code or software technology within
the Windows 98 product as Microsoft designed it." Fisher,
1/6/99am, at 9:21 - 10:4.
- Microsoft's Cole testified that, "at a minimum," a user who invokes
Microsoft's Add/Remove procedure to remove a software product
"might expect the visible aspects of the program or update or
whatever it happens not to be there anymore, so in appearance it
might be gone from the end user's perspective." Cole Dep.,
1/13/99, at 394:4-9 (GX 1465).
(c) Microsoft used its operating system
monopoly to deny OEMs the ability to
license or sell Windows with Internet
Explorer uninstalled
143. Although Microsoft provided, through the "uninstall" capability, a ready
means for users to remove Internet Explorer from Windows 95, Microsoft refused to
permit OEMs to obtain, or license to their customers, Windows with Internet Explorer
uninstalled.
143.1. Microsoft denied Gateway's request for a version of Windows from
which Internet Explorer had been uninstalled.
- In a letter to Gateway addressing Gateway's earlier statement, "We
want IE to have uninstall (for as much of the code as can be
removed without disabling the system)," Microsoft responds by
saying that "Internet Explorer technologies are an integral part of
Windows 98 and cannot be uninstalled . . . . Consequently, the
concept of an 'uninstall' lacks practical significance in this context."
GX 1073, at MS98 0204593.
143.2. Microsoft prohibited the OEMs from selling PCs with Windows
installed and Internet Explorer uninstalled.
- See supra Part V.B.2.c; ¶¶ 127, 129, 132, 135.
143.3. Microsoft also prevented OEMs from removing the Internet
Explorer icon or any other aspect of the browser.
- See infra Part V.C.1.a; ¶ 177; Part V.C.2.a.(1); ¶ 199.
e. Microsoft also tied Internet Explorer to Windows 98
144. Microsoft determined that it could better exclude Netscape both by
continuing its welding of browser and operating system and by making the products
further inseparable. It did so with Windows 98.
(1) Microsoft concluded that defeating Netscape
required it to tie its browser more tightly to the
operating system
145. In late 1996 and early 1997, Microsoft designed and tested, and
considered shipping, a version of Windows 98 that, like Windows 95, was merely
bundled with Internet Explorer 3 components, rather than more tightly "integrated" with
Internet Explorer 4.
- In December 1996, David Cole and his Internet Explorer development
team discussed "de-coupling" the Internet Explorer 4 browser from the
Windows shell. According to Cole, "After thinking about this for the past
couple of days, it's clear to me (and others) that we must de-couple the
Browser from ActiveDesktop and the shell integration features.
ActiveDesktop and the new shell UI must be a completely optional
component for users and corporations. The default is the IE 4 browser
without the shell enhancements. If the user installs the new shell, they will
have some things to learn and pay a performance price. By coupling
these together, I think the overall effort has suffered. We've got a
compromised new shell design that tries to be too Windows 95 shell
compatible in my view. We don't have HTML on the desktop because we
are worried about performance. But even in compatibility mode,
performance will degrade and there will be differences that could stall
adoption of the browser platform. . . . What I really want is a browser and
ActiveDesktop which do not change shell32 at all, or at most some
carefull [sic] hooks are added and we ship it everywhere. I don't
understand why most of ActiveDesktop can't be done without any shell
changes at all." A member of the development team responded that "It's
good to have a decision like this. We need to investigate hard what we
will loose [sic] if we don't update shell32.dll even in the full IE 4.0 install --
that's an option we've never considered. It will definitely simplify our
testing metrix and is a good way to cut dev/testing cost." GX 46.
- In March 1997, Jim Allchin reported to Paul Maritz on the status of
Internet Explorer 4 and Memphis and listed options, including "drop IE 4
from Memphis and NT 5. There is a strong push to do this. We are
wasting hundreds of people's time on builds that don't work, etc. Frankly,
we may have to do this anyway to make progress. If we drop it, then we
know we must either go out without IE 4 in the final or we have to be
honest in that both systems will take perhaps a half year slip because we
would have to fix the quality/performance/size later and go through beta
tests much later." GX 110.
- In March 1997, Megan Bliss asked Carl Stork whether "IE 4 and Memphis
are joined at the hip." Stork responded, "We do not have closure on the
issues below at present . . . IE 4 is not being developed as joined to
Memphis at the hip -- at present Memphis is an afterthought. It is not one
of the four main test platforms for IE4. We are being encouraged by the
IE4 team to release a Memphis Beta 1 with the old shell. We need to
rethink the plans & make sure we have a plan that makes sense. Today I
would not tell anyone that it is possible to ship an integrated IE4/Memphis
product in 1997." GX 160.
- Bill Veghte conceded that Microsoft considered shipping Windows 98 with
Internet Explorer 3 instead of Internet Explorer 4 because there was
OEM demand for hardware-related improvements (e.g., USB support) that
were ready for inclusion in Memphis prior to the time Internet Explorer 4
was ready. He also testified that they released outside Microsoft a
version of Memphis without Internet Explorer 4. Veghte Dep., 1/13/99, at
783:2 - 786:8 (GX 1477).
- Chris Jones also testified that Microsoft shipped a pre-beta version of
Windows 98 to hardware manufacturers that did not have Internet
Explorer 4. Jones Dep., 1/13/99, at 536:8 - 537:3.
- Jonathan Roberts wrote to Allchin, Dunie, and Stork in March 97 to
discuss options regarding the proposed bundling of Internet Explorer 4
and Windows 98. One of those options was to "De couple Memphis and
IE and ship Memphis in July/August and connect with IE in the OEM
channel when it ships." Carl Stork responded that "Currently IE4 is so
immature (and big & slow & compat-bug prone) that it is impending our
self hosting process. We find tons of bugs but so many are in IE
components that our test & repro efforts are becoming meaningless on
the OS. We are also finding more and more resistance on the team to
install the builds because things don't work. I am at the stage where I do
not recommend that we release anything with IE4 integrated under the
name of a Memphis beta. Customers would experience too many
problems and the performance would be unacceptable as well - it would
be so bad as to blemish the reputation of Microsoft and of Windows . . . .
More importantly, at this point it is getting in the way of valid development
testing & repro work for Memphis." Roberts summed up the exchange
with the following: "I'm depressed. I wasn't aware things were so bad
with IE. This makes the following trade-off very painful, Hardware support
for Spring machines and some TCO benefits vs IE penetration." GX 355,
at MS7 003001.
146. Microsoft eventually concluded, however, that in order to win the "browser
war" it needed to create a stronger tie between Internet Explorer and Windows than its
OEM licensing practices achieved with Windows 95. The contemporaneous documents
show that Microsoft's decision to further bind Internet Explorer and Windows 98 was
driven, not by the technical merits of any such integration, but instead by a strategic
desire to drive up Internet Explorer's market share vis-a-vis Netscape Navigator.
- In December 1996, James Allchin wrote Paul Maritz an e-mail entitled
"concerns for our future": "Ensuring that we leverage Windows. I don't
understand how IE is going to win. The current path is simply to copy
everything that Netscape does packaging and product wise . . . Maybe
being free helps us, but once people are used to a product it is hard to
change them. . . . My conclusion is that we must leverage Windows more.
Treating IE as just an add-on to Windows which is cross-platform losing
our biggest advantage -- Windows marketshare. We should dedicate a
cross group team to come up with ways to leverage Windows technically
more . . . We should think first about an integrated solution -- that is our
strength." GX 47; GX 655, at MS7 003375 (one of the "Objectives for
Memphis Release" is to "provide ship vehicle for strategic internet
components").
- On January 2, 1997, Allchin wrote to Maritz: "You see browser share as
job 1. . . . I do not feel we are going to win on our current path. We are
not leveraging Windows from a marketing perspective and we are trying to
copy Netscape and make IE into a platform. We do not use our strength -- which is that we have an installed base of Windows and we have a
strong OEM shipment channel for Windows . . . . I am convinced we have
to use Windows. This is the one thing they don't have. . . .We have to be
competitive with features, but we need something more: Windows
integration. . . . If you agree that Windows is a huge asset, then it follows
quickly that we are not investing sufficiently in finding ways to tie IE and
Windows together." "Memphis must be a simple upgrade, but most
importantly, it must be killer on OEM shipments so that Netscape never
gets a chance on these systems." GX 48.
- Maritz responded to Allchin's January 2, 1997 email (GX 48) by agreeing
"that we have to make Windows integration our basic strategy" and
proposing that Microsoft hold the release of Memphis (Windows 98) to
"sync" with IE4. GX 49. Allchin agreed to that plan, arguing that instead
of "letting people think about whether they should choose
Nav/Communicator vs. IE," Microsoft "should move the argument to
Windows (Memphis and NT 5.) and drive it because of ZAW [Zero
Administration Windows], etc. as the reason to use IE." GX 50. Allchin
also argued that integration "is the only thing that makes sense, even if
OEMs suffer." GX 50.
- In a January 1997 internal MS presentation on the "NC and Java
Challenge," in a section called "response summary," the first bullet point is
"Increase IE share - integrate with Windows." GX 51, at MS7 005536.
- In an email to Bill Gates and Paul Maritz on February 18, 1997, Allchin
wrote that "I am convinced the path we're on is the wrong one. We are
playing into Netscape's strengths and against our own. . . . We focus
attention on the browser battle where we have little market share instead
of focusing the battle at integrating things into Windows where we have
market share and a great distribution channel." GX 354.
- Christian Wildfeuer, reporting on the result of a focus group study in
February 1997 of the upcoming release of Windows 98, observed that
most of the study group were "Navigator users. They said they would not
switch, would not want to download IE 4 to replace their Navigator
browser. However, once everything is in the OS and right there,
integrated into the OS, 'in their face' so to speak, then they said they
would use it b/c there would be no more need to use something
'separate.' The stunning insight is this: To make them switch away from
Netscape, we need to make them upgrade to Memphis. . . . It seems clear
that it will be very hard to increase browser market share on the merits of
IE 4 alone. It will be more important to leverage the OS asset to make
people use IE instead of Navigator." GX 202, at MS7 004343.
- Jonathan Roberts wrote to Allchin, Dunie, and Stork in March 1997 to
discuss options regarding the proposed bundling of Internet Explorer 4
and Windows 98. Roberts framed the issue as a "trade-off between
ensuring we have new device support in the OEM channel for the Spring
line of machines and generating twenty or so million more dollars in RUP
upgrades versus driving IE 4.0 penetration and a simpler customer
proposition. Based on my understanding of the company priorities, we
should opt for the plan of record and keep them synced . . . Hold Memphis
for IE 4.0 and ship in August-December. Pros: This is absolutely the best
way to drive IE 4.0 penetration. Customer feedback, including that from
over 200 folks in over 15 focus groups, indicates that people want the two
to be tied together. If they are de coupled, then Navigator has a good
chance of winning. In a browser battle, victory will go to the incumbent."
GX 355, at MS7 003000.
- In a January 7, 1997 e-mail to Allchin, Maritz argued that Microsoft should
hold Windows 98 for IE 4 even if it pushed the release date back to
August or September. "The major reason for this is to combat Netscape.
We have to position the browser as going away, and do deeper
integration on Windows. The stronger way to communicate this is to have
a new release of Windows and make a big deal out of it. We will thus
position Memphis as Windows 98. IE integration will be the most
compelling feature of Memphis." GX 53; see also Allchin, 2/3/98pm, at
27:12-17 (agreeing with Maritz's email).
- Kumar Mehta reported in March 1997 that "based on all the IE research
we have done" his feeling "is that it is a mistake to release memphis
without bundling IE with it." Because "IE users are more likely than other
browser users to get it with their computers, . . . effectively we would be
taking away the distribution channel of almost a quarter of all IE users."
Moreover, "80% of those who do not use IE say they have no plans to
switch to it. Which means that if we take away IE from the o/s, most nav
users will never switch to us. Also from all our research with IS and web
professionals we know that they eventually expect us to win the browser
war because Ie will be bundled with the operating system and they will
have no real reason to purchase navigator." Jonathan Roberts responded
that Mehta's report "validates why it is important to keep IE with
Memphis." GX 205; GX 736, at MS98 0128504.
- In an e-mail to Allchin on March 20, 1997, Roberts wrote that "Internet
Explorer has a much stronger chance of winning once it is integrated into
the operating system. An integrated browser makes Netscape a nonissue
-- a superfluous product for all but the most committed Netscape user."
GX 355, at MS7 003002.
- In June 1997, Chris Jones sent a memo to Bill Gates entitled "How to get
to 30% share in 12 months." Among other things, Jones wrote: "We will
bind the shell to the Internet Explorer, so that running any other browser is
a jolting experience." GX 334, at MS98 0104683.
- In July 1997, Microsoft executive Moshe Dunie, commenting on a
proposal to stop shipping the Windows 98 shell with Internet Explorer 4
after the release of Windows 98, noted that such a proposal "would
certainly increase significantly Win98 upgrade sales. I know there is the
browser share counter argument ... But it is an intriguing thought..." He
received the following response from Paul Maritz: "It is tempting, but we
have to remember that getting browser share up to 50% (ore more) is still
the major goal." (ellipses in original). GX 113.
- In December 1997, Allchin wrote to Cole, Dunie and other executives that:
"We have several goals from my perspective as a company -- no matter
where the work is done. That is why this is tough. We have to continue to
win against Netscape on the browser. This means that we need to
consider downlevel and xplatform solutions. In addition, it is possible
(although that is yet to be proven to me) that we might have to ship more
frequently than once per year. And at the same time we need more
integration with Windows -- both technically and marketing-wise. We
need that for business reasons (ignoring the perception issue of the DOJ).
I see this as critical. This is a hard balance, but I feel that we need to
slant things much more toward Windows while we still accomplish the
other goal against Netscape." GX 480.
(2) To accomplish this objective, Microsoft made the
browser and the Windows 98 operating system
more difficult to separate by, among other things,
eliminating the "uninstall" capability and
hindering users from making other browsers the
default
147. To achieve its objective of further impeding browser rivals, Microsoft made
Internet Explorer 4 and Windows 98 more difficult to separate.
147.1. With Windows 98, Microsoft continued to supply Windows and
Internet Explorer in a form in which the underlying web browsing routines and other
routines have already been combined in the same DLLs.
147.2. The only functional difference between Windows 98, on one hand,
and Windows 95 combined with Internet Explorer 4, on the other hand, is a few features
that Microsoft easily could separately supply and which can now be obtained by
combining Windows 95 and Internet Explorer 5.
- See infra Part V.B.3.c.(1).(b); ¶¶ 159-161.
147.3. There are, however, other differences between Windows 95 and
Windows 98. Among other things, Microsoft eliminated the end user's ability to
"uninstall" Internet Explorer from Windows 98, despite retaining the uninstall option for
numerous other features.
- Professor Felten testified that, although the Web browsing
experience in Windows 95 OSR 2.5 and Windows 98 is very
similar, Microsoft does not provide a mechanism for removing
Internet Explorer Web browsing from Windows 98. Felten Dir. ¶¶
35-37, 52.
- GX 1366 is a series of screen shots of Windows 98's
"Add/Remove" function, showing dozens of functions that can be
added or removed by the user, including, among other things,
internet tools, desktop wallpaper, mouse pointers, dial-up
networking, virtual private networking, and hyper terminal. Internet
Explorer is not one of them.
- James Allchin testified that Microsoft provides a ready means of
removing many files and features that Microsoft considers to be
"integrated" features of Windows (Allchin, 2/2/99pm, at 5:2-5) such
as the TCP/IP stack (Allchin, 2/2/99pm, at 7:12-15) and dial-up
networking, Netmeeting, and the personal web server (Allchin,
2/2/99pm, at 10:3 - 11:11).
- When asked to estimate "how many of the components of Windows
98 can be readily removed by procedures that Microsoft makes
available," Allchin testified that "the number is going to be quite
high, if you consider all of the approaches for, you know, which
drivers or file systems you're using and everything. So, you know,
one of the great things about Windows is it's so configurable . . . ."
Allchin, 2/2/99pm, at 11:12-22.
147.4. Microsoft was well aware that its customers wanted the ability to
remove web browsing functionality from Windows 98 but nonetheless chose to
eliminate that feature in order to force adoption of Internet Explorer.
- Gateway specifically requested that Microsoft provide a way to
uninstall Internet Explorer from Windows 98, in part because it was
"concerned that the installation of the full MS product (including
channels) results in a much slower system performance if the
customer chooses an alternate browser after full installation on
IE4." Microsoft refused. GX 1073, at MS98 0204593 (4/24/98
letter from Microsoft to Gateway).
- . In response to a CID, Gateway stated that:
- redacted -
GX 652, at ATR 30008 (sealed).
- Joe Belfiore testified that he was concerned that the omission of an
add/remove option for Internet Explorer in Windows 98 would
create a "customer satisfaction issue," in part because some users
would have applications that were incompatible with Internet
Explorer 4. Belfiore Dep., 1/13/98, at 366:8-11.
- OEMs uniformly believed that they had no choice but to license
Windows 98. Ransom testified that Packard Bell must "pre-install
100 percent of its consumer machines with Windows 98" because it
is "the only viable choice." Ransom Dep. (played 12/16/98pm), at
68:25 - 69:5; see also supra Part II.A; ¶ 15.1 (collecting similar
testimony from Compaq, IBM, Gateway, and Hewlett-Packard
executives). In addition, Microsoft's licenses for Windows 98 forbid
OEMs from removing Internet Explorer or its icon. GX 1190
(sealed); GX 660 (sealed); GX 458 (sealed); see also infra Part
V.C.1.a; ¶ 177 (collecting cites to other OEM licenses). Thus,
OEMs were forced to distribute Internet Explorer.
147.5. Microsoft also designed Windows 98 to override the user's choice
of default browser in certain circumstances.
- Professor Felten testified: "In all versions of Windows released
prior to OSR 2.5, the Default Browser is activated whenever the
user asks to initiate Web browsing." Felten Dir. ¶ 50.
- Professor Felten also testified, however, that in some cases
"Windows 98 uses IE 4 Web browsing even if the user has
specified another browser as the Default Browser. There are
several situations in which this can occur. First, when the user
initiates the Web Help function described in paragraph 36, and
chooses the option of clicking on the 'Support Online' link, the
system will always initiate IE Web browsing, instead of launching
the Default Browser, to go to the Support Online Web site.
Second, certain menus in Windows Explorer contain URL
Shortcuts created by Microsoft. A user who selects the 'Home
Page' or 'Search the Web' URL Shortcut found in the 'Go' menu in
Windows Explorer will always initiate IE Web browsing to go to the
particular Web site, rather than launching the Default Browser.
Third, if a user places a Web page on the Active Desktop, and then
clicks on a Web link on that page, this action will again initiate IE
Web browsing even if the user has designated another browser as
the Default Browser. Finally, Windows Explorer allows a user to
type into the Address Bar a command to search the Web. Typing
the word "Go" followed by a phrase or word that a user wants to
search for on the Web will initiate IE Web browsing to display the
response to the search request, regardless of the user's choice of
Default Browser. This is an example of what the industry refers to
as "hard-coding," in this case, forcing the use of IE Web browsing."
Felten Dir. ¶ 51.
- Both Professor Felten and Professor Farber described the
significant problems for PC users created by Internet Explorer
ignoring or overriding their choice of Navigator as their default
browser. For example, Professor Farber testified that "the way
Microsoft packages and distributes the Internet Explorer caused
real problems. My personal experience, I think, is a good example
of that. I tried to use Netscape. I keep trying to use it and,
periodically, I install it and I keep seeing IE pop up in funny places
and interfere with it. And so as a product, it is very difficult to use.
I am not a person that wants to use multiple browsers. I focus on
one, like I focus on one word processor. It's just too difficult to use
one and then suddenly when error occurs, you're faced with
another one." Farber, 12/9/98am, at 53:3-16; Felten, 12/14/98am,
at 27:11-19, 29:11-17; Felten, 12/14/98pm, at 14:7-11.
3. There is no technical or economic justification for Microsoft's
tying of Internet Explorer and Windows
148. Microsoft contends that its forced licensing of Internet Explorer is justified
by numerous benefits that depend either on what it calls its "integrated" design or its
contractual restrictions. But contemporaneous evidence shows that Microsoft's
conduct was motivated by a desire to thwart rivals and protect its operating system
monopoly rather than to benefit consumers, and other evidence demonstrates that
Microsoft's justifications are pretextual.
a. Microsoft's "welding" of its browser thwarted the
substantial demand for Windows without an Internet
browser
149. As an initial matter, Microsoft's refusal to supply either Windows 95 or
Windows 98 without web browsing, and its contractual prohibition on OEMs supplying
such a product, thwarted consumer demand for a browserless OS.
- See supra Part V.B.1.b.(4); ¶ 111.
150. This reduced the value of Windows to customers who preferred a
browserless operating system. Indeed, as explained in detail below, Microsoft's tying
arrangement inflicted on a number of customers substantial inefficiencies and
consumer harm.
- Dr. Warren-Boulton testified: "Even if Internet Explorer is preferred by
some users, it is not preferred by all users. Consumer welfare is
maximized when the market is responsive to consumer demand, not when
a firm with monopoly power over one product requires purchasers also to
take an unwanted product or makes it difficult or costly for them to obtain
a related product they desire." Warren-Boulton Dir. ¶ 158.
b. There is no technical reason for Microsoft's refusal to
meet demand for Windows without web bowsing
151. There is no reason -- other than its campaign to protect its operating
system monopoly through weakening browser rivals -- for Microsoft's steadfast refusal
to meet the demand for Windows without Internet Explorer. Microsoft easily could have
offered, or permitted OEMs to offer, the option of Windows 95 or 98 without web
browsing.
(1) Microsoft easily could have supplied Windows 95
without web browsing
151.1. First, no technical reason can explain Microsoft's refusal to license
Windows 95 without Internet Explorer 1 or 2.
151.1.1. The version of Internet Explorer (1.0) that Microsoft
included with the "plus pack" and the original OEM version of Windows 95 was a
separate, executable program file supplied on a separate disk. Web browsing thus
could be installed or removed without affecting the rest of Windows 95's functionality in
any way. The same was true of Internet Explorer 2.0.
- Professor Felten testified, with respect to Windows 95, that
he has "determined that removing IE1 from this version of
Windows is easily accomplished by removing the IE1
program file (sometimes called an 'executable') and
removing any icons on the Windows desktop and Windows
Start menu items that refer to IE1. After doing this, a user
cannot browse the Web without adding more software to the
system, but the functionality of the operating system is
unaffected." Felten Dir. ¶¶ 21, 22 (same for IE2).
- In a communication directed to OEMs on July 3, 1995,
Microsoft indicated that it had "decided" to include Internet
Explorer (among other things) in the OEM release of
Windows 95. Microsoft acknowledged that it would have
been "possible for the OEM to integrate these tools into their
manufacturing process themselves," but Microsoft said that
it was pre-installing Explorer to "save each OEM the time
and effort" that would require. GX 36.
- Glenn Weadock testified: "The operating system doesn't
need a browser to work, as Microsoft showed, when it
released the original retail version of Windows 95, which, as
we discussed earlier, does not contain a browser."
Weadock, 11/16/99pm, at 92:16-22.
151.1.2. Microsoft, moreover, created an easy way to remove
Internet Explorer 1.0 and 2.0 from Windows 95 after they had been installed, via the
"Add/Remove" feature in the Windows 95 "Start Menu." This, too, demonstrates the
absence of any technical reason for Microsoft's refusal to supply Windows 95 with web
browsing.
- Professor Fisher testified that "Microsoft has argued that it
must force OEMs to take IE because the absence of IE may
undermine the quality of the operating system, to the
detriment of users. However, several facts contradict this
suggestion. For example, Microsoft provided ways to
remove IE in Windows 95--a function that would most likely
not have been provided if it led to a decrease in the quality
of the operating system." Fisher Dir. ¶ 159.
151.2. Second, there is no technical reason for Microsoft's refusal to
license Windows 95 to OEMs without web browsing, either by providing a version of
Windows 95 with Internet Explorer 3 or 4 uninstalled or by permitting OEMs to uninstall
Internet Exlorer 3 or 4.
151.2.1. Microsoft's decision to provide an "uninstall" procedure for
Internet Explorer 3 and 4 to end users, and to promote Internet Explorer on the basis of
that feature, shows that there was no technical or quality-related reason for refusing to
permit OEMs to use the procedure. Microsoft would not have permitted end users to
uninstall Internet Explorer, and consumers would not have demanded such an option, if
that process fragmented or degraded the other functionality of the operating system.
- Professor Fisher testified: "Microsoft provided ways to
remove IE in Windows 95 -- a function that would most
likely not have been provided if it led to a decrease in the
quality of the operating system." Fisher Dir. ¶ 159.
- Celeste Dunn of Compaq testified that when Compaq was
planning to remove the Internet Explorer and MSN icons
from the desktop, Microsoft tested Compaq's Windows
configuration and had not detected any technical problems.
Dunn Dep., 10/23/98, at 187:12-25 (DX 2566).
151.2.2. Microsoft's agreement in January 1998 to provide OEMs
an uninstall option also demonstrates that there was never any bona fide technical
justification for Microsoft's refusal to license Windows 95 with Internet Explorer
"uninstalled."
- Jon Kies testified that Packard Bell/NEC took advantage of
the January 1998 stipulated remedy to offer some of its PC
models without Internet Explorer Kies Dep. (played
12/16/99am), at 6:11-19.
- Professor Fisher testified that "OEMs would not negotiate to
remove IE if the operating system would be adversely
affected, since a poorly operating computer would reflect
poorly on the OEM and would be likely to increase the
number of customer support calls; also, large customers
would not request an operating system with IE removed if
they felt this system would be adversely affected." Fisher
Dir. ¶ 163.
(2) Microsoft easily could have supplied Windows 98
without web browsing and enabled OEMs and
users to "uninstall" the browser
152. As with Windows 95, there is no technical justification for Microsoft's
refusal to meet demand for a browserless version of Windows 98.
153. As Professor Felten demonstrated, Microsoft could easily supply a version
of Windows 98, without the ability to browse the web, to which users could add the
browser of their choice. In fact, Professor Felten's prototype removal program,
although only a concept program designed in a relatively short period of time and
without the benefit of Microsoft's internal expertise, produces precisely that result when
run on a computer with Windows 98 installed.
- Professor Felten testified that his "analysis demonstrates that it is possible
for Microsoft to divide Windows 98 into two programs, one that replicates
the function of the current version of Windows 98 except that Web
browsing is removed, and another that adds IE 4 Web browsing to the first
program, such that an OEM or user who installed the two programs in
sequence would end up with software functionally identical to today's
version of Windows 98. Microsoft, with its intimate knowledge of its own
products, would have little difficulty performing this task." Felten Dir. ¶ 66.
- Professor Felten also testified that his "prototype removal program
removes Internet Explorer. It removes the ability to browse the Web, and
it prepares the machine to accept the installation of another web browser.
So, if you're in that state where IE Web browsing has been removed and
nothing has been put in its place, then all of the Web-browsing functions,
features are not there; and, in particular, the ability to display a Web page
inside an embedded subwindow is gone . . . ." Felten, 12/14/98pm, at
46:14 - 48:2.
- Professor Felten testified that his "programs demonstrate that Microsoft
can deliver a version of Windows 98 from which the IE web browser has
been removed, and they can deliver that in a way which does not affect
the non web-browsing functionality of Windows 98. . . . Microsoft can then
produce an IE installation program which puts the system back, in effect,
to what it is in today's Windows 98." Felten, 6/10/99am, at 9:4-12.
154. Professor Felten's program does not degrade the performance or stability
of Windows 98 in any way.
154.1. Professor Felten testified repeatedly and credibly that he had been
using a Windows 98 computer, on which his program had been run, for more than
seven months with no discernible loss in performance or stability.
- Professor Felten testified that "I should tell you that for seven and a
half months now I have been using a PC from which Internet
Explorer has been removed and Netscape substituted -- that's
since the 23rd of April -- on my primary desktop computer at work.
And since I'm a computer scientist, I use that machine pretty
intensely. I have see no problems in that time. My primary
desktop computer at home I have been using Windows 98 in the
same configuration with Web browsing removed and Netscape in
place since the midle of August. My testimony in this case was
written on that machine, and I have never seen a problem -- other
than the Windows Update issue which I described to you before,
your honor." Felten, 12/14/98pm, at 42:13 - 43:2; Felten,
12/14/98pm, at 43:15-20 (same); Felten, 12/14/98pm, at 52:1-17
(same).
154.2. Professor Felten also testified that he ran several performance
testing programs provided by Microsoft and discovered that removing Internet Explorer
from Windows 98 via the prototype removal program actually yielded a modest
performance gain.
- Professor Felten testified that: "Microsoft turned over to us a set of
ten performance measurement programs that they use for
measuring performance of various Microsoft software, in particular
measuring performance related to what Mr. Allchin calls the core IE
DLLs." Felten, 6/10/99pm, at 13:17-22.
- Professor Felten testified that, "in these performance tests, what
we found was on the whole, removing the Internet Explorer
browser from Windows makes Windows a little faster." Felten,
6/10/99pm, at 14:17-19.
- Professor Felten testified that "there were ten tests, and on one of
the tests there was no statistically significant difference between
the two systems. On six of the tests there was a performance
improvement due to removing Internet Explorer -- the Internet
Explorer browser. And on three of the tests there was a slight
performance slowdown due to removing the IE web browser. And I
want to point out the three slowdowns are considerably smaller
than the six performance improvements on the other test. So, on
the whole, what we see is a slight performance improvement due to
removing IE." Felten, 6/10/99pm, at 15:2-13.
- Professor Felten testified that Windows 98 uses less dynamic
memory (RAM) after the prototype removal has been run, which
has a positive effect on system performance. Professor Felten
testified that "with the browser, the amount of memory allocated
after boot was 35.6 megabytes. And in the other scenario, with the
IE web browser removed, the amount of allocated memory was
29.8 megabytes. That's a difference of about six megabytes, or
about 20 percent, in the memory use of Windows." Felten,
6/10/99pm, at 20:16-21.
154.3. Most of the performance problems that Microsoft alleges about
Windows 98 after Professor Felten's program had been run were merely
acknowledgments that Professor Felten had, in fact, successfully removed web
browsing from the system. For example, Microsoft argues that Professor Felten's
program removes the user's ability to type in a web page from the "Start" menu or to
place content from a web page on the "Active Desktop." Felten, 12/14/98pm, at 29:1-8; Felten, 12/14/98pm, at 30:19 - 31:7. Such observations demonstrate merely that
Professor Felten's program does what it was intended to do.
154.4. Microsoft attempted to demonstrate that Professor Felten's
program degrades the general performance of Windows 98 in ways unrelated to web
browsing, but the video demonstration that Microsoft offered as evidence did not prove
what it purported to prove.
154.4.1. On the video tape, Microsoft employee Yusuf Mehdi led
what appeared to be a guided demonstration of a Windows 98 machine connecting to
Microsoft's Windows Update web site. DX 2161. Mehdi said that Microsoft had "not
made any other changes to this computer or Windows 98 except to run Dr. Felten's
program as he describes in his expert report and his written direct testimony." DX 2161
(played 2/1/99pm), at 5:13-20. He also said that the computer was taking an unusually
long time to complete that operation because of "performance degradation that has
occurred because of running the Felten program."
- Mehdi stated that "As you can see, at the bottom of the page
here, we're actually connecting out to the Internet and
fetching that data. It's taking a very long time, however--unusually long--to access that web site. That's a result of
the performance degradation that has occurred because of
running the Felten program." DX 2161 (played 2/1/99pm), at
7:12-18.
- Mehdi stated that "Dr. Felten chose to let customers access
this one web site which is done using the IE code in
Windows 98 including MSHTML, URLMON, and WININET
among other files. However, Dr. Felten's changes make
access very slow." DX 2161 (played 2/1/99pm), at 7:21-25.
- Mehdi stated that "as I have already demonstrated in
showing how slowly the Windows Update site loaded, the
performance of the government version of Windows 98 is
much slower." DX 2161 (played 2/1/99pm), at 15:17-20.
154.4.2. In fact, however, almost nothing about the purported
demonstration was accurate and truthful.
154.4.2.1. In the first place, Microsoft's sponsoring witness
for the videotape, James Allchin, acknowledged that there were serious discrepancies
in the appearance of certain title bar screens on the "demonstration PC" in the video.
As a result, it initially appeared to him that, contrary to the claim made in the video,
Professor Felten's program had not even been run on that machine though he later
produced a different explanation.
- Compare GX 1688 (screen shot from unmodified
Windows 98 machine attempting to access Windows
Update, showing ""Microsoft Windows Update -
Microsoft Internet Explorer" at top) with GX 1689
(screen shot from a machine on which Professor
Felten's program had been run, showing "Microsoft
Windows Update - Windows 98" at top), and GX 1692
(screen shot from DX 2161, showing "Microsoft
Windows Update" at top).
- Allchin initially stated on cross-examination that it
appeared that Professor Felten's program had
apparently not been run on the demonstration
machine at all. Allchin, 2/2/98am, at 27:8-18
(testifying that "from what I'm seeing here right now, I
believe that that was done on a pre-Felten system,
although the point still stands. He has performance
problems and the Windows Update doesn't work, but
I believe, from what I'm seeing here, they filmed the
wrong system."); Allchin, 2/2/98am, at 28:23-24 ("In
this particular case, . . . I did not think the Felten
program had been run.").
154.4.2.2. In addition, as Mr. Allchin conceded on cross-examination, Microsoft's representation that it had "not made any other changes to this
computer or Windows 98 except to run Dr. Felten's program as he describes in his
expert report and his written direct testimony," was false (DX 2161 (played 2/1/99pm),
at 5:13-20). To the contrary, the videotape demonstration was apparently compiled by
splicing together footage from several different machines, some of which had both
extensive additional software installed and several unexplained manual changes to the
Windows Registry.
- Allchin testified that "I believe some of those
machines had Office on it, for example, and some of
them had some of the other browsers that were done.
Those weren't all the same machine, and they all
didn't have exactly the same thing on it." Allchin,
2/2/99am, at 29:24 - 30:5.
- When asked whether "some of the machines--at least
some of the time somebody had manually changed
the registry," Allchin answered: "Yes. There is a part
of the film that shows that they had apparently rerun
the test a couple of times filming, and they had
added--it's very easy to add empty entries to the
registry." Allchin, 2/2/99am, at 36:5-23.
- Allchin conceded that the registry changes shown on
the tape would not be there if you installed Felten's
program on a virgin machine and did nothing else.
Allchin Dep. (played 2/2/98am), at 40:16-21.
- Allchin conceded that the statement on the tape, that
nothing had been done to the machine but Professor
Felten's program, was untrue, because "they had
been through a rehearsal." Allchin, 2/3/99pm, at
57:5-19.
- Allchin conceded that even though the videotape
narration claims that, other than running Professor
Felten's program, "we have not made any other
changes to this computer," in fact the number of icons
visible on the desktop changes several times during
the video, clearly demonstrating that changes had
been made or that more than one PC must have
been used for the so-called "demonstration." Allchin,
2/3/99pm, at 64:3-19.
154.4.2.3. Allchin also conceded on cross-examination that
it would not in any event be technically possible to measure the kinds of alleged
performance degradation under the circumstances purportedly depicted on the video.
The entire premise of the demonstration was, therefore, inaccurate and misleading.
- Allchin, 2/2/99am, at 21:17-22 ("The test that we
know shows performance has to be done in a
controlled circumstance. You cannot prove the
performance slowdown when you're connected to the
Internet. You can only prove it in a controlled
situation, which is how we test the performance
degradation.").
c. There is no technical reason for Microsoft not to meet
demand for Windows 95 or Windows 98 without web
browsing by offering further separation between the
browser and the operating system
155. Beyond its plain ability to enable OEMs and users to "uninstall" Internet
Explorer, there is no technical reason for Microsoft's refusal to offer OEMs and users
the option of further separation between the browser and the operating system.
(1) Microsoft easily could supply versions of
Windows 95 and Windows 98 without the routines
that provide web browsing and still offer users
the same alleged benefits of its "integrated"
features and design
156. Although Microsoft contends that removing the routines that supply only
web browsing from Windows 95 or Windows 98 will deprive consumers of the benefits
of its "integrated" design and features, this contention is deeply flawed. First, such
"integration" could never supply any meaningful benefits to consumers who do not wish
to browse the web using Internet Explorer. Second, supplying an unbundled version of
Windows 98 to those consumers would not prevent Microsoft from offering an
"integrated" version to those consumers that desired it. Because of the malleable
nature of software, the "integration" necessary to produce any such benefits could be
achieved just as effectively by OEMs or end users installing a separately distributed
product. Whether Microsoft chooses to call that product a "browser" or an "operating
system upgrade" is, from both a technical and an economic perspective, immaterial.
(a) Bundling the browser with the operating
system is inefficient for users that do not
want the browser
157. Bundling browsing-only routines into large system DLLs is inefficient for
users who do not want web-browsing functionality.
157.1. Microsoft has never contended (and could not plausibly contend)
that the presence of browsing-only routines in its large DLLs improves system
performance even if those routines are never invoked by any code path on the system.
Routines that are not executed are simply dead weight and degrade system
performance.
- Professor Felten testified that: "Any code in a DLL that supports
only one function of the DLL may be removed without endangering
other functions of the DLL. For example, code that supports only
IE Web browsing functions may be removed without endangering
any non Web browsing functions of Windows 98." Felten Dir. ¶ 62.
- See infra Part V.B.4.c.(1); ¶ 170.
157.2. It can thus be efficient to place routines that are used only for web
browsing into large system DLLs only if the system is designed to deliver web browsing
functionality. And it is, by the same token, not inefficient for Microsoft to disaggregate
browsing-only routines from files like SHDOCVW.DLL in the versions of Windows 98
that it delivers without web-browsing functionality.
- Professor Felten testified that Microsoft split certain DLLs from
Internet Explorer 4 to Internet Explorer 5, which shows "of course,
that these DLL files are not indivisible and they are not fixed. And
so arguments that say that `A' and `B' are in the same DLL, and,
therefore, we cannot separate them, are not correct." Felten,
6/10/99am, at 51:23 - 52:2.
- Hadi Partovi testified that Microsoft has moved functions in one
DLL into different DLLs in succeeding versions of the product.
Partovi Dep., 1/13/99, at 659:7-23.
- Professor Felten testified that SHDOCVW.DLL "is a great example
of the point I'm trying to make about packaging of functions into
files. This SHDOCVW file is really a bundle of separate functions.
It contains some functions having to do with displaying the Start
menu. It contains some functions that have to do specifically with
Web browsing, and it contains some general user interface
functions as well. And to talk about this file as doing one thing or
being part of one product is really incorrect." Felten, 12/14/98am,
at 60:18 - 61:2.
157.3. Bundling routines into large system DLLs in fact creates
substantial inefficiencies for users who do not wish to use the functionality that those
routines deliver.
157.3.1. Installing software on a system that the end user does not
desire and will not use degrades performance by unnecessarily consuming system
resources, increasing the likelihood of software conflicts, and increasing the complexity
of the user interface. Those problems are exacerbated when the unwanted software is
integrated into the operating system because operating system code is often loaded
into the "working set" in dynamic memory, whereas unused applications typically sit
dormant on the hard drive.
- See infra Part V.B.4.c.(1); ¶¶ 164.1-2.
157.3.2. Microsoft understood that its decision to "integrate"
Internet Explorer into Windows 98 would in fact substantially degrade the performance
of Windows 98 for those users who desired to browse the web with Netscape
Navigator, or not at all.
i See supra Part V.B.2.3; ¶¶ 145-47.
157.3.3. As Microsoft recognizes, bundling new functionality into
the operating system can also make testing difficult and can slow the rate of innovation.
- In August 1996, Hank Vigil sent Paul Maritz an e-mail
entitled "Gravity or Anti-Gravity" and observed "Once
something has been pulled into the OS, the requirements of
quality, breadth of compatibility and scale mean that lots of
dependencies and trade-offs happen. The net result is that
the monolithic code base ships on long cycles after lots of
testing. There is also a tendency to meet all needs: be
everything to every consumer. Despite the advantages of
integrating more and more functionality into the OS, there
seem to be areas that can/would benefit by breaking out of
the OS so that they can develop richer functionality faster.
This allows for groups to discover, re-define and exploit
customer needs in ways that are hard when teams believe
that OS gravity is the central law." GX 157, at MS98
0167387.
- Brad Silverberg, commenting on the e-mail, agreed that
"This is a very good and important point. To me, the optimal
strategy is something in between: key components evolve
and improve and get delivered independently of the OS
release cycles, and then synch up when there is an OS
release, providing additional integration. Clearly the needs
for many components require that they release in much
faster cycles than the OS itself can. The most obvious
example is the browser. Yes, it will be integrated into the os,
and ie4 integrates deeply enough that it takes over the os's
UI; but it is on a much faster release schedule. We would
be dead if we had to synch with OS's." GX 157, at MS98
0167387.
- Maritz testified that: "There is a cost to integrating things into
your operating system. It means more work to be done,
more things to be tested, more software to be written."
Maritz, 1/27/99pm, 47:6-12.
- Jonathan Roberts wrote to James Allchin, Moshe Dunie, and
Carl Stork in March 1997 to discuss options regarding the
proposed bundling of IE 4 and Windows 98. One of those
options was to "De couple Memphis and IE and ship
Memphis in July/August and connect with IE in the OEM
channel when it ships." Stork responded that "Currently IE4
is so immature (and big & slow & compat-bug prone) that it
is impeding our self hosting process. We find tons of bugs
but so many are in IE components that our test & repro
efforts are becoming meaningless on the OS. We are also
finding more and more resistance on the team to install the
builds because things don't work. I am at the stage where I
do not recommend that we release anything with IE4
integrated under the name of a Memphis beta. Customers
would experience too many problems and the performance
would be unacceptable as well - it would be so bad as to
blemish the reputation of Microsoft and of Windows. . . .
More importantly, at this point it is getting in the way of valid
development testing & repro work for Memphis." GX 355.
157.3.4. And even Microsoft's own engineers have expressed
skepticism about Microsoft's decision to bundle more and more unrelated features into
the Internet Explorer DLLs.
- In an August 1997 e-mail, Christian Fortini wrote: "We have
to stop adding non-browsing features into Trident and start
taking some of the existing ones out. We should shrink the
core Trident code base down to a very compact (and fast)
HTML rendering and manipulation engine and hopefully limit
the number of people in this code base." GX 1377, at MS7
004591. "Trident" is Microsoft's code name for the file
MSHTML.DLL. Felten, 6/10/99am, at 46:23-24.
- Commenting on GX 1377, Professor Felten testified that
Fortini "appears to think that there are features in there that
are not related to browsing, and he's advocating taking them
out. . . . And he seems to say that, if that is done, that will
cause the HTML rendering engine to be more -- to be faster
and more compact. In other words, he seems to think that
it's desirable for technical reasons." Felten, 6/10/99am, at
47:4-12.
(b) Tying the browser to the operating system
is not necessary to achieve the benefits
sought by users who want both the
operating system and the Internet Explorer
browser
158. Microsoft is entirely free to offer a bundled version of the operating system
and the browser to OEMs and users that want it; it does not need to require OEMs and
users to take that version in order to offer it to those that want it.
- See infra Part V.B.3.d.(3); ¶ 165.2.
159. Moreover, even if Microsoft were unable to offer a bundled version, and
even if it were most efficient for web browsing routines to be placed in large DLLs,
operating system functionality and web browsing functionality can still efficiently be sold
or distributed separately. Microsoft could deliver web browsing functionality separately
to those that desire it, in the form of updated DLLs.
159.1. As previously explained, the malleable nature of software has two
important implications.
159.1.1. First, except at the extremes, software routines need not
reside in the same file to function together in a perfectly seamless or "integrated"
fashion. The organization of routines into files (including DLLs) is thus largely a matter
of design discretion, as opposed to engineering necessity.
- See supra Part V.B.2.d.(2)(a); ¶ 131.
159.1.2. Second, even if placement of certain routines in the same
files has engineering benefits, it is not necessary for those routines to be shipped
together to achieve that benefit. Users can be supplied with files containing some of
the routines and, should they also desire the others, can obtain a different file
containing the additional, related routines.
- See supra Part V.B.2.d.(2)(a); ¶ 126.4.
159.2. Therefore, there is no technical reason why Microsoft could not
ship even fully "integrated" web browsing functionality as a separate product that could
be installed on Windows 98.
- James Gosling testified that "regardless of whether a particular file
is installed on a computer with the original operating system, or
separately by a computer manufacturer, or by an end user
installing a program, the computer will operate in the same
manner." Gosling Dir. ¶ 42
- Professor Farber testified that: "Microsoft claims in its memoranda
filed with this Court that certain 'efficiencies' result from its
'integration' of some of the files (or DLLs) that are included in its
Internet Explorer (IE) product as part of Windows 98. . . . The
claims that efficiencies exist from this combination of functions are
misleading. While the combination may offer certain efficiencies,
these same efficiencies can be achieved without bundling of the
Web browser software with what Microsoft calls its Windows
operating system. This is because there are no technical barriers
that prevent Microsoft from developing and selling its Windows
operating system as a stand alone product separate from its
browser software . . . . Windows 98 (like all other software)
necessarily consists of modules which are malleable and
separable. There are no technical efficiencies for users achieved
by combining Microsoft's browser software with the remainder of
the software sold as Windows 98 that could not be achieved by
writing two programs in a manner that later could be loaded and
'integrated' either by the retail end user (i.e., just as end users
install any other application that runs on Windows) or by an OEM."
Farber Dir. ¶ 24.
- Professor Felten testified that "the nature of software is such that it
is easy to aggregate unrelated functions into the same file, or to
'integrate' separate products into a single product. The mere fact
that two functions are implemented in the same file, or that two
products are 'integrated' into a single product, does not imply that
they must be implemented in this fashion; because of the nature of
software, functions can be separated into distinct files, or
'integrated' products can be separated into distinct products without
any loss of capability." Felten Dir. ¶ 31.
159.3. Microsoft concedes that the version of Internet Explorer separately
distributed over the Internet accomplishes its "integration" in precisely that way.
- See infra ¶ 159.4.
- Carl Stork testified that Microsoft distributed Internet Explorer 3
separately from the Windows 95 because "Internet Explorer 3
represented significant customer improvements over previous
generations of Internet Explorer. And we wanted to provide that to
as many of our customers as we could. We have in the past
released advances to components that are part of the operating
system separately from the operating system as well. Another
example to that would be DirectX, which has frequently both been
made available on the Web as well as to be shipped with
applications. And there are others as well where I could cite the
same thing." Stork Dep., 8/11/98, at 40:19 - 41:7 (DX 2594).
159.4. As Microsoft's Jim Allchin conceded, all of the benefits offered by
the "integration" of Internet Explorer with Windows 98 can already be achieved by an
end user who installs the most recent, separately downloaded version of Internet
Explorer onto a version of Windows without Internet Explorer.
159.4.1. Separate delivery of Internet Explorer 4 and the original
browerless retail release of Windows 95 provides nearly all the web-related features of
Windows 98; as Mr. Allchin conceded when asked about 19 separate features of
Windows 98, Microsoft's decision to include the routines that supply those features in
Windows, rather than in a separate browser product, is simply a choice about
"distribution."
- When asked whether a user could achieve the "integration
of Internet technologies" accomplished by Windows 98 by
"combining a retail version of Windows 95 and a retail
version of Internet Explorer 4, both purchased separately,"
Allchin answered: "Yes. IE is replacing core Windows files,
and it becomes a modified Windows system that has this
integration in it." Allchin, 2/1/99pm, at 37:15-25.
- When asked whether Windows 98 was therefore "just a
distribution vehicle" for the technologies that Microsoft also
distributed as Windows 95 and Internet Explorer 4, Allchin
answered: "It's the same code out of Windows." Counsel for
the United States then asked: "It's the same code, and all
we're talking about are different distribution vehicles, in your
words; correct, sir?" Allchin answered: "Yes, that's what I
said, yes." Allchin, 2/1/99pm, at 39:18-25.
- When again asked whether "a user who had purchased
Windows 95 at retail and who added IE 4 purchased at retail
would have exactly the same experience," Allchin again
answered: "Yes, for exactly the same reason, i.e., he is
replacing core system files, no matter how you got it."
Allchin, 2/1/99pm, at 41:9-14.
- Allchin agreed "that you can get those benefits either by
buying Windows 98 or by having purchased an original retail
version of Windows 95 to which you added IE 4 either
downloaded or bought from retail or gotten in some other
way." Allchin, 2/1/99pm, at 45:9-25.
- Carl Stork testified that the Internet Explorer 4 team
developed its "set of technologies" for several different "ship
vehicles," one of which was "a retail upgrade for Windows
95," and another of which was inclusion in Memphis. Stork
Dep., 1/13/99, at 772:1-6.
- In a February 1997 summary of the results of Internet
Explorer 4 and Windows 98 focus groups, Christian
Wildfeuer discussed the reaction to the new "WebView" user
interface available with both products: "Interestingly, they
attributed these new features to Windows and not to Internet
Explorer, and this despite the fact that we repeatedly
hammered home the message that they would get all that in
IE 4 for free, if they downloaded it off the Web." GX 202
(emphasis added).
159.4.2. The remaining features can be obtained by combining a
separately-obtained Internet Explorer 5 and a version of Windows 95 on which Internet
Explorer 4 has been installed.
- Professor Felten testified that Allchin "mentioned three
features: HTML Help, Update Windows, and WebTV for
Windows" that were available in Windows 98 but not to a
Windows 95 user with Internet Explorer 4. Felten,
6/10/99am, at 18:18-19. Professor Felten further testified
that the HTML Help and Windows Update functionality are
delivered by the version of Internet Explorer 5 that Microsoft
is currently making available separate from Windows 98.
Felten, 6/10/99am, at 19:10-16.
- Professor Felten testified that the separately downloadable
version of Internet Explorer 5 does not include the WebTV
functionality "but that does not mean that it could not. In
fact, if you look at Windows 98, you'll see that WebTV for
Windows is an optional feature, which means the user has
the option to install it or not. And if the user has installed it,
the user can take it away at any time." Felten, 6/10/99am, at
19:20 - 20:2.
159.5. Microsoft's contention that a user cannot get the same benefits
from combining Netscape with Windows is beside the point; the important point is that
Microsoft does not have to bundle Windows and Internet Explorer in order for those
users who want both to get the benefits of both.
- Professor Felten testified that Allchin's assertion -- that
installing Navigator on top of the original retail version of
Windows 95 results in losing 19 or 20 different features
available on the integrated Windows 98 -- is "not really
related" to the issue of whether Internet Explorer has to be
included with Windows. Felten continued: "If you want to
understand the relationship between Windows 98 and
Internet Explorer, you can't do it by looking at the
relationship between two different products, Windows 95
and Netscape Navigator. So I don't see the relevance of
that to any argument that IE has to be delivered with
Windows 98." Felten, 6/10/99am, at 16:13-23.
- Instead, Professor Felten testified that "the relevant
comparison is what happens when you combine the original
retail version of Windows 95 -- that's the one that came
without any browser -- what happens when you combine
that with IE 4 distributed separately, or perhaps IE 5
distributed separately." Felten, 6/10/99, at 17:2-6.
159.6. Industry participants, including Microsoft, routinely describe
software products as seamlessly "integrated" even when they are not shipped together
or even produced by the same company.
159.6.1. Microsoft, for example, describes Office as "integrated"
with the operating system and each of its separate components, even though the
functionality supplied to the end user is identical whether the components are
purchased together or separately.
- Robert Muglia testified that Microsoft Office is "an integrated
package" including distinct applications known as Word and
Excel, which were "designed to be integrated" together into
Office, but that Microsoft nonetheless distributes Word and
Excel separately. "The way I might say that is that Office is
an integrated package overall. It was designed to be
integrated. We produced, because our customers would like
us to produce it, a separate word-processing program that
we derived from the overall integrated Office package and a
separate spreadsheet program." Muglia, 2/26/99pm, at
67:17 - 70:3.
- In response to Muglia's comments about Office, Professor
Felten testified: "In this instance, Microsoft makes Word and
Excel available separately for those users who want them.
Or for those users who want both, Microsoft provides a
single box they can buy which gives them a single install.
So, in other words, Microsoft can give the user the choices
they want . . . The same is true with regard to Internet
Explorer and Windows. Microsoft could provide a single
install for those users who want both Windows and Internet
Explorer, without taking away the other choices such as
buying only Internet Explorer." Felten, 6/10/99pm, at 12:1-17.
159.6.2. Intuit describes a browser as "integrated" into Quicken,
even though Intuit must obtain a browser from another company.
- William Harris testified that in early 1995 Intuit was
interested in "the possibility of bundling a browser and with
some light integration." This meant creating a "mechanism"
"by which, within the Quicken product, one could instantiate
the browser and instruct the browser as to the URL that
should be displayed." Harris, 1/4/99pm, at 8:23 - 9:5.
159.7. Other operating system and browser vendors deliver similar
benefits, and describe their products as seamlessly "integrated," even though they can
be distributed and installed separately.
159.7.1. For example, the Caldera Open Linux product, which
Allchin himself demonstrated, provides "integrated" features yet is completely
removable and replaceable, just like any application installed on top of the operating
system.
159.7.1.1. Caldera OpenLinux, combined with the "KDE"
browser, provides "integrated" features similar to those delivered by the combination of
Windows 98 (or, as noted above, the combination of the retail version of Windows 95
and Internet Explorer 4 or 5).
- Professor Felten testified that the video
demonstration produced by Allchin "claimed to show .
. . that Caldera OpenLinux shipped a browser, which
Mr. Allchin characterized as integrated, and that that
browser had some of the features that Mr. Allchin
said were benefits of the integration of IE into
Windows. In other words, it claimed to show that
Caldera was, in some sense, acting like Microsoft in
achieving these benefits supposedly by putting in an
integrated browser." Felten, 6/10/99am, at 23:7-14.
- During James Allchin's cross-examination, the
government produced a still screen shot of
Microsoft's video presentation that demonstrates that
the combination of Caldera OpenLinux and the KDE
browser provides integrated features similar to those
offered by Windows 98 and IE, including 1) single-window navigation (also known as unified viewing)
between the Web and local files, including the use of
back and forward buttons to let the customer manage
local files and folders, as well as internet content; 2)
unified favorites list; and 3) unified history list. See
GX 1707 (still screen shot of Microsoft video demo
played in the record at Allchin, 2/1/99am,at 61:1 -
66:21); see also Felten, 6/10/99am, at 23:20 - 24:21
(examining GX 1707).
159.7.1.2. The KDE browser is entirely separate from the
OpenLinux operating system; it is produced by a different company; it is easily replaced
by another operating system installed on top of OpenLinux; consumers can uninstall it
at any time.
- The KDE browser is produced separately from the
OpenLinux operating system, by a different company.
Felten, 6/10/99am, at 26:5-14.
- The KDE browser can be easily replaced with other
browsers installed separately on top of OpenLinux,
and will then deliver the same integrated functionality.
Felten, 6/10/99am, at 26:15-18.
- OpenLinux customers can choose not to install the
KDE browser or can uninstall it at any time. Felten,
6/10/99am, at 26:22-25.
- The KDE browser provides similar integration when
installed separately on top of other operating systems
to which it is ported but with which it is not bundled.
Felten, 6/10/99am, at 26:19-20 ("KDE browser runs
on other operating systems, such as Solaris, HP-UX
and IRIX"); Felten, 6/10/99am, at 27:1-11 (KDE
browser provides integrated features if installed on
top of other operating systems).
- For all of these reasons, Professor Felten testified,
"The Caldera example contradicts" Allchin's testimony
(Allchin Dir. ¶ 3) "because the KDE browser is an
add-on product and it comes from a third party, and
yet it achieves these benefits of integration that Mr.
Allchin says can only be achieved by bolting the
browser onto the operating system." Felten,
6/10/99am, at 28:4-8. He continued: "What we see
with Caldera is a pair of products, if you will -- the
Linux and the KDE browser -- which work well
together, and are integrated in that sense, but are not
inseparable." Felten, 6/10/99am, at 29:9-12.
159.7.2. The Be OS product also provides integrated features
using a removeable, replaceable browser application installed on top of the operating
system.
- During its video demonstration, Microsoft employee Vinod
Vallipolil stated: "The demonstration will show that the Be
OS includes browsing and multimedia functionality, which
are built directly into the operating system, and that no third
party code is required in order to exercise this functionality."
Allchin, 2/1/99am, at 58:11-15.
- GX 1771 (a series of screen shots that shows that the
browser on Be, Net Positive, is an application listed under
the "apps" directory which can be removed by clicking on it
and dragging it to the trash can; removal results in reduction
of size taken up by applications from 4.3 megabytes to 3.0
megabytes -- a reduction of 1.3 megabytes); see also
Allchin, 2/2/99am, at 13:5 - 19:13 (Boies walks through GX
1771 with Allchin).
- The Net Positive browser can be removed from the Be OS.
Removing the Net Positive browser from the applications
directory frees up 1.3 megabytes of RAM on the Be OS
applications directory. Allchin, 2/2/99am, at 13:5 - 19:13;
GX 1771.
- Although Be's help system will not function fully in the
absence of a browser, the help system will work if another
browser is installed after Net Positive is removed. Allchin,
2/2/99am, at 20:5 - 21:4.
160. Accordingly, even if Microsoft's design creates benefits for some users,
forcing all of its customers to take an "integrated" browser is wholly unnecessary to
achieve those benefits; Microsoft's decision to force users to take the browser in order
to get the operating system is, as Mr. Allchin put it, simply a choice about "distribution."
- Professor Fisher testified regarding the two senses of "integrated": "One
of them is to call two software items integrated if they run seamlessly
together. . . The second is integration in the sense that it is impossible or
very difficult to split it apart. Now, as to whether that is anticompetitive, I
think for that one has to think about some more. The consumer benefit
doesn't come from . . . the fact that code is designed in that form. The
consumer benefit comes from seamless operation. Microsoft, in Windows
95, designed Internet Explorer, particularly Internet Explorer 4.0, and
Windows 95 to work seamlessly togther and be integrated in that form.
And there is, you know, evidence that they could have perfectly well
designed Windows 98 and Internet Explorer to also work seamlessly
without having the what I have referred to the other day as the welded
feature, the difficulty of taking it apart feature. If that is so, then I think
yes, it probably was anticompetitive . . . because they could have done it
in a way that is less restrictive." Fisher, 1/12/99pm, at 4:5 - 6:20.
- Professor Fisher testified: "there is substantial testimony from Microsoft
witnesses . . . that the advantages to consumers from the combination of
Windows 98 and IE . . . could be achieved just as substantially as in
having Windows 95 and IE separately without the two of them being, so to
speak, so tightly welded as they are in Windows 98." Fisher, 6/1/99am, at
42:17-23. That testimony, Professor Fisher explained, led him to
conclude that "there are no particular consumer benefits from getting the
things together as opposed to getting them separately, but there are no
particular advantages -- I mean economic cost advantages to Microsoft
from delivering them together, as opposed to just delivering them
separately." Fisher, 6/1/99am, at 44:23 - 45:3.
(2) Microsoft's forced licensing of its browser is not
necessary to provide OEMs and users with other
benefits, such as new file formats and data
protocols
161. There is also no technical reason to force users to take the browser with
the operating system in order to supply them with the other consumer benefits Microsoft
identifies. Each of those benefits can be supplied separately by the browser and the
operating system.
161.1. Non-browsing features. Microsoft points out that Windows 98
includes non-web browsing features such as DVD and USB support (Allchin Dir. ¶¶
119-120). But there is no reason to force users to take routines that supply web
browsing to obtain these features.
- Professor Felten testified that there are "plenty of things in
Windows 98, unrelated to browsing, that are not in Windows 95.
Plenty of Features. For example, support for new kinds of
hardware devices." Felten, 6/10/99am, at 34:7-10. Professor
Felten further testified that there is no technical reason to require
users to take Internet Explorer to obtain these features. Felten,
6/10/99am, at 39:11-15.
161.2. Support for new Internet protocols and data formats. Nor does
providing support for new data formats and Internet protocols provide, as Microsoft
suggests (Allchin Dir. ¶¶ 212-213), a justification for forcing users and OEMs to take
Microsoft's browser.
161.2.1. File formats are commonly supplied separately. For
instance, Adobe Acrobat Reader, a third-party application provided entirely separately
from Windows, provides support for the popular "PDF" file format commonly used for
viewing text retrieved over the Web.
- Professor Felten testified: "One example is the Adobe PDF
viewer. `PDF' stands for portable document format. And it's
a very commonly-used format for describing documents so
that you can move them from computer to computer, print
the, view them on your screen and so on. And Adobe
makes available some programs with names like PDF
reader or PDF viewer that are widely used for viewing and
printing PDF documents. . . . It may be provided with some
OEM's, but I don't know of any instance in which an OEM is
forced to take it or a user is forced to take it." Felten,
6/10/99am, at 42:11-23.
- Professor Felten testified: "again, in general, there's really
no connection between offering users the ability to support a
new protocol or a new format -- there's no connection
between that and forcing them to take any particular
software product." Felten, 6/10/99am, at 43:3-6.
161.2.2. Similarly, Microsoft need not force customers to take its
web browser to supply support for HTML.
161.2.2.1. Including only an HTML rendering engine in its
operating system -- an option Microsoft specifically considered but rejected -- would
suffice.
- Microsoft has a file which, among other things,
contains an HTML rendering machine; that file is
MSHTML.DLL. Felten, 6/10/99am, at 44:20-22.
When asked whether there were any inherent reason
to put different functions -- HTML rendering and other
functions -- together in the same DLL, Professor
Felten testified: "there are many ways in which
functions can be grouped into DLL's. . . . You might
make a decision based on whether things fit in a
certain way," like you would with organizing grocery
bags, "but fact that the ice cream and the carrots are
in the same bag doesn't necessarily mean that they
are related in some way. . . . It's possible to put
unrelated functions into the same DLL." Felten,
6/10/99am, at 45:6-19.
- Specifically regarding MSHTML.DLL, Professor
Felten testified: "You have the function of an HTML
rendering engine in there, and you have other things
as well. Those things could be separated, but they
are not." He further explained that the rendering
engine could be taken out of the DLL, or alternative
the other stuff could be taken out. Felten, 6/10/99am,
at 45:21 - 46:4.
- In a March 1997 e-mail, similarly, Allchin discussed
the idea of separating the "shell" from the "browser"
(IE 4). In doing so, he presented some development
options, including the following: "Move the shell -- but
not the browser -- to the OS team. This was my
recommendation before as you know. It may not be
the thing you want to do for other reasons, but it is the
right thing to do for the OS (both Memphis and NT).
IE 4 would just plug into the environment." GX 616
(emphasis added).
- See also supra Part V.B.3.c(1)(c); ¶ 161.
161.2.2.2. Indeed, other operating system vendors -- which lack
Microsoft's monopoly power -- include a separate HTML rendering engine, even when they
bundle a removable browser.
- Apple's Avadis Tevanian testified that in the Mac OS
"there are multiple HTML renderers," including "one
independent of the Netscape Navigator and Internet
Explorer technologies." Tevanian, 11/5/98pm, at
67:20 - 68:4.
d. The post hoc economic justifications Microsoft's
witnesses have advanced for tying Internet Explorer to
Windows are contrary to the evidence
162. Microsoft's economic justifications for forcing users to take the browser
with the operating system, and for making the two difficult to separate, cannot be
squared with the evidence.
(1) Microsoft's conduct was not plausibly designed
or intended to increase demand for Windows
163. Microsoft's tying of Internet Explorer to Windows, and the decision to make
it non-removable, was not intended to increase demand for Windows.
163.1. Although creation of an appealing new complement can increase
demand for a product, the value of a product is maximized by helping consumers use
the complement of their choice, including complements (like browsers) produced by
other firms.
- Professor Fisher testified that "if browsers are a complement to
operating systems such that the sale of browsers that can be used
with Windows will increase demand for Windows, it should not
matter who makes the complement." Fisher Dir. ¶ 129(b).
- Professor Fisher testified that if Microsoft were genuinely trying to
maximize demand for Windows, "I don't know that Microsoft would
have an interest in promoting the Netscape browser, but Microsoft
would surely have no interest in restricting its distribution, since
people who wanted to use the Netscape browser with Windows
would be happier people with Windows. To some extent, it would
increase the sale of Windows." Fisher, 6/1/99am, at 66:3-8.
163.2. Microsoft did not, however, seek to aid consumers who wanted a
non-Microsoft browser. To the contrary, it took deliberate, active, and costly steps to
impair the distribution and usage of Netscape Navigator, including raising the costs to
both OEMs and consumers of supporting and using Navigator through its tying
arrangement and related contractual restrictions.
- The tying of Internet Explorer to Windows itself made it more
difficult for users to obtain other browsers. See infra ¶¶ 169-171.
- Professor Fisher testified that "if browsers are complements to
operating systems such that the sale of browsers that can be used
with Windows will increase demand for Windows, it should not
matter who makes the complement. But Microsoft cared greatly
who made the browsers used with Windows." Fisher Dir. ¶ 129(b)
- Professor Fisher testified that "Microsoft even tried to discourage
Netscape from offering Netscape's browser for use with Windows -
an action inconsistent with browsers being a complement to
Windows, whose distribution Microsoft wanted to maximize."
Fisher Dir. ¶ 129(c).
- Professor Fisher testified that "Microsoft was preoccupied not with
increasing total sales of browsers but with Microsoft's share of
browser sales. Indeed, Microsoft studied, and tried to implement,
ways to disable Netscpe and reduce total browser sales. That
conduct doesn't 'make sense from a business standpoint' if
browsers are viewed as a means of increasing sales of Windows.
But this conduct makes good sense if browsers are viewed as a
competitive threat to Microsoft's Windows monopoly." Fisher Dir.
¶ 129(e).
- When asked whether Microsoft did, in fact, attempt to restrict
distribution of the Netscape browser, Fisher testified: "Oh, you bet
you. To take a particular example, in its contracts with ISPs,
Microsoft doesn't merely require that the ISP ship some minimum
number -- I think it's usually around 85% -- of Internet Explorers to
the ISP subscribers. That requirement alone would have permitted
the ISP to ship both IE and Netscape Navigator. The contracts, in
fact, require that the ISP not ship more than, in this example, 15
percent of other browsers to the ISPs. That's a restriction on
Netscape. If Microsoft were really interested in selling Windows, it
wouldn't have any interest in doing that. And it can't have any
interest in doing that to protect its, quote, sales of IE, end quote,
because it doesn't have any, quote, sales of IE, end quote. It's a
no-revenue product." Fisher, 6/1/99am, at 66:12-25.
- Dr. Warren-Boulton testified that "Microsoft's . . . efforts to increase
IE's share by excluding Netscape and making it more difficult for
users to obtain Netscape's browser could only reduce the value of
its operating system to consumers." Warren-Boulton Dir. ¶ 187.
(2) Microsoft's tie-in and related restrictions were not
reasonably necessary to preserve the integrity of
the Windows platform
164. Microsoft's argument that tying the browser to the operating system is
reasonably necessary to preserve the "integrity" of the Windows platform (Allchin Dir. ¶
85; Kempin Dir. ¶ 29) is pretextual.
164.1. First, concern with the "integrity" of the platform cannot explain
Microsoft's original decision to tie Internet Explorer to Windows 95 because Internet
Explorer 1 and 2 did not contain APIs.
- See infra Part V.G.6.a; ¶ 312.1.
164.2. Second, concern with integrity of the platform cannot explain
Microsoft's refusal to offer OEMs the option of Windows 95 or 98 with Internet Explorer
uninstalled or its equivalent because APIs, like all other shared files, are left on the
system when Internet Explorer is uninstalled.
- Professor Felten testified that "I have implemented the prototype
removal program to continue to support the ability of ISVs to use all
of the shared program libraries shipped with Windows 98. Such
shared program libraries are left substantially unchanged, though
they are no longer used in the course of Web browsing without the
addition of another software program such as AOL's access
software or Intuit's Quicken personal finance software. Microsoft
could have produced a version of Windows 98 without Web
browsing in a way that did not adversely affect the functionality of
ISV applications." Felten Dir. ¶ 56.
- Professor Felten testified that leaving shared files in place
"conforms to the ordinary way in which software application
programs are removed," as well as to the instructions that Microsoft
itself gives to application developers. See Felten Dir. ¶ 57 (citing
GX 431, Microsoft's Handbook for Applications, p. 29).
- Dr. Warren-Boulton testified that "Microsoft's objective of supplying
ISVs with a consistent platform does not provide an economic
justification for biasing OEMs' choice of which browser to feature"
because "Microsoft's design decision was arbitrary; Microsoft could
have put 'platform files' [such as shared files] entirely in the
operating system and not included any such files in its browser
product." Warren-Boulton Dir. ¶ 165; id. ¶¶ 166-167.
164.3. Third, Microsoft's concern that offering OEMs the choice as to
whether to install certain browser-related APIs would fragment the Windows platform
(Kempin Dir. ¶ 29) is insubstantial because OEMs, which operate in a competitive
market, have ample incentives to include APIs (including non-Microsoft APIs) for
applications their customers demand.
- Professor Fisher testified that Microsoft's concern with offering
developers a stable, up-to-date platform is insubstantial because
"it's not obvious that those APis have to be Microsoft's API's for
there to be a stable set of API's offered to developers." Fisher,
6/3/99am, at 21:21 - 22:20.
164.4. Fourth, even if there were some potential benefit from forced
licensing of a single set of APIs to all OEMs, any such justification could not apply in
this case because Microsoft itself perpetuates fragmentation of the platform.
- Dr. Warren-Boulton testified: "There are millions of PCs running
earlier versions of Windows releases that lack the latest versions of
Windows 95 or Windows 98. To ensure that the software they
develop runs no matter which version of Windows a PC contains,
ISVs commonly redistribute necessary shared program libraries
with their software. In short, Microsoft's own practice of continually
updating its platform means that application developers must
repliate part of the platform with the software they distribute and,
therefore, that the effect on an OEM removing certain parts of the
'platform' is likely to be small." Warren-Boulton Dir. ¶ 170.
164.4.1. Because Microsoft frequently releases new APIs with its
updated versions of Internet Explorer and Windows, the installed base of Windows PCs
has very different sets of APIs. Microsoft's practice of continually updating those APIs
perpetuates this fragmentation.
- Professor Fisher testified that "the Microsoft APIs are not, in
fact, stable. They change. And ISVs have to keep
embedding pieces of the appropriate APIs into their own
software and shipping it out." Fisher, 6/3/99am, at 22:11-14.
He also testified that "ISVs have to redistribute IE code
anyway because Microsoft has put so many different
releases out there." Fisher Dir. ¶ 165.
- Dr. Warren-Boulton testified: "There are millions of PCs
running earlier versions of Windows releases that lack the
latest versions of Windows 95 or Windows 98. To ensure
that the software they develop runs no matter which version
of Windows a PC contains, ISVs commonly redistribute
necessary shared program libraries with their software. In
short, Microsoft's own practice of continually updating its
platform means that application developers must replicate
part of the platform with the software they distribute. . . ."
Warren-Boulton Dir. ¶ 170.
- John Gailey, Director of Engineering for Novell, declared
that: "Because Microsoft is constantly changing and
updating the system services provided by Windows 95,
Novell bundles some of those operating system services
with GroupWise in order to ensure that all users have
available to them the latest version of the system service
GroupWise is calling upon." Fisher Dir. ¶ 165 (quoting
Gailey Decl. 11/17/97).
164.4.2. ISVs have adapted to this reality by redistributing needed
APIs with their applications in order to ensure that the APIs get installed on the user's
PC; to facilitate this, Microsoft makes the APIs it ships with Internet Explorer available
to third party developers for distribution with their products.
- Microsoft's David Cole testified that many ISVs redistribute
Internet Explorer in order to "upgrade the operating system
to the level they need to run their application. . . . That's a
very common practice." Cole Dep., 1/13/99, at 390:20-24.
- James Allchin testified that "Microsoft does license
developers whose products rely on Internet Explorer
technologies to ship them with their products so that older
versions of Windows can be upgraded to the necessary
level of functionality." Allchin Dir. ¶ 135. Allchin testified
that "we do this for graphics and everything else. We take
the system to the installed base, provide pieces of software
that upgrade it. Games that run easily on Windows 95
wouldn't operate correctly on Windows 95 without additional
software . . . called DirectX. So, we either have updated
Windows 95, we put it on the web site or even let people,
ISVs, vendors, ship that software. So it's all about how far
we take it about adding new functionality to the system for
the installed base." Allchin, 2/1/99pm, 47:10-19; see also
Maritz Direct ¶ 171; Jones Dep., 1/13/99, at 535:6 - 536:6.
- William Harris testified that: "The combination, or separation,
of software products or components will almost always have
some potential advantages and some potential
disadvantages. For Intuit, in particular, distributing a
browser with our products, rather than as part of the
operating system, has some real advantages. The fact that
Intuit currently distributes a version of Internet Explorer with
Quicken is illustrative of this point. Intuit has exerted
hundreds of hours testing and verifying that Quicken will
operate with the specific version of Internet Explorer that
comes with Quicken. If a Quicken customer does not
already have a compatible version of Internet Explorer,
Quicken will install the version of Internet Explorer that
comes with Quicken. This ensures that Quicken will work
the way it was intended and tested." Harris Dir. ¶ 85.
- Professor Fisher testified that "ISVs have to redistribute IE
code anyway because MS has put so many different
releases out there." Fisher Dir. ¶ 165; see also Felten,
6/10/99am, at 61:10-14.
- Glenn Weadock testified that applications that update
Windows DLLs are common. Weadock, 11/17/98am, at
25:15 - 26:10.
- Robert Muglia testified that, because of differences in
implementations of Java, ISVs similarly redistribute Java
virtual machines with their Java programs to ensure that
those programs will function properly. Muglia Dir. ¶¶ 87,
104, 107.
164.4.3. This is equally true of Microsoft, which distributes APIs --
including those distributed with Internet Explorer -- with a number of Microsoft
applications.
- Muglia testified that Microsoft will distribute Internet Explorer
with Office 2000: "There is updated Windows functionality
that we need to take advantage of. So, to make sure that
functionality is on the user's computer, we are distributing it
with Office 2000, just like other ISVs have the option to do."
Muglia, 2/26/99pm, at 67:3-11; see also GX 727 (Microsoft
Office 97 redistributes all of Internet Explorer 3, and installs
certain Internet Explorer components that it needs to
function properly).
- Professor Felten testified that Microsoft itself redistributes
various Internet Explorer components with its application
products Office 97, Money 98, Money 99, Frontpage 98,
Visual Studio 6.0, MSN, and Plus 98. Felten, 6/10/99am, at
62:13-15.
- GX 2220 (series of screen shots captured from the beginning of the
installation process for Visual Studio) ("Visual Studio comes
with the latest version of Internet Explorer 4.01. The
updated version of Internet Explorer 4.01 is an essential
component of Visual Studio 6.0 Enterprise Edition and
installation is required."); see also Felten, 6/10/99am, at
67:12 - 71:4.
164.5. Although Gordon Eubanks testified that redistributing software
components was inefficient for ISVs, he admitted that Symantec, in fact, routinely
redistributed Internet Explorer components in its ordinary business.
- Eubanks testified that, until recently, Symantec "shipped a
rendering engine with Norton Utilities so we could display
HTML because we couldn't count on every customer having
a browser." Eubanks, 6/16/99pm, at 76:25 - 77:3.
-
Eubanks testified that "in the past, we did distribute this
HTML rendering DLL, and it was used by components of
Norton Utilities." Eubanks, 6/16/99pm, at 74:5-8.
(3) Microsoft's quality-related justifications are
pretextual
165. Microsoft's quality-related justifications are similarly pretextual.
165.1. As explained, Microsoft's design of Windows 98 delivers to end
users no technical benefit (aside from the ability to browse the web) that could not be
achieved in a version without web browsing.
- See infra Part V.B.2.e.(3)(b); ¶¶ 151-154.
165.2. Microsoft's concern that meeting demand for Windows without
Internet browsing will degrade product quality or its reputation is insubstantial in any
event because, if consumers genuinely prefer the version of Windows bundled with
Internet Explorer, they would choose it in the market.
- Professor Fisher testified that "if consumers prefer seamless
operation, they would chose it in a competitive market," and that
"competition leads to a consumer-driven arrangement of . . . what
gets produced, what gets distributed and so on. If welding it
together actually provided benefits, then consumers would choose
the welded version as opposed to a separate version, and they
would be willing to pay more." Fisher, 1/12/99pm, at 6:3 - 7:7.
- Professor Fisher testified that: "The consumer gets the same
benefits if it . . . acquires those two things separately. In that event,
there is no reason why Microsoft shouldn't offer them typically
separately throughout and let consumers decide, if those are really
good benefits, that they want to acquire them." Fisher, 6/1/99am,
at 44:6-12.
- Dr. Warren-Boulton testified that if removing Internet Explorer from
Windows 98 "would affect the way in which Windows 98 would
operate either by itself without a browser or with Netscape, then
that would mean in the market people would look at that product
and say 'I don't like that product very much.' And what would
happen is people wouldn't choose it, and so the outcome of the test
would be, in fact, that people would say I don't want to have the
two products separately." Warren-Boulton, 11/24/98pm, at 22:25 -
23:7.
- Dr. Warren-Boulton testified that "I don't think that Microsoft -- there
is any reason why Microsoft cannot inform the customers that this
is a Windows 98 product that does not have IE." Warren-Boulton,
11/24/98pm, at 23:15-20.
165.3. Ensuring ease of preinstallation and configuring the product
similarly cannot justify Microsoft's tie-in.
165.3.1. OEMs can provide the benefit of preinstalling and
configuring to end users just as effectively as Microsoft, and those OEMs that want
Microsoft to perform that service can obtain a bundled version from Microsoft.
- When asked whether "some customers might like to open up
a new PC, plug it in, and get connected to the Internet
easily," Professor Felten testified: "Absolutely. I think also
that those customers would prefer to have the browser of
their choice pre-installed on the system. And I'm not saying
that OEM's should not be free to do that. The whole point
that I'm trying to make in my testimony and the whole point
of the prototype removal program is trying to make is that
that choice could be provided to end users, to OEM's and all
along the supply chain so that users can have what they
want." Felten, 12/14/98pm, at 28:2-13.
- Professor Farber testified that none of his testimony "denies
the possible convenience or preference of some users for
'one stop shopping' for bundled products such as the current
version of Windows 98 sold as one product by Microsoft.
Those OEMs and retail end users who may find this
convenience outweighs any technical inefficiencies
described here can certainly still choose to buy Windows 98
in the form it now exists." Farber Dir. ¶ 28.
- Muglia conceded that the inconvenience of multiple setup
procedures for customers is no reason not to offer an
unbundled version of Microsoft Office; "Sure, of course not.
Again, we're just providing choices for customers. We're
saying if people want to buy just a word processor or
spreadsheet, they have the option to do so." Muglia Dep.
(played 6/10/99pm), at 11:17-21.
- IBM's John Soyring testified that PC suppliers "generally
have ample ability to include applications such as a browser
with an operating system and load this combined set of
products on their machines." Soyring Dir. ¶ 21; see also id.
¶ 22 ("[PC suppliers] can install browsers on the operating
system on machines they ship, so long as they are given
appropriate information by the browser supplier and are
given any information specific to the particular operating
system involved by the browser supplier or the operating
system supplier.").
- When asked whether the convenience of a single installation
procedure was "a compelling rationale requiring users to
take Internet Explorer along with Windows," Professor
Felten answered: "No, it's not. If a user wants both
Windows and Internet Explorer, they -- Microsoft can offer
them that option with the single install. But a user who only
wants Windows without Internet Explorer can get that in a
single install." Felten, 6/10/99pm, at 5:19 - 6:1.
- Professor Felten testified that "Microsoft could provide a
single install for those users who want both Windows and
Internet Explorer, without taking away the other choices
such as buying only Windows or only Internet Explorer."
Felten, 6/10/99pm, at 12:14-17.
165.3.2. Microsoft gives users the option of installing or
uninstalling other programs that it distributes with Windows or with other software
bundles.
- A series of screen shots from Windows 98 shows that the
operating system has menus from which various software
programs, unlike IE, can be installed or uninstalled from
Windows. GX 1700.
- WebTV for Windows can be optionally installed or removed
from Windows 98 despite the fact that it is delivered along
with the integrated Internet Explorer 5. Felten, 6/10/99, at
19:24 - 20:20 (WebTV falls under "add/remove" control
panel on Windows 98).
- Microsoft provides single installation with "no assembly
required" optionally in the case of Microsoft Office, allowing
users to choose whether to obtain all the programs bundled
together in Office at the same time or separately. See supra
¶ 165.3.1; Felten, 6/10/99pm, at 11:22 - 12:17.
4. Microsoft's tying of Internet Explorer to Windows has caused
significant exclusionary effects and consumer harm
166. Microsoft's coercive binding of Internet Explorer to Windows raised the
costs of using other browsers, facilitated Microsoft's objective of preventing Netscape
from developing into a viable threat to the applications barrier to entry, and thereby
harmed consumers and aided Microsoft in its objective of preserving its operating
system monopoly.
a. Installing a second product in a software category
imposes costs on OEMs
167. Microsoft's conduct raised the costs to OEMs of carrying Netscape or other
browser products.
(1) Increased technical support costs
167.1. OEMs bear essentially all the customer support costs for the
computers they sell, including those related to Windows, even though Windows is
Microsoft's product.
- Microsoft's licensing agreements with PC OEMs require them to
- redacted - GX 418 (Toshiba) (sealed); see also GX 410 (DEC), at
MS98 0008841 (sealed).
- Joseph Kanicki testified that Dell bears support costs for the
computers it sells. Kanicki Dep., 1/13/99pm, at 342:5-7.
- Gayle McClain testified that Gateway provides customer support
for the machines they sell, and that "a new user could call
regarding almost anything," including being confused by clutter on
the desktop. McClain also testified that there is no mechanism for
Microsoft to reimburse them for any of those support costs.
McClain Dep., 1/13/99pm, at 616:15 - 617:12.
- both the hardware and software on the PCs it sells. Rose,
2/18/99pm, at 41:23 - 42:11; see also Rose Dir. ¶ 15 ("Many of our
consumer customers do not care to know the constituent
components (or even the components' brand names). Rather, they
are buying the 'out of the box' experience. They want their
computers to be simple and easy to use right out of the box, and
they look to the Compaq brand to make that experience a good
one.").
- Soyring testified: "Software suppliers often provide better pricing to
PC suppliers if the PC supplier responds to the support calls from
customers and handles the initial analysis of potential problems.
This activity can range from simple to very complex and can even
involve review and analysis of the source code for the software
involved." Soyring Dir. ¶ 20.
167.2. Adding a second product in a given category, including a browser
or the visible means of accessing the browser, can significantly increase those support
costs.
- When asked whether Compaq's support costs would go up if they
installed more than one browser on a PC, Rose testified that "I
would expect that as the number of multiple things go up, the
support costs would go up." Rose, 2/18/99pm, at 42:12-22. Rose
testified that Compaq had evaluated the relative costs and benefits
of preinstalling more than one product in any particular software
category, and reached the following conclusions: "'That, one, it's
expensive; puts a greater cost burden on Compaq; adds more
complexity; causes confusion to the customers, particularly
consumer customers, that don't have any personal computing
experience.'" Rose Dep. (read 2/18/99pm), at 45:25 - 47:13; see
also Rose, 2/18/99pm, at 47:25 - 48:14.
- Soyring testified that, "even if there is customer demand for
another browser, the PC supplier has to consider the cost of the
second browser. Even if the other browser supplier offers it to the
PC supplier for free, the PC supplier will incur substantial additional
costs, including additional testing, distribution and support costs."
Soyring Dir. ¶ 27.
- Soyring also testified: "Whenever manufacturers install an
additional program, there usually is incremental cost, not just the
fee for the license, but training their staff to build the image that's
preloaded on the hard disk, or whatever media they choose, to
train their support staff, because typically contracts with software
manufacturers require that the first two levels of customer contact,
if there's a support problem, is with the P.C. manufacturer;
therefore, there is additional training cost. There's costs that go
into their marketing programs to explain the various products that
they've installed and the value to their customer. So yes, there are
additional costs." Soyring, 11/18/98am, at 70:15 - 71:1.
- Microsoft's Gayle McClain testified that multiple icons, or
"redundancy of function in various places," is confusing to end
users (McClain Dep., 1/13/99pm, at 623:6 - 624:11), and that
Gateway wanted to remove icons from Windows 98 because of
concern about clutter on the screen. McClain Dep., 1/13/99pm, at
614:5 - 615:9.
- According to John Kies, Senior Product Manager for the Packard-Bell/NEC Versa Notebook product line, Packard-Bell/NEC would
not preinstall Navigator if Internet Explorer is already preinstalled
because "It wouldn't make sense to have two very large programs
installed using up the hard disk drive and it might be confusing to
the end user as to why two of the same applications were
included." Kies Dep. (played 11/17/98am), at 68:13-21.
- Based on this testimony and other evidence, Professor Fisher
concluded that "some OEMs preferred to load only one browser to
avoid user confusion and the resulting consumer support costs,
and to avoid increased testing costs." Fisher Dir. ¶ 150.
- Dr. Warren-Boulton explicated OEM testimony that having Internet
Explorer increases the costs, and reduces the benefits, of a second
browser (Warren-Boulton 11/24/98pm, at 59:18 - 59:25); further, he
testified that Internet Explorer support costs are significant costs to
OEMs. Warren-Boulton, 11/24/98pm, at 26:19 - 27:10.
(2) Additional testing costs
167.3. Preinstalling a second product in a given software category can
also increase the OEM's testing costs.
- Dr. Warren-Boulton testified that if OEMs purchase Windows
"already with IE on it, then you've got to test to make sure that your
system is compatible with the Windows IE bundle. If it didn't come
with IE on it, then you just would have to test it with Netscape. So
the point is that, I think as Mr. Kempin has pointed out, if you use
IE, you only need to test it once. If you want to use Netscape,
you've got to test it twice." Warren-Boulton, 11/30/98am, at 14:16 -
24.
- Professor Fisher testified that "some OEMs preferred to load only
one browser to avoid user confusion and the resulting consumer
support costs, and to avoid increased testing costs." Fisher Dir. ¶
150.
- Weadock testified: "It is certainly also important that we avoid the
testing costs associated with supporting a dual browser end-user
environment or an environment in which users click one place and
run one browser and click somewhere else and run another
browser. That increases an organization's testing costs, because
now they have to -- they can't rely on Navigator being the only
browser that users will activate." Weadock 11/17/98am, at 74:3-11.
(3) Opportunity costs
167.4. Preinstalling a second application in a given software category
also takes up scarce and valuable space on the computer's hard drive and desktop.
- According to John Kies, Senior Product Manager for the Packard-Bell/NEC Versa Notebook product line, Packard-Bell/NEC would
not preinstall Navigator if Internet Explorer is already preinstalled
because "It wouldn't make sense to have two very large programs
installed using up the hard disk drive and it might be confusing to
the end user as to why two of the same applications were
included." Kies Dep. (read 11/17/98am), at 68:13-21.
- Stephen Decker testified that Compaq stopped preinstalling
Netscape on the computers it sells because "with the inclusion of
Internet Explorer from Microsoft, that category is already filled
because of the inclusion of that product as part of the operating
system, and then also to actually license the additional browser
that would involve both time by Compaq to put that particular
agreement in place, we would have another product that would
take up real estate on our hard drive and, you know, there
potentially would be some additional licensing fees, and we would
have to pay for that technology. . . ." Decker Dep. (read
2/18/99am), at 61:8-21.
- Professor Fisher testified that "some OEMs viewed the desktop
and/or disk space as scarce real estate and were generally
reluctant to preinstall more than one software title in each
functional category." Fisher ¶ 151.
b. Microsoft's tie-in and associated contractual restrictions
raised the costs to OEMs of, and thus deterred OEMs
from, preinstalling Netscape and other non-Microsoft
browsers
168. Microsoft's conduct has deterred OEMs from loading Netscape (and other
browser rivals) and thus significantly contributed to Microsoft scheme to raise rivals'
costs and gain browser usage share.
- A Microsoft OEM sales manager, Candace Grisdale, responded as
follows to a May 1998 news article suggesting that Hewlett-Packard might
bundle Navigator on all its PC lines: "HP we've known was close to NSCP
but each time we've asked them of their plans, they have said they do not
want to carry the burden of two browsers, unless the customer segment
demands it." GX 323.
- Mal Ransom testified that Netscape approached Packard Bell about
preinstalling Navigator. Packard Bell seriously considered do so, but
decided not to because Packard Bell did not want to carry the burden of
two browsers on its machines. Ransom Dep. (played 12/16/98pm), at
74:12 - 75:6.
- Compaq removed Netscape once it was compelled by Microsoft to restore
the Internet Explorer icon to the desktop because of the increased costs
of supporting a second browser. See infra Part V.C.1.b.(2); ¶ 179.
- Professor Fisher testified that, "since Microsoft's tying arrangement
ensures that IE is on every Windows PC, the result is a significant
exclusionary effect that ensures that IE is the only browser on most PCs
shipped by OEMs." Fisher Dir. ¶ 152; see also Fisher, 1/6/99pm, at 12:21
- 13:2 (OEMs don't find it profitable to install Navigator because Internet
Explorer is already there).
- Dr. Warren-Boulton testified that a significant exclusionary effect can be
inferred from the fact that IE users are more likely than Netscape users to
have gotten their browser through the OEM channel. Warren-Boulton,
11/24/98pm, at 58:16 - 59:12.
- See generally infra Part VII.A (describing both the raising of rivals' costs
and its impact on OEM carriage of Navigator).
c. Microsoft's conduct similarly raised the costs to end
users of employing non-Microsoft browsers
169. Microsoft's conduct similarly raises the costs to end users of employing
non-Microsoft browsers.
(1) It is undesirable for a consumer who wants one
type of browser to have a different browser pre-loaded on his PC
170. For a consumer who desires to use a particular browser, the existence of a
different browser pre-loaded on the PC is not only superfluous but also, for several
reasons, undesirable.
- Professor Felten testified: "Providing code that people actually use is
efficient, but providing code that is not being used, or packaging code that
is not being used, or code that the user does not want along with code the
user does want, as Microsoft has done in this case, is not efficient and
makes things worse for the users." Felten, 12/14/98am, at 51:14-21.
- Professor Felten also testified that "in general if you know that the user
does not want something, it can only be inefficient to force them to take it.
And you will note that with respect to Internet Explorer in Windows 98, we
are talking about forcing the user to install software that they don't want
onto their hard disk. When they boot Windows, that software they don't
want is loaded into the memory. And as Professor Farber explained, in
some cases that software is even run, and Internet Explorer pops up even
though the user doesn't want it. It's certainly inefficient to do that. I also
want to point out that when I talk about efficiency, as a computer scientist,
I'm using it in the broadest sense. That is, I'm including -- I'm including
inefficiencies that develop, for example, because of user confusion
because of unwanted behavior. Mr. Weadock talked about the cost in
support calls and lost productivity because of those cases, and those also
apply in the case of Windows 98 and IE." Felten, 12/14/98pm, at 11:16 -
12:9.
- Weadock testified: "If an organization wants to standardize on a specific
browser for reasons of cross-platform access, then they may want to
choose a browser like Navigator, as opposed to a browser like Internet
Explorer, in which case they don't want Internet Explorer on the machine."
Weadock, 11/17/98am, at 43:5-10.
170.1. Unused software on a PC, particularly a program as large as a
modern browser, takes up scarce and valuable space on the user's hard drive and may
therefore increase hardware resource requirements.
- Professor Felten testified: "Forcing some users or OEMs to take
software they do not want is inefficient, since the unwanted
software needlessly uses resources such as disk space and
memory, and increases the complexity of the user interface by
cluttering it with unwanted icons, menu items, and programs."
Felten Dir. ¶ 67.
- Weadock testified that "if a user wants Windows 98, doesn't want
Internet Explorer and wants Netscape, then the commingling of
code between Windows 98 and IE may result in a total package
that uses more memory and more disk space than if Microsoft had
not commingled Internet Explorer and Windows 98. So, it's not a
benefit for everybody." Weadock, 11/16/98pm, at 44:16-23.
- Weadock testified: "The required hardware resources can increase
significantly when an operating system integrates application
software," and that "Windows 98 requires a great deal more disk
space and significantly more memory than Windows 95, largely
due to Internet Explorer software." Weadock Dir. ¶ 32d.
170.2. The drain placed on a system by additional, undesired software
increases when that software is "integrated" in some fashion with the computer's
operating system because code used by the operating system is more likely to
consume critical dynamic memory, or RAM, in addition to storage space on the hard
drive. For this reason, "integrating" certain additional software into the operating
system is also more likely to cause stability problems.
- Weadock testified: "The likelihood of an application failure affecting
the operating system may increase when code is shared between
the two." For example, "I have noticed in my own experiments with
Windows 98 that the failure of an Internet Explorer Window can
cause the entire desktop to malfunction." Weadock Dir. ¶ 32a.
- Weadock testified: "An application that modifies operating system
files could create (and, in the case of Internet Explorer, has been
documented in some cases to create) conflicts with other
applications and with company-developed applications." Weadock
Dir. ¶ 32b.
- Finally, Weadock testified: "It may become more difficult to enforce
security when an operating system integrates application software."
For example, "[s]ecuring the system against users running
programs that management doesn't want them to run becomes
more difficult as application software is folded into the operating
system." Weadock Dir. ¶ 32e.
- John Soyring testified that "integration could be inefficient and
disadvantageous to customers" because, for example, "integration
generally increases the size of the operating system, and therefore,
the size of the hardware required to run it effectively. In addition, it
may slow the use of other applications, and may provide function
which certain customers do not want." Soyring Dir. ¶ 25.
- James Gosling testified that, "in Windows 98, . . . Microsoft
apparently loads some browser-related files into memory even
when the user may never need that functionality . . . . In essence,
Microsoft simply shifts the time required to load the browser code
from when it is first needed by the user to every time the computer
boots up." Gosling Dir. ¶ 37.
- Felten testified that, "as more memory gets used up, the system
starts having to do complicated things to keep all the programs
running, and so use of more memory generally translates into
reduced performance." Felten, 6/10/99pm, at 18:16-20.
- In an email from Jonathan Roberts to Bill Gates and others in July
1997, Roberts points out that a 16 MB Navigator user "will have a
much slower experience with 98 than 95," because "if they access
help or an HTML page while in Explorer or in My Computer they will
be loading the IE HTML rendering engine and significantly
increasing the working set." GX 725.
- In December 1996, David Cole and his Internet Explorer
development team discussed "decoupling" IE4 from the Windows
shell, in part because "if the user installs the new shell, they will
have some things to learn and pay a performance price. By
coupling these together, I think the overall effort has suffered.
We've got a compromised new shell design that tries to be too
Windows 95 shell compatible in my view. We don't have HTML on
the desktop because we are worried about performance. But even
in compatibility mode, performance will degrade and there will be
differences that could stall adoption of the browser platform." GX
46.
- Professor Felten also testified: "There are two costs that come
from loading unnecessary code into memory. First of all, it takes
time to read that code off the disk, and that means that the
response time of some operation is slower because you spend
extra time loading this data into memory. Also, the unwanted code
takes up space in memory, and memory space is a limited
resource. Something else may have to get moved out of memory
or something else may not be able to work because the system has
run out of memory. The user ultimately might be forced to run out
and buy more memory or upgrade their PC in order to get enough
memory to loaded the unwanted code along with the code they
actually want." Felten, 12/14/98pm, at 57:10-22.
- AOL's Barry Schuler testified that
- redacted -
Schuler Dep., 5/5/99, at 136:21 - 137:6 (DX 2810A)
(sealed).
170.3. Unused software can also increase consumer confusion and
support costs by needlessly increasing the complexity of the user interface.
- Professor Felten testified that "in general if you know that the user
does not want something, it can only be inefficient to force them to
take it. . . . I also want to point out that when I talk about efficiency,
as a computer scientist, I'm using it in the broadest sense. That is,
I'm including -- I'm including inefficiencies that develop, for
example, because of user confusion because of unwanted
behavior. Mr. Weadock talked about the cost in support calls and
lost productivity because of those cases, and those also apply in
the case of Windows 98 and IE." Felten, 12/14/98pm, at 11:15 -
12:9.
- Weadock testified that it is "generally accepted practice among IT
managers in businesses large and small to put the least amount of
software on a computer that will do what their users need to do.
You just save all kinds of costs that way, all the way from resource
use to support and training." Weadock, 11/17/98am, at 41:17 -
42:16.
- Weadock testified that commingling operating system and
application code can create user confusion. For example,
Microsoft's fusion of Windows Explorer and Internet Explorer may
confuse consumers as to whether they are viewing local or remote
data. Weadock Dir. ¶ 32c.
(2) The hard-coding of Internet Explorer makes users
less likely to use Netscape with Windows 98
170.4. Microsoft's decision to frustrate the user's choice of default
browser in Windows 98 by forcing the use of Internet Explorer in certain situations
decreases the value of other browsers to end users.
- Professor Felten testified that a user can install Netscape Navigator
on Windows 98, "but there is a problem using it. And Professor
Farber described this the other day. He talked about what
happened when he tried to install Navigator on his Windows 98 PC.
And Internet Explorer kept popping up in his face, as he put it.
That's the sort of phenomenon that I refer to in my testimony as
hardcoding access to Internet Explorer 4. And so that hardcoding
does prevent the user from using Netscape Navigator in all the
cases where they would like to." Felten, 12/14/98am, at 27:8-19;
see also Felten 12/14/98am, at 29:6-17 ("there are also situations
where IE pops up when a user does not want it"); Felten,
12/14/98am, at 44:12-17 (although "Windows 98 gives the user
more choices than Windows 95 did, . . . there's one area in which it
gives the user less choice, which is web browsing").
- Dr. Warren-Boulton testified that he personally bought a Windows
98 machine, but "never had the intention of using IE, and I wound
up installing Windows Navigator." Nonetheless, "IE keeps popping
back up again. It is, perhaps, people more technically adept would
be able to avoid it, but I can't seem to avoid running into IE."
Warren-Boulton, 11/24/98pm, at 30:22 - 31:4.
- Dean Schmalensee conceded that, if the integration of a software
product into the operating system degrades the performance of a
different product or makes it less convenient for users to use that
product, that would be a harm to consumers. Schmalensee,
1/19/99am, at 39:21 - 40:3.
170.5. Microsoft's "hard-coding" of Internet Explorer to Windows 98 is
particularly burdensome for organizations that want to standardize on a non-Microsoft
browser.
- When asked whether organizations remain "free to standardize on
Netscape Navigator if they want," Weadock answered: "Not in the
case of Windows 98. They are not free to fully standardize on
Navigator because Windows 98 enforces certain user actions to
bring up Internet Explorer, so I would say no, they were completely
free to standardize on a product if the operating system imposes
requirements to use an alternative nonpreferred product in some
cases." Weadock, 11/17/98pm, at 7:17 - 8:2.
- When asked how the fact that Windows 98 comes with Internet
Explorer affects a corporation that has standardized on Navigator,
John Kies of Packard Bell/NEC stated that "it would require the end
user to uninstall Internet Explorer 4.0 or install next to it Netscape
Communicator. And this -- then they would have two browsers on
it, where most companies would just prefer to support one browser
and, once again, go back into the training issue. And most
corporations would not -- would prefer not to have any items in the
user interface that they're not ready to support." Kies Dep. (played
11/17/98am), at 72:8-19; see also Kies Dep. (played 11/17/98am),
at 72:8 - 73:7.
- In a presentation entitled "Why Internet Explorer 5.0?" Scott Vesey
of Boeing wrote, "We do not have a choice. Internet Explorer will
be installed as a component of our next generation desktop
operating system. The extent to which we might be able to
disengage it needs to be determined. Operating System
integration. Microsoft is unlikely to back away from their
commitment to integrating the Web browser into the operating
system." Under "Risks," Vesey noted, "Two browsers on all Wintel
mahcines. IE comes in the operating system and is available for
use as a browser. Netscape would have to be separately installed.
Not able to fully disengage browser. May be able to remove IE
icon from the desktop (to be determined). Windows Explorer can
still be used to access internet protocols: (ex. http and ftp). Difficult
to enforce Netscape as 'The Browser' on the Wintel environment."
GX 638, at TBC 000412.
- A "Win98 Browser choice matrix" prepared by Vesey included the
following option (among others): "Accept Win98 as is with full
Internet Explorer integration. Use IE as the 'Standard' browser.
Install Netscape Communicator as 'Alternate' browser." Under
"Impact/Risk" of this option, Vesey wrote "Possible DLL and
registry contention issues. User confusion about what browser
should be used. Developers will develop sites that require one or
the other browser requiring users to switch browser depending on
what site is being accessed." Another option presented is
"Remove Internet Explorer Desktop icon, disable browser function
for Web served documents" and "Install Netscape Communicator
as 'Standard' browser." The Impact/Risk of this option is "The
extent to which it will be possible to disable Internet Explorer." GX
633.
- In that same document, Vesey suggests that: "Installing Netscape
as an alternative browser may lead to user confusion about what
browser to use. Some sites may require a specific browser. We
have been working to minimize this possibility but this risk still
exists. --- Removing Explorer (or disabling the web access
capabilities) will likely prove to be impractical or impossible. This
will leave us with 3 likely alternatives: Install both browser, allow
user to select which browser to use. -- Install only Explorer, require
all web sites to be completely neutral. -- Remove Explorer (or
disable all web access capabilities) and install Communicator for
web access." GX 633.
- An internal Boeing document describing Boeing's "Browser
Decision History" discusses planning for 1999, and notes, "Internet
Explorer v5 will be so deeply embedded in NTW 5, Office 2000, &
Outlook 2000 that we will not be able to extricate it -- Therefore
both Netscape and Internet Explorer browsers will be installed on
Windows desktops -- Netscape will continue to be the standard
web browser, next step will be to determine how to constrain use of
Internet Explorer as the browser -- Continue to evaluate 1999
browser direction." GX 631.
- An internal Boeing presentation entitled "Enterprise-wide Web
Browsers for the Desktop" by Scott Vesey in March 1998, stated:
"Installing both web browsers may: Confuse users about which
browser to use. Increase end user support costs." GX 635, at 11;
see also GX 637.
170.6. Because of these costs, many firms will have to choose between a
preferred non-Microsoft browser like Netscape and Windows 98.
170.6.1. Many customers feel strong pressure to use Windows 98
as their operating system for various reasons, including a desire to remain compatible
with other users and a desire to use new hardware or peripherals that Windows 95
does not support. Many of these customers are likely to forgo installing Netscape
Navigator (or other browsers) on their computers.
- Weadock testified that "many customers (depending on their size
or profile) feel strong pressure to use Windows 98 for various
reasons, including the following: (a) The organization's customers,
suppliers, or clients are likely to use it, and business reasons exist
to use the same software that customers, suppliers, or clients use.
(b) Windows 98 brings new technological benefits, such as a more
efficient file system; support for new types of hardware, such as
high-capacity optical disks; support for new hardware devices, such
as printers and network cards; better control of power-saving
features on both desktop and portable PCs; Year 2000 compliance;
and a number of new and enhanced housekeeping utilities that
Microsoft claims (and I concur) can reduce support costs. . . . (c)
The organization relies on hardware that is being discontinued by
the manufacturer and replaced by hardware that doesn't work with
Windows 95 but that does work with Windows 98. . . . At some
point in the life cycle of an outdated operating system, computer
hardware manufacturers tend not to devote resources towards
making their newest products compatible with that outdated
operating system." Weadock Dir. ¶ 42.
- Boeing's Scott Vesey testified that, "in the long term," Boeing could
not continue to use Windows 95 but would eventually "have to
move forward in a new operating system version as hardware is
not supported by Windows 95." Vesey, 1/13/99, at 280:13-16.
- Vesey wrote that "The main reasons for moving to Internet Explorer
5.0 in Q2/3 of 1999 of the 18-month tactical plan are: We do not
have a choice. . . . The integration between Internet Explorer and
the desktop operating system cannot be fully disabled. . . . Our only
choice is whether we will install two browsers or just install Internet
Explorer." GX 637 (emphasis in original).
170.6.2. A number of corporations have chosen instead the costly
option of forgoing Windows 98 and the non-browser related benefits it provides in order
to use the browser of their choice and have reverted to the original retail release of
Windows 95 (which does not include Internet Explorer).
- Glenn Weadock testified that "some companies are resisting, or
electing not to use, Windows 98 largely or in part because it would
force them to have a two-browser desktop (for example, Chrysler,
where the Manager of Performance and Cost Management stated
that two browsers would increase support costs)." Weadock Dir. ¶
41.
- Weadock testified that many organizations have gone back to the
retail version of Windows 95, even though doing so entails various
costs, because "they have the greatest control over what
applications they can install onto it, because it is the cleanest
version of Windows 95. It doesn't contain software that they don't
want. And, in particular, it doesn't contain Internet Explorer, which
they may not want." Weadock, 11/17/98am, at 62:12-20; see also
Weadock, 11/17/98am, at 27:9-20 ("Boeing went back to the
original retail version of Windows 95")
- A survey conducted by Compaq in February 1998 of 283 PC
decision makers at US companies found that "About 80% of
companies wipe or reformat the hard drives of new desktops. . . .
The operating system re-installed most often are OSR2 and the
retail version of Windows 95. Large businesses lean more toward
the retail version of Windows 95." GX 1242, at 34.
d. Microsoft's conduct has caused other significant
inefficiencies and consumer harm
(1) Microsoft's commingling of the browser and
operating system reduces system performance
171. Microsoft's commingling of the code that supplies browsing and other
operating system functionality reduces system performance for customers that do not
desire to browse the web using Internet Explorer.
- See supra Part V.B.4.c(1); ¶ 169.
- Professor Felten testified that Windows 98 uses approximately 20% less
dynamic memory after the prototype removal program has removed
Internet Explorer web browsing, measurably improving performance. See
supra Part V.B.3.b(2); ¶ 154.2; see also Felten, 12/14/98pm, at 56:3-13.
- In an e-mail from Jonathan Roberts to Bill Gates and others in July 1997,
Roberts states "Even with the option to turn off the default loading of
Active Desktop, Windows 98 is inescapably most appealing to the Internet
Explorer user. Of course, the plan is Win 98 with Web integration
converts a huge base, but a die hard 16MB Nav user is hard to move. If
they access help or an HTML page while in Explorer or in My Computer
they will be loading the IE HTML rendering engine and significantly
increasing the working set. This means, that in many scenarios, the 16
MB nav user will have a much slower experience with 98 than 95." GX
725.
- Paul Maritz conceded that "in certain circumstances, applications in
general, not just Netscape's browser, can run slower on Windows 98
versus 95 in memory-constrained situations; in other words, running a
machine with smaller amounts of memory." Maritz, 1/27/99pm, at 4:7-16;
see also Maritz, 1/27/99pm, 4:17-23.
- Chris Jones's notes from a November 1997 offsite meeting among
Internet Explorer project team members report that "Performance overall,
in particular with integrated shell, is a problem. The IE 4 browser, while
fast, is simply too big for customers to install and adopt, both in terms of
memory usage (working set), and also in terms of disk footprint (install
size). The integrated shell adds additional requirements, and customers
are not deploying on 32 MB NT systems." GX 364, at MS7 004719.
- Gateway expressed concern to Microsoft in April 1998 that "the
installation of the full MS product (including channels) results in a much
slower system performance if the customer chooses an alternate browser
after full installation on IE4." GX 320.
- Weadock testified that if "we look at Windows 98, we see a situation
where Internet Explorer can't be removed, it takes up memory resources;
it takes up disk space. If a company can't remove that and then obtains --
to use your word -- Netscape Navigator, because there is so much RAM
and disk and processor overhead already associated with the
nonremovable Internet Explorer in Windows 98, they can't obtain
Navigator and put it to work on their system without a substantial
performance penalty, as I think one of the Microsoft employees in one of
the e-mails that I've seen expressed concern about." Weadock,
11/16/98am, at 63:1-11.
- Professor Farber testified that "combining applications with an operating
system into a single product available with all functions combined imposes
technical inefficiencies for OEMS, other software developers and retail
end users, including redundancy, performance degradation of unused
software and increased risk of 'bugs;' and . . . any function provided by an
operating system (as distinct from higher level files) that does not satisfy
the criteria of simplicity, general applicablitiy and accessibility reduces the
efficiency of the operating system environment and the applications that
use it." Farber Dir. ¶ 27.
(2) Microsoft's commingling of the browser and
operating system causes undesirable system
complexity, incompatibilities and security
concerns
172. Microsoft's commingling of the browser and operating system also
introduces undesirable system complexity and incompatibilities with other software.
- A November 1997 internal Microsoft memo from Brian Hall quotes
participants from Internet Explorer user focus groups as saying: "why do
we need to see local files through our web browser? It's like a whole other
version of windows explorer in a web browser. Need one or the other,
don't need both." GX 218.
- One of the recommendations of an ISP focus group consulted by
Microsoft was: "Turn off the Active Desktop. Didn't like that a browser
introduced UI changes -- they didn't want to be in the business of training
poepl [sic] how to use the UI when it is really a part of the OS." GX 375.
- Weadock testified that: "User confusion can result from combining
application code and operating system code, for example, as Microsoft
has done with Windows 98 and the "single Explorer." Weadock Dir ¶ 32c.
"An application that modifies operating system files could create (and, in
the case of Internet Explorer, has been documented in some cases to
create) conflicts with other applications and with company-developed
applications." Weadock Dir. ¶ 32b; see also Weadock 11/17/98am, 37:24
- 38:13.
- Professor Felten testified that "in general if you know that the user does
not want something, it can only be inefficient to force them to take it. And
you will note that with respect to Internet Explorer in Windows 98, we are
talking about forcing the user to install software that they don't want onto
their hard disk. When they boot Windows, that software they don't want is
loaded into the memory. And as Professor Farber explained, in some
cases that software is even run, and Internet Explorer pops up even
though the user doesn't want it. It's certainly inefficient to do that. I also
want to point out that when I talk about efficiency, as a computer scientist,
I'm using it in the broadest sense. That is, I'm including -- I'm including
inefficiencies that develop, for example, because of user confusion
because of unwanted behavior. Mr. Weadock talked about the cost in
support calls and lost productivity because of those cases, and those also
apply in the case of Windows 98 and IE." Felten, 12/14/98pm, at 11:15 -
12:9.
- Professor Farber testified that "combining applications with an operating
system into a single product available with all functions combined imposes
technical inefficiencies for OEMS, other software developers and retail
end users, including redundancy, performance degradation of unused
software and increased risk of 'bugs'; and . . . any function provided by an
operating system (as distinct from higher level files) that does not satisfy
the criteria of simplicity, general applicablitiy and accessibility reduces the
efficiency of the operating system environment and the applications that
use it." Farber Dir. ¶ 27.
- In a Boeing planning document written in July 1998 on deployment of 5.0
level browsers starting in 1999, Scott Vesey noted the following about
IE4: "The problem for Internet Explorer 4.0 is that it modified the Windows
95 operating system DLL's. Several conflicts were identified with
commercial software packages. All known conflicts have been resolved,
ether by the software vendor or by a Microsoft patch. However, concern
exists for conflics with Boeing custom written applications. Due to these
concerns and because there was not a compelling tecnology reason to
change the standard web browser vendor, the Netscape browser was
selected as the browser standard." GX 637; see also GX 632; GX 634;
GX 635; Vesey Dep. (played 11/17/98am), at 90:6-14.
- Professor Felten testified that "giving users or OEMs the choice of what
Web browsing software, if any, to have on their systems is technically
efficient. Although some users or OEMs may benefit from bundling
together separate software products desired by those users or OEMs,
significant inefficiencies may also arise for others. Forcing some users or
OEMs to take software they do not want is inefficient, since the unwanted
software needlessly uses resources such as disk space and memory, and
increases the complexity of the user interface by cluttering it with
unwanted icons, menu items, and programs. Had Microsoft originally
designed a version of Windows 98 without Web browsing, this version
would have been significantly smaller than the version Microsoft actually
released." Felten Dir. ¶ 67.
173. The bundling of a browser or other application software with an operating
system increases the vulnerability of the system to viruses or unauthorized usage.
- The following colloquy took place between the Court and Professor
Felten: "THE COURT: . . . Are there any security issues involved in a
choice of a browser or whether to get a browser at all? . . . It seems self-evident to me, but maybe it's not, that the presence of a browser
increases the risks of penetration by a virus or something like that. THE
WITNESS: Certainly. If you are in the position of, say, a computer
systems administrator in a large organization and you're concerned that
you less-trained users might accidentally introduce a virus or something
like that -- you might well choose to not have browsers on your users'
computers in order to prevent that means of spread of virus." Felten,
6/10/99am, at 39:18 - 40:7.
- Weadock testified: "It may become more difficult to enforce security when
an operating system integrates application software." For example,
"[s]ecuring the system against users running programs that management
doesn't want them to run becomes more difficult as application software is
folded into the operating system." Weadock Dir. ¶ 32e.
e. Dean Schmalenseee's testimony that Microsoft's
conduct did not result in significant competitive and
consumer harm is unreliable
174. Microsoft's testimony that consumers benefitted from its tying arrangement
and associated contractual restrictions (E.g., Schmalensee, 6/21/99am, at 10:5 - 11:5,
36:18 - 37:11), rests on faulty assumptions.
174.1. First, Dean Schmalensee conceded that he did not investigate
whether Internet Explorer could be removed from Windows 98 or why Microsoft made it
non-removable.
- When asked whether he investigated what functions of Windows
98 were removable, Schmalensee answered, "Absolutely not. It
seemed to me that the key issue was whether users had a choice
as to which software they employed, and whether it had to be
removed or just not used seemed to me completely immaterial."
Schmalensee, 1/20/99am, at 5:24 - 6:5; see also Schmalensee,
1/20/99am, at 10:5-9 (same).
- When asked whether he looked at "any internal Microsoft
documents to determine why Microsoft decided not to make
Internet Explorer removable from Windows 98," Schmalensee said:
"No. I did not. I inquired what they did, not -- as we've said several
times, not what they said about what they were doing."
Schmalensee, 1/20/99am, at 14:19-24.
174.2. Second, Dean Schmalensee also conceded that he did not
investigate the consequences of commingling software code in the same files.
- Schmalensee conceded that he has no "quantitative answer" to the
question of how much shared code there is between Windows 95
and Internet Explorer, and Windows 98 and Internet Explorer.
Schmalensee, 1/19/99pm, at 31:4-8. See generally Schmalensee,
1/19/99pm, at 32:10-18 (similarly conceding that he does not know
how many DLLs contained shared code in Windows 95 and
Windows 98).
- When asked what percentage of the Internet Explorer code in
Windows 98 is shared by other operating system functions,
Schmalensee testified that "I don't have that breakdown. I don't
know." Schmalensee, 1/20/99am, at 11:11-14; see also
Schmalensee, 1/20/99am, at 12:9-14 ("I can't answer that question.
And I didn't investigate it, because it's not important for the reasons
I've given you several times.").
174.3. Third, Dean Schmalensee implicitly assumes (Schmalensee Dir. ¶
232), contrary to the evidence, that a user can consistently enforce his or her choice of
default browser in Windows 98.
- See supra Part V.B.2.e(2); ¶ 147.5.
174.4. Fourth, Dean Schmalensee's and other Microsoft witnesses'
contention that Microsoft has merely offered an additional choice of browsers
(Schmalensee, 1/21/99pm, at 37:24 - 38:8, Schmalensee, 6/21/99am, at 37:4-7) is
wrong. While the development of Internet Explorer itself provided additional options to
users, its tying of the browser to the operating system denied users the option of
forgoing Internet Explorer and increased their costs of using other browsers.
- See supra Part V.B.4; ¶¶ 166-175.
C. Microsoft imposed a variety of other anticompetitive restraints on
the OEM channel in order to impede rivals such as Netscape
1. Microsoft imposed exclusionary restrictions on OEMs' ability
to modify the Windows desktop and start-up sequence
175. In addition to its tying arrangement, Microsoft forced OEMs to agree to
restrictions on their ability to modify the Windows desktop and start-up sequence. As
with its tying arrangement, Microsoft's OEM desktop and start-up restrictions raised
rivals costs, harmed consumers, and cannot be explained except as part of Microsoft's
effort to blunt the threat to its operating system monopoly posed by non-Microsoft
browsers.
a. Microsoft imposed the "Windows Experience"
restrictions in response to, and in order to stop, OEMs'
featuring Netscape Navigator more prominently than
Internet Explorer.
176. When Microsoft first released Windows 95, OEMs customized the Windows
95 desktop and start-up sequence in various ways designed to meet consumer
demand; and they did so with Microsoft's tacit or express consent.
176.1. OEMs operate in a competitive environment; to thrive, OEMs must
supply what their customers demand, differentiate their products, minimize their support
costs, and offer PCs that are easy to use.
- Professor Fisher explained: "OEM's are, in some sense, the
representatives of the consumer for certain purposes. They are in
competition with each other. They gain if they deliver what end
users actually want. They wouldn't care about the restrictions on
them if they didn't think that it mattered in their dealings with
consumers." Fisher, 6/2/99am, at 22:1-6.
- Dr. Warren-Boulton testified: "As Microsoft has acknowledged,
OEMs are in the business of satisfying their customers. They are
exceedingly unlikely to market a product that does not meet
consumer demand." Warren-Boulton Dir. ¶ 160.
- Garry Norris testified that IBM sought to "differentiate PC's that
were shipping from IBM versus those shipping from competitors" in
order to ensure that new users, "a large percentage" of whom "are
novices," understand "how to navigate . . . their PC." Norris,
6/7/99pm, at 42:14-25. "The reason" IBM "invested the millions of
dollars . . . was, in fact, to differentiate the IBM Aptiva 'out of box'
experience and end-user experience with that of our competitiors.'"
Norris, 6/7/99pm, at 56:10-18. Norris further testified that, if IBM
"designed a confusing or a poor" product, IBM would be negatively
impacted "in the market." Norris, 6/7/99pm, at 69:2-14.
- A Hewlett-Packard executive told Microsoft: "PC's can be
frightening and quirky pieces of technology into which" consumers
"invest a large sum of their money. It is vitally important that PC
suppliers dramatically improve the consumer buying experience,
out of box experience as well as the longer term product usability
and reliability." GX 309.
- Compaq's Celeste Dunn testified that the PC consumer segment is
"a really tough segment of the market, because there is a lot of
price point pressure." Dunn Dep., 10/23/98, at 26:19-25 (DX
2566).
176.2. With the release of Windows 95, OEMs sought to achieve these
objectives -- and to meet consumer demand -- by, among other things, customizing the
Windows start-up sequence to include tutorials and other features that run in the "boot-up" or "start-up" process before the user is presented with the Windows desktop.
- IBM. IBM developed an "out-of-box" (Norris, 6/7/99pm, at 46:18 -
48:25) experience that included the "IBM Welcome Center" in the
start-up sequence that instructed users on basic features of their
PCs, such as "how to use a mouse, how to run Windows 95" and
"what an icon is." Norris, 6/7/99pm, at 42:9 - 43:9. IBM placed the
Welcome Center in the start-up sequence because usability studies
showed that "a large percentage of new users are novices" and
IBM "wanted to ensure that they understood how to navigate, so to
speak, through their PC." Norris, 6/7/99pm, at 42:14 - 43:3. By
making new users "more self-sufficient" before they reached the
Windows desktop for the first time, the Welcome Center helped
IBM reduce support costs. Norris 6/7/99pm, at 59:1-15; GX 2191
(the Aptiva "Windows Guide" heightened "the consumer's
experience with new features of their computer system"); GX 2187
("If anything our welcome program helps acquaint the user with the
machine, mouse, Windows, and gently delivers them to the
Windows desktop.").
- Hewlett-Packard. Hewlett-Packard also developed an "out-of-box
experience" that included placing several screens in the start-up
process. Romano Dep., (played 12/16/98pm), at 38:7-14.
- Dell. Dell PCs automatically completed the set-up of all software
during the boot-up sequence. Kanicki Dep., (played 12/16/98pm),
at 65:3-18.
- Compaq. Microsoft allowed Compaq to "add Compaq welcome
screens and non-MS advertisements." GX 292.
176.3. Although Microsoft's original Windows 95 licenses prohibited
certain OEM modifications,
- Microsoft's Windows 95 licenses typically provided:
- redacted -
- But a Microsoft letter to Hewlett-Packard acknowledged that HP
permitted users a choice of an alternative desktop "shell," in
addition to the graphical user interface provided by the ordinary
Windows desktop. GX 294.
177. By contrast, when OEMs customized their PCs in ways that threatened
Microsoft's objective of gaining browser usage share, Microsoft moved quickly both to
enforce and to augment its restrictions.
177.1. Microsoft construed its licenses to prohibit OEMs from removing
the Internet Explorer icon from the desktop.
A Microsoft document reports that, in July 1995: "Some" OEMs
"want to remove the" Internet Explorer "icon from the desktop -- our
response should be that this is not allowed." GX 296.
Microsoft forced Compaq to restore the Internet Explorer icon to
the Windows desktop. See infra Part V.C.2.a(1); ¶¶ 200.2-5.
Kempin conceded that Microsoft denied OEM requests to remove
the Internet Explorer icon. Kempin, 2/25/99pm, at 64:6-12
(Gateway asked if it could remove the Internet Explorer icon, but
Microsoft denied the request. "We basically said, 'we thought we
had an agreement that' Gateway 'cannot change our code and our
work.'").
177.2. Microsoft also became very concerned when OEMs found other
ways to highlight non-Microsoft browsers and, in response, set out to impose new, more
rigid restrictions.
- On January 6, 1996, Bill Gates wrote Joachim Kempin, head of
Microsoft OEM sales, that: "Apparently a lot of OEMs are bundling
non-Microsoft browsers and coming up with offerings together with
Internet Service providers that get displayed on their machines in a
FAR more prominent way than MSN or our Internet Browser";
reminding Kempin that "[w]inning Internet browser share is a very
very important goal for us," Gates also requested "an analysis of
the top OEMs of what they are doing with browsers." GX 295.
- The resulting January 16, 1996, analysis showed that "other
internet solutions are not only on every hard disk, but are invariably
positioned on the desktop more strongly by OEMs than any MS
offerings" and that a number of OEMs had configured their PCs to
boot automatically into desktop shells that featured Netscape
Navigator but not Internet Explorer. GX 297 ("Compaq Presario
9546. Tabworks [Compaq's alternative shell that was presented to
users automatically] is on the system, but we started up in the
Win95 shell. Perhaps someone played with this box at the store
before us getting it. MSN is not on the desktop; AOL is on the
desktop. IE is on the desktop as well. If you start up Tabworks,
AOL, CIS, INN, Netscape, Prodigy, CNN (AOL's internet access
service) are all top level tabs. MSN and IE are not, nor are they in
the Tabworks index.").
- Less than a week later, Kempin wrote in his semi-annual report on
Microsoft's OEM Sales that a top goal for the upcoming months
was to "Make OEMs Support our Internet Efforts" and that one
thing Microsoft "Missed in 1st Half of FY 1996" was "Control over
start-up screens, MSN and IE placement." GX 401, at MS6
600019, -23.
- Kempin conceded at trial that he intended to solve the problem of
Internet Explorer placement through control over start-up screens.
Kempin, 2/25/99am, at 43:4-15.
177.3. Accordingly, throughout 1996 and into 1997, Microsoft forced
OEMs to agree, either as amendments to existing Windows licenses or in new Windows
licenses, to a series of restrictions on their ability to customize the Windows desktop
and boot-up sequence that Microsoft called the "Windows Experience."
177.3.1. Among other things, these restrictions:(1)
177.3.1.1. Prohibited OEMs from removing any icons,
folders, or start-menu entries that Microsoft places on the Windows desktop, including
the Internet Explorer icon;
177.3.1.2. Prohibited OEMs from modifying the Windows 95
start-up sequence;
177.3.1.3. Prohibited OEMs from configuring programs,
including alternative OEM shells -- alternative desktop screens and user interfaces, in
lieu of the Windows desktop -- as well as programs to complete the installation of
software, to launch automatically at the completion of the Windows start-up sequence;
and
177.3.1.4. Prohibited OEMs from adding icons or folders to
the Windows 95 desktop that were not similar in size and shape to Microsoft supplied
icons.
177.3.2. OEMs acquiesced in these restrictions because they had
no choice.
- Hewlett-Packard's John Romano wrote Microsoft, in
response to the imposition of the Windows Experience, that:
"If we had a choice of another supplier, based on your
actions in this area, I assure you [sic] would not be our
supplier of choice." GX 309.
- Dean Schmalensee conceded that OEMs lack any practical
alternative to Windows and thus must accept Microsoft's
terms. Schmalensee, 1/20/99am, at 33:3 - 34:9.
177.4. Microsoft broadened its OEM restrictions when it introduced
Internet Explorer 4.0, which included a feature known as the "Active Desktop," in order
to prohibit OEMs from using the Active Desktop to feature third-party brands on the
desktop.
- Microsoft introduced with Internet Explorer 4.0 a feature known as
the "Active Desktop." Schmalensee Dir. ¶¶ 465-467. This feature,
which overlays the standard Windows desktop, permits OEMs to
place on the desktop items that are both more interactive and
significantly larger than the icons placed on the standard Windows
desktop. Kempin Dir. ¶¶ 49-50; DX 2163 (demonstrating the Active
Desktop).
- Although the Active Desktop enables OEMs "to promote
prominently" particular products or services (Kempin Dir. ¶ 49)
GX 1201, at MS98
0102862 (Appendix A, OPK Requirements and Restrictions has
identical language as GX 458). Thus, while an OEM could place a
large item in the center of the desktop that reads, "click here to
learn more about your PC," DX 2163 (Kempin's videotape),
b. Microsoft's restrictions significantly increased the costs
to OEMs and end users of preinstalling or using non-Microsoft browsers
(1) Microsoft intended its restrictions to facilitate
winning the browser war
178. The origins of the Windows Experience (see supra ¶¶ 176-177)
demonstrate that Microsoft's purpose in enforcing and augmenting its OEM restrictions
was to gain browser usage share and thereby thwart the threat to its monopoly
presented by non-Microsoft browsers; other evidence confirms Microsoft's exclusionary
purpose.
178.1. Microsoft's restrictions imposed significant costs and inefficiencies
on OEMs and, ultimately, on OEMs' end-user customers.
- IBM. Although Microsoft representatives told IBM "that they
thought" IBM's "Welcome Center enhanced the end user's
experience," Microsoft forced IBM to scrap its out-of-box
experience. Norris, 6/7/99pm, at 61:14 - 63:25. IBM expected
support costs to rise as a result of the inability to offer its tutorial
screens in the start-up sequence. Norris, 6/7/99pm, at 49:20 -
50:15 (testifying that IBM anticipated support "costs going up" as a
result of Microsoft's restrictions); Norris, 6/7/99pm, at 70:16 - 71:17
(explaining that complying with Microsoft's restrictions would cause
IBM to "lose" its "investments, restart, and redevelop" its "Access
Aptiva" feature). Microsoft's restrictions also precluded IBM from
implementing other features useful for consumers, such as
"information screens assisting customers with installations such as
warning to avoid power cycling during setup." GX 2191. Although
IBM told Microsoft that its restrictions made "the system . . . harder
to use" and would confuse users, Norris, 6/7/99pm, at 55:19-25;
GX 2141 (July 30, 1996 letter from IBM to Kempin, stating
"restriction will negate the system tutorial features that are currently
booting and will make the system harder to use, particularly for
computer novices."); GX 2193 (notes of August 13, 1996,
IBM/Microsoft; IBM tells Microsoft that restrictions a "major blow to
brand and value"), Microsoft refused to lift all but trivial restrictions.
Norris, 6/7/99pm, at 65:7 - 66:20.
- Hewlett-Packard. Hewlett-Packard reported in a March 1997 letter
regrading Microsoft's restrictions that "Microsoft's mandated
removal of all OEM boot-sequences and auto-start programs for
OEM licensed systems has resulted in significant and costly
problems for the HP Pavilion line of retail PCs." GX 309. As "a
direct result" of Microsoft's restrictions, HP's "support calls went up
by approximately ten percent." Romano Dep., (played
12/16/98pm), at 41:20 - 42:6. The letter continued:
"From the consumer perspective, we are hurting our industry and
our customers. PC's can be frightening and quirky pieces of
technology into which they invest a large sum of their money. It is
vitally important that the PC suppliers dramatically improve the
consumer buying experience, out of box experience as well as the
longer term product usability and reliability. The channel feedback
as well as our own data shows that we are going the wrong
direction. This causes consumer dissatisfaction in complex
telephone support process, needless in-home repair visits and
ultimately in product returns. Many times the cause is user
misunderstanding of a product that presents too much complexity
to the common user."
"We strongly protested the changes last fall and were flatly refused
any leeway. We also, as of October of 1996, began to request
what we would be able to do in the future. We were continuously
put off as to what the future policy would be. But we were led to
believe that we would be granted some options."
"Our Customers hold HP accountable for their dissatisfaction with
our products. We bear for the cost of returns of our products. We
are responsible for the cost of technical support of our customers,
including the 33% of calls we get related to the lack of quality or
confusion generated by your product. And finally we are
responsible for our success or failure in the retail PC market." GX
309.
- Compaq.
- redacted -
GX 120 (sealed); see also Dunn Dep.,
10/23/98, at 130:15 - 134:2 (DX 2566) (sealed)
GX 756, at COM-8-000048 (sealed). Compaq more recently expressed concern that
the "Windows desktop experience" means "(1) loss of revenue to
Compaq for ISP subscriptions, (2) loss of revenue due to inability to
send customers to our portal, (3) customer dissatisfaction due to
confusion over the best way to get onto the Internet, and (4)
customer dissatisfaction due to non-functioning ISPs." GX 326
(9/98). See also infra Part V.C.2.1(a); ¶ 200.5 (Compaq voluntarily
adhered to the restrictions).
- Dell. Dell determined that compliance with the Windows
Experience would increase technical support costs, "confuse and
disappoint repeat customers," and result in the "perception of Dell
will be negatively effected." GX 307.
- Gateway. Gateway concluded that Microsoft's restrictions
"jeopardized" its "direct relationships with" its "customers" and
"limits choice for the majority of end users." GX 302. Gateway also
wrote Microsoft that it needed "to be able to remove icons if the
customer does not choose those options" in order "to remove as
much clutter from the screen as possible." GX 319. The document
summarized: "Generally, Gateway wants to have flexibility on
anything associated with the Internet. We want MS to provide us
with the technology, not make decisions and choices for us, or our
customers." GXs 319, 320 (same).
178.2. The costs imposed on OEMs and consumers by Microsoft's
restrictions reduced the value of Windows and impaired Microsoft's goodwill; Microsoft
would not rationally have inflicted these costs on OEMs and itself unless it expected a
corresponding gain through excluding browser rivals.
- Warren-Boulton Dir. ¶¶ 189-94.
178.3. Microsoft selectively enforced its Windows Experience restrictions,
further evidencing the anticompetitive purpose behind these restrictions.
- Kempin admitted that there is "no general rule" concerning when
Microsoft enforces its restrictions; it examines each situation "in a
pragmatic way." Kempin, 2/25/99pm, at 93:24 - 94:3.
- Microsoft granted certain requests for exceptions to its restrictions
but has generally declined to grant exceptions that would
undermine its objective of making it difficult for OEMs to feature
browser rivals. See infra ¶¶ 186-187.
- Kempin explained the exclusionary purpose of the selective
enforcement in a September 17, 1996, memorandum to Bill Gates,
in which Kempin, noting that OEMs had acquiesced in the
Windows Experience "only reluctantly" and had pushed back in a
number of areas, concluded that Microsoft should relax certain
restrictions but "at the same time . . . ensure that no other
advertising other than promoting the OEMs name brand can be
done during" the "'extended' boot process." GX 304. To allow
OEMs to fully control the boot-up sequence, Kempin wrote that it
would undermine Microsoft's "strategic objectives," which included
promoting Microsoft's "platforms." GX 304.
- Reflecting Kempin's description of how the Windows Experience
furthered Microsoft's "strategic objectives," Microsoft planned to
move the Internet Connection Wizard to the Windows 98 start-up
sequence because doing so "increases the likelihood that an end
user gets the option to sign up for solutions that promote IE before
they get into the desktop or any customized shell that features
other browser solutions." GX 176A.
178.4. Microsoft viewed alternative OEM shells as an important facet of
the threat non-Microsoft browsers posed to its operating system monopoly, not only
because shells could offer other browsers favorable promotion, but also because
alternative OEM shells could hasten the adoption of Netscape's APIs; Microsoft thus
had a particular anticompetitive reason to impose the restrictions.
- Brad Chase testified that Netscape was "'also trying to control and
define user interfaces, which is another important element of an
operating system.'" Fisher Dir. ¶ 89 (quoting Chase Dep., 3/25/98,
at 39:6-8).
- Jim Allchin testified: "'The stated goal of Netscape was to replace
the user interface of Windows were you couldn't see it and to
create a new set of API's that developers would write to. It's
exactly what we see Sun doing as well.'" Warren-Boulton,
12/1/98am, at 39:20-23 (quoting Allchin Dep., 3/19/98, at 119:21-25).
- Professor Fisher testified: "The browser could also threaten the
operating system monopoly by providing an alternative user
interface that would reduce users' reliance on the dominant
operating system interface. By providing a popular alternative user
interface, browsers could reduce consumers' resistence to non-Windows operating systems and enable businesses that utilize the
alternative interface to use different operating systems without
increasing their training and support costs. In turn, this would
reduce Microsoft's power to exploit the value of its interface real
estate by requiring other companies to promote Microsoft's
products through exclusive agreements." Fisher Dir. ¶ 89.
(2) Microsoft's restrictions significantly raised the
costs of its browser rivals, thereby impairing their
ability to compete and harming consumers
179. Microsoft's restrictions, in particular its prohibition on removing the Internet
Explorer icon (and other visible means of end-user access to Internet Explorer, such as
its entry on the start menu), raised the costs to OEMs of preinstalling rival browsers on
the PCs they sell.
179.1. Including icons for additional browsers on the Windows desktop
increases clutter and customer confusion, which translates into increased support costs
the OEMs must bear.
- Gateway. Gateway wanted to remove the Internet Explorer icon
from both Windows 95 and Windows 98 because Gateway
determined that removing the icon would reduce its support costs
should a user choose a Gateway-supplied option to install
Netscape's bowser. Von Holle Dep., 1/13/99, at 304:17 - 305:12,
310:24 - 312:21. "General usability studies," a senior Gateway
executive testified, "indicate that the less cluttered the desktop . . .
the less confusing it is for the customer to use the product." Von
Holle Dep., 1/13/99, at 305:9-12; see also GX 320 ("We want to
remove as much clutter from the screen as possible."); GX 652, at
ATR-30008 (sealed)
Kempin, 2/25/99pm, at 63:9 - 64:15 (acknowledging Gateway
raised these concerns and that its request was denied).
- Packard Bell. John Kies testified that Packard-Bell likely would not
preinstall Netscape's browser if the Internet Explorer icon were not
on the desktop, among other reasons, because "it might be
confusing to end users as to why two of the same applications
were included." Kies Dep., (played 12/16/98am), at 17:4-11.
Indeed, Packard Bell's Mal Ransom testified that to load Netscape
if Internet Explorer is on the desktop, Netscape would have to
"bring" Packard Bell "a business proposition in the way of financial
remuneration." Ransom Dep., (played 12/16/98pm), at 78:13 -
79:8.
- Compaq. Compaq's Steven Decker testified that Compaq stopped
preinstalling Netscape because "with the inclusion of Internet
Explorer from Microsoft, that category is already filled because of
the inclusion of the product with the operating system, and then
also to actually license the additional browser that would involve
both time by Compaq to put that particular agreement in place" as
well as having "another product that would take up real estate on
our hard drive" as well as any "additional licensing fees." Decker
Dep., 2/18/99am, at 61:2 - 63:19.
179.2. These cost increases caused by Microsoft's restrictions distort
competition between Internet Explorer and other browsers and in some cases are
significant enough to deter OEMs from preinstalling Netscape Navigator altogether.
- Microsoft emphasized to OEMs the increased costs of supporting
and testing a second browser and Kempin testified that this
argument was "'sometimes successful'" in "'persuading OEMs' that
they don't really need to distribute another browser.'" Kempin,
2/25/99am, at 62:25 - 65:6 (quoting Kempin's deposition).
Kempin's attempt to distance himself from this admission by
asserting that "OEM's at the end of the day ship what end users
demand and a small amount of support costs will not make them
basically change their plan," Kempin, 2/25/99am, at 65:19 - 66:7, is
contradicted by other evidence regarding OEM installation of
browsers. See infra Part VII.A.2.b; ¶ 363.4 (OEM bundling of
Netscape has decreased dramatically since Microsoft embarked on
its predatory campaign).
- Compaq, as explained, stopped preinstalling Netscape on its PCs
when Microsoft forced Compaq to restore the Internet Explorer icon
to the Windows 95 desktop. See supra ¶ 179.1; infra ¶ 203.
179.3. The contrary and shifting testimony of John Rose (2/18/99am, at
63:20 - 64:6, 67:1-24) -- that Compaq removed Netscape not, as Steve Decker testified,
because the "category was filled," but rather because of its obligations to AOL and/or
concerns regarding compatibility -- is inconsistent with his other testimony, is not
supported by the documents on which it purports to be based, and is not credible.
- Rose had no first hand knowledge concerning why Compaq
removed Netscape (Rose, 2/18/99am, at 67:1-24) and did not even
know until his cross-examination that Compaq ceased preinstalling
Netscape on consumer lines in 1996. Rose, 2/18/99am, at 62:14-17; Rose, 2/19/99am, at 52:2-7.
- Rose's testimony that Compaq generally does not load two
applications in a similar software category on its personal
computers because doing so "puts a greater cost burden on
Compaq; adds more complexity; causes confusion to the
customers, particularly consumer customers, that don't have any
personal computing experience" (Rose 2/18/99pm, at 45:25 -
48:14) supports Decker's testimony that Compaq removed
Navigator because with Internet Explorer, the category was filled.
- Microsoft points to documents which show only that AOL might
have objected to Compaq's bundling of Netscape with another
Internet access provider, not that AOL would have objected to
Compaq's inclusion of Netscape (or another browser) in other
ways. DX 2261, at COM-13-000040 (Compaq-AOL agreement)
(Compaq's obligation was to feature the AOL service and GNN
(defined in the agreement as a "direct Internet service provider,"
not a browser) "above all other Online Services within the user
interface of its Products in accordance with the attached
addendum"; the addendum refers only to "online services" and
"online services folder," not to browsers) (emphasis added).
- Moreover, nearly all the documents relied on by Rose, in support of
his assertion that Compaq's removal of Navigator was in response
to AOL, in fact support the conclusion that AOL's complaints were
limited to other online service/ISPs, not browsers, and did not
directly apply to Compaq's positioning of Netscape as a separate
application. DX 2376 (5/13/96 letter from Ted Leonsis of AOL to
Celeste Dunn) ("The terms were set to insure that both services
were featured in their respective categories, AOL as the online
choice and GNN as the Internet choice. We believe that the
current positioning of Netscape/Spry Internet service does not
recognize GNN as the featured product."); DX 2378 (5/22/96 letter
from Celeste Dunn to Ted Leonsis of AOL) ("In reviewing our plans
for AOL on our new consumer line, I am confident that we have
met and exceeded the obligations in our contract for both AOL and
GNN. Both AOL and GNN are featured above all other online
services as defined in the contract."); DX 2374 (6/3/96 letter from
Ted Leonsis of AOL to Celeste Dunn) ("Your letter states 'since we
do not include Internet access in the same category as a full-service online service, we have included an additional Internet
service provider icon on our desktop.' If Compaq is including such
an additional icon, then it must go to GNN . . . . Moreover,
providing this icon to any other provider, including Netscape
bundled with an ISP, would violate part 2 because the language
contemplates placement of such a provider in an innocuous (i.e.,
Other) grouping. Your letter's reference to reasons for GNN not
occupying this spot are, frankly, unacceptable -- we do not see
where Netscape's browser fits into our agreement except that
where it is bundled with an ISP (other than GNN), the two are to be
relegated to the 'Other' grouping." (emphasis added)).
180. The prohibition on removing the Internet Explorer icon similarly increases
the costs to end users of standardizing on a non-Microsoft browser because of
increased support costs and thus reduces the value of Windows to those customers.
- See supra Part V.B.4.c; ¶¶ 169-170.
- Boeing's Scott Vesey explained that "[h]aving two web browsers on the
desktop will confuse users" and increase support costs. GX 637. "End
user confusion," he explained, could manifest itself in several ways,
including, "applications or sites that require a particular browser," "specific
configuration settings" that do not "work within one browser or the other,"
or "an expectation that . . . a particular browser would function in a
particular way . . . based on experience." Vesey Dep., 1/13/99, at 290:4-21. Vesey further testified that, to encourage the use of a single browser,
Boeing "probably would" delete the desktop icon of the second browser.
Vesey Dep., 1/13/99, at 289:13 - 290:3. (explaining that to avoid these
costs, large end-users end up paying firms to remove the Internet
Explorer icon in order to enforce a corporate standard on another
browser, a cost they need not bear if they standardize on Internet
Explorer).
- Glenn Weadock testified that eliminating clutter can reduce end user
confusion. Weadock, 11/17/98am, at 34:16 - 35:3.
181. Microsoft's prohibition on altering the start-up sequence also impeded
Netscape's ability to promote its browser because the Windows start-up sequence is an
important point of contact with the features and software available on a PC.
- Garry Norris testified that IBM concluded, based on usability studies, that
its "Welcome Center" features was much less likely to be used if it were
merely one of many icons on the Windows desktop. Norris, 6/7/99pm, at
58:4-20.
- Hewlett-Packard's John Romano testified that an icon on the desktop is
not as "effective" as placement in boot-up sequence. Romano Dep.
(played 12/16/98pm), at 54:13 - 55:7.
- Microsoft's Steve Bush expressed concern that OEMs, if permitted to
supply their own registration mechanism, could "interject themselves into
the first boot customer experience and offer customers valued added
services before a Microsoft proposition of similar services." GX 313.
- Kempin wrote that it was vital to Microsoft's "strategic objectives" to
"ensure that no other advertising other than promoting the OEMs name
brand can be done during this 'extended' boot process." GX 304.
- Microsoft employees expressed concern that a Windows tutorial program
"won't attract end users' or OEMs' attention" if it is "not part of the boot
sequence or in a very prominent position." GX 742.
- Microsoft moved its Internet Connection Wizard from the desktop to the
start-up sequence in Windows 98 because it believed it would be more
likely to be used in the start-up sequence. "Adding the ICW (Internet
Connection Wizard) to the welcome experience should help drive up the
percent of new internet sign-ups which originate from Windows." GX 764,
at MS98 0114749; see also McCLain Dep., 1/13/99, at 622:22 - 623:4
(conceding that users are "less likely to invoke the Interent Connection
Wizard from the desktop" if they have already accessed it from the
"'Welcome to Windows' screen"); id. at 625:10-20 (conceding that having
the signup screen appear before the "Welcome to Windows" screen
"forces the customer to go through that option"); GX 176 ("In order to
protect our position on the desktop and increase the likelihood that IE gets
the prominent position with the end user we should move the Sign Up
Wizard into the boot up sequence some where.").
- John Soyring of IBM, who noted that he is "quite familiar with computer
operating systems," testified that only end users with sufficient technical
skills are able to reconfigure their machines so that they boot up directly
into an alternate browser shell. When Soyring himself tried it with an IBM
Thinkpad, the browser was on the machine's hard drive and not in "an
obvious place for me." He "searched through some folders to find it," and
it took some time. Soyring, 11/18/98pm, at 18:2 - 19:18.
- Dr. Warren-Boulton testified that "the start-up sequence is a particularly
effective vehicle for promoting software or services, especially for firms
interested in promoting their services to first-time PC purchasers."
Warren-Boulton Dir. ¶ 122. He further explained: "The principal reason
why it is advantageous to have a 'first to market position' in the boot-up
sequence is simple: It is apparently quite difficult to induce new browser
users to switch to a competing browser" a fact "confirmed by Microsoft's
own documents, which conclude that 'it is very hard and expensive to
make people switch [browsers].'" Warren-Boulton Dir. ¶ 123 (quoting GX
39).
182. Microsoft's restrictions reduced OEMs' ability to promote Netscape in other
ways.
- Dr. Warren-Boulton testified that Microsoft's prohibition on OEMs
automatically booting their PCs into an alternative OEM shell, the original
impetus for the Windows Experience, inhibited OEM efforts to highlight
other browsers over Internet Explorer. Warren-Boulton Dir. ¶¶ 177-78.
- Professor Fisher testified that Microsoft's prohibition on changing the size
and shapes of icons further has "the effect . . . to limit the ability of OEMs
to promote other browsers by, for example, highlighting the existence of
another browser with a large desktop icon." Fisher Dir. ¶ 148.
183. By raising the costs to OEMs and end users of employing alternative
browsers, and by preventing OEMs from booting directly into alternative OEM shells,
Microsoft's restrictions facilitated the protection of its operating system monopoly.
- See infra Part V.A.2.b; ¶¶ 362-363.
184. Microsoft's restrictions caused other substantial consumer harm.
- Microsoft's restrictions compelled OEMs to remove tutorial programs,
auto-booting alternative OEM shells, and other features that permitted
OEMs both to differentiate their products and to make the PCs they sell
easier to use. See supra ¶ 179.1.
- Microsoft's restrictions resulted in a dramatic increase in OEM support
costs (see supra ¶ 179.1), which both indicates substantial consumer
confusion and which, in the competitive market in which OEMs sell, can
be expected to be passed on to consumers in the form of higher prices.
- Microsoft's restrictions constrained users' choice of web browsing
software and made more costly their ability to standardize on a particular
web browser. See supra ¶ 179.1.
185. Microsoft's contention that its restrictions have not materially raised the
costs to OEMs of featuring Netscape Navigator or otherwise impeded Netscape's ability
to gain promotion and distribution through the OEM channel is unpersuasive.
185.1. Testimony by Microsoft's witnesses, in particular Dean
Schmalensee, that Microsoft's contracts permit OEMs to promote other browsers on the
portion of the Windows desktop that Microsoft permits OEMs to control (Schmalensee
Dir. ¶ 358 ("OEMs can add any additional programs they wish, including Web-browsing
software."); see also id. ¶¶ 361-63; Kempin Dir. ¶ 19) ignores the ways in which
Microsoft's contracts in practice raise OEMs' costs of including other browsers and
thereby, as a practical matter, deter OEMs from doing so.
- Dean Schmalensee conceded that OEMs in practice take
advantage of only "a small fraction of the Windows desktop
available to them." Schmalensee Dir. ¶ 365; DX 2163 (Kempin's
videotape demonstration); A Gateway Account Manger also
admitted that OEMs do not add many icons to the Windows
desktop. McClain Dep., 1/13/99, at 615:13 - 616:14.
- As explained, including icons for additional products -- particularly
when a comparable product in the same category is already
included, as is the case for browsers -- imposes significant, and in
some cases prohibitive, costs. See supra ¶ 179.1.
185.2. Microsoft's argument that any raising of rivals' costs is
insubstantial because a number of OEMs install multiple browsers, including the
Encompass browser and Netscape Navigator (Kempin Dir. ¶¶ 21-22 (testifying that "a
number of major OEMs . . . currently preinstall Netscape's web browsing software" and
that "more recently, leading OEMs . . . have begun installing browsing software
developed by Encompass, Inc.")), is wrong.
185.2.1. First, the Encompass browser is not a true web browser,
but rather consists of a shell that is built "on top" of Internet Explorer and relies on
Microsoft's APIs. Microsoft allows OEMs to customize such shell browsers in
conjunction with ISPs. The costs to OEMs of supporting the Encompass browser and
thus are lower than the costs to OEMs of supporting Netscape while not being
permitted to remove the Internet Explorer icon.
- Kempin's videotape (DX 2163) shows several OEM-customized Encompass browsers. OEMs collect most of the
ancillary revenues from this product. See infra Part
V.G.6.b; ¶ 313.2.2.1.
- Dr. Warren-Boulton explained that the contrast between
Microsoft's sacrifice of ancillary revenues, through its shell-browser, and Netscape's efforts to collect such revenues "is
a nice example of the distinction between what Netscape is
trying to do in the browser market, which is to make money,
and what Microsoft's trying to do in the browser market,
which is to control the technologies." Warren-Boulton,
12/1/98am, at 14:8-11.
185.2.2. Second, even though some OEMs continue to offer
Netscape on some of their PCs, the OEM restrictions caused the number of OEMs
offering Netscape, and the number of PCs on which they offered Netscape, to decline.
- See infra Part VII.A.2.b; ¶ 363.4.
185.3. Mr. Kempin's testimony that OEMs could run another operating
system prior to initiation of the Windows start-up sequence without running afoul of
Microsoft's restrictions, and could feature a rival browser either there or in the "BIOS" --
the software that turns on the computer and runs prior to invoking Windows (Kempin ¶
43 (license agreement allows OEMs to "run a small operating system such as Caldera's
DR-DOS from the BIOS before Windows starts, and launch all sorts of programs")) --
ignores the evidence from OEMs and Microsoft that extensive customization through
either of these mechanisms would have been too costly to be a practical alternative.
- John Romano testified that Hewlett Packard did not place its
Personal Page Program before the boot-up sequence because of
the "cost and time" in creating a "technically viable" program.
Romano Dep., 12/16/98pm, at 42:12 - 44:2.
- A Microsoft document from "Brad Chase's Online files" concludes
that "most OEMs won't go through the hassle to develop" a utility
that runs before Windows boots. GX 176A.
- Each OEM featured in Joachim Kempin's videotape had only a
brief "splash screen" in the BIOS. DX 2163.
- Compaq's Celeste Dunn testified that the BIOS "supports the
hardware peripherals" and that "to run things like commercials or
other things" in the BIOS one would need "to have the resources of
the full operating system available to you, plus the things supported
by the BIOS." Dunn Dep., 10/23/98, at 190:15 - 191:6 (DX 2566).
185.4. While Microsoft witnesses stress that OEMs are permitted to
include icons on the desktop that enable the user, with a few mouse-clicks, to install
Netscape as the default browser or to launch an alternative desktop featuring Netscape
that appears each time the user subsequently boots the PC (Kempin, 2/24/99am, at
70:13-19 (testifying that "the OEM is free to display whatever it wishes, and can even
replace the Windows 98 desktop entirely")), these options are substantially less cost-effective for both OEMs and non-Microsoft browser providers than the options Microsoft
blocked.
- Hewlett-Packard dropped its alternative OEM shell because it "was
not as effective of a way to" reach customers when placed "after
the bootup sequence"; on the desktop, "there was much less
likelihood" that consumers would successfully invoke it. Romano
Dep. (played 12/16/98pm), at 37:1-8, 54:7-22.
- End users are much less likely to obtain the browser they use
through these methods than through the channels Microsoft
foreclosed. See infra Part VII.A.2.c; ¶ 366.
c. Microsoft's recent relaxation of some of its restrictions
eliminates neither the most anticompetitive restrictions
nor the restrictions' past effects
186. Microsoft over time relaxed certain requirements of the "Windows
Experience."
- In approximately March 1998, Microsoft permitted some 50 OEMs to
include the ISPs of their choice in Microsoft's Internet Connection Wizard.
Kempin Dir. ¶ 12.
- Through letters issued in late May and early June 1998, Microsoft granted
seven OEMs the right to add their own Internet Connection Wizard and
registration server to the boot-up process. Kempin Dir. ¶12.
- Microsoft permitted certain OEMs to include other features in the start-up
sequence. GX 1121, at MSV 000752 (sealed)
- redacted -
GX 1201 (Windows 98 OPK), at
MS98 0102861-62 (describing the limits on OEMs' ability to customize the
Windows desktop and start-up sequence).
187. These changes, however, do not eliminate the restrictions' anticompetitive
character.
187.1. First, the restrictions had a substantial effect when in place.
- See infra Part VII.A.2.b; ¶ 363.
187.2. Second, these changes did not remove the restrictions that most
significantly raise the costs to OEMs of featuring Netscape, such as the prohibition on
removing the Internet Explorer icon from consumer systems, the prohibition on
providing alternative OEM shells that load automatically at the end of the start-up
sequence, and restrictions on featuring Netscape.
- Microsoft's letters to the OEMs allowed them to add their own
registration wizard and ISP sign-up process provided that such
features "shall contain no third party advertising or product
promotions other than those from the [OEM] or the ISP being
signed up." E.g., GX 1195, at MS98 011385 (Letter to Packard
Bell/NEC May 27, 1998).
- Kempin testified that OEMs may feature both Internet Explorer and
OEM-branded browser shells built on top of Internet Explorer, but
they are barred from featuring a third-party browser, such as
Netscape, in the ISP sign-up process. Kempin, 2/25/99am, at 6:1-15.
- Although Kempin at one point testified that OEMs were barred only
from advertising Netscape, not from providing a choice of Netscape
(Kempin, 2/25/99am, at 13:12 - 17:16), he testified both earlier and
under questioning from the Court that Microsoft's letters prohibited
OEMs from offering Netscape in the ISP sign-up process. Kempin,
2/25/99am, at 6:1-15; Kempin, 2/26/99am, at 11:5 - 12:9.
187.3. Although Microsoft has subsequently granted one OEM,
Gateway, permission to offer users the choice of Netscape in the Gateway.net ISP
sign-up process (Kempin, 2/26/99am, at 5:22 - 8:13), Microsoft continues to make that
option costly by prohibiting Gateway from removing the Internet Explorer icon.
- Microsoft continues to prevent Gateway from removing the Internet
Explorer icon. See supra ¶ 179.1.
- Gateway executives testified that it now offers the Netscape option
only to Gateway customers who preselect Gateway.net as their ISP
and pay subscription fees that offset some of the costs caused by
the inability to remove the Internet Explorer icon. Brownrigg Dep.,
1/13/99, at 126:6 - 128:1; Von Holle Dep., 1/13/99, at 311:21 -
312:21.
d. Microsoft's justifications for its restrictions are
pretextual and belied by the evidence
188. Microsoft's witnesses, principally Joachim Kempin, advanced a number of
justifications for its OEM restrictions. These justifications are inconsistent with
Microsoft's actual conduct and the contemporaneous evidence and are pretextual.
(1) Microsoft's purported concern with consistency
of the user experience cannot explain its
restrictions
189. Microsoft's suggestion that its restrictions preserve a "consistent
experience" for end users (Kempin Dir. ¶ 10) is not credible because Microsoft allows
OEMs to take numerous actions that are inconsistent with its professed interest in
preserving a "consistent experience."
- Microsoft permits OEMs to ship Windows 98 with the Active Desktop and
its channels either on or off. Poole, 2/8/99am, at 11:14-17 (DX 2115,
Poole's videotape demonstration). The experience of a user booting up a
machine with the Active Desktop (including the channel bar) enabled
could be very different from that of a user booting into the "standard"
Windows desktop.
- Microsoft permits several OEMs to customize the Windows start-up
sequence in various ways, so that each PC presents users with a different
experience, including, for example, providing very different introductory
tutorials, OEM-related advertisements, and registration screens. DX 2163
(Kempin video). Compaq even has a lengthy "movie" promoting Compaq
in the start-up sequence. DX 2163.
- Microsoft permits the largest OEMs to include their own Internet
Connection Wizard in the start-up sequence. Kempin Dir. ¶¶ 57-58.
- Microsoft itself recognizes that its conduct is inconsistent with its effort to
justify the restrictions; as one Microsoft employee wrote upon learning that
Kempin had granted OEMs greater flexibility with respect to the start-up
sequence, "the reaction from DaveHe and the antitrust team was
negative. Changes like this undermine our whole case in defense of
Windows Experience." GX 379 (emphasis in original).
- When Microsoft permits OEMs to promote products more prominently
than Internet Explorer, it specifically prohibits OEMs from promoting third
party brands, including Netscape. GX 1201 (no third party branding in the
Active Desktop); GX 1159, at TM 000057 (OEMs cannot advertise or
promote third party branding in ICW or reg. wizard in boot-up sequence).
This distinction between OEM brands and third party brands has no
relationship to the "consistent experience" of end users because different
OEMs promote their brands in different ways. DX 2163 (Kempin video).
- Kempin testified that Microsoft prohibits third party advertising in the ISP
sign up process because that "process . . . . should not be interrupted by
any advertising." Kempin, 2/24/99pm, at 79:18-19 (emphasis added).
- redacted - 1190 (sealed), at MS98 008924;
Warren-Boulton Dir. ¶ 175.
- Kempin elsewhere made clear that Microsoft's purported concern with
"consistency" easily gives way when Microsoft seeks to accomplish its
"strategic" goals. GX 304. Although Microsoft told OEMs that "all OEM's
of Windows 95 are treated equally by" Microsoft "in meeting the same
OPK requirements for shipping Windows 95 PC's to ensure the customer
of a consistent experience" (GX 294), in practice Microsoft favored
Compaq by permitting it to customize the start-up sequence in Windows
95 in ways it denied to other OEMs (Romano Dep. (played 12/16/98pm),
at 52:10 - 54:9), and by permitting it to include its own ISP sign-up
process before other OEMs were allowed to do so. Kempin, 2/24/99pm,
at 42:22 - 43:9.
190. In light of this real-world conduct, there is no reason to conclude that
Microsoft's OEM restrictions were intended to, or did, preserve a consistent user
experience.
- "The fact that Microsoft has granted exceptions to these restrictions to
certain OEMs suggests that the concern for quality, speed, and
consistency is not Microsoft's primary motive for enforcing these
restrictions." Fisher Dir. ¶ 166.
- By allowing OEMs to vary the user's initial experience in significant
respects, "Microsoft makes plain that maintaining a consistent user
experience is a minor concern and easily gives way when OEMs create
value by differentiating their products: for example, by pre-installing a
particular set of applications. There is no reason why it should not
similarly give way when OEMs believe that end-users will find their
machines more attractive when they come with a non-Micorosoft browser,
rather than Internet Explorer, pre-installed." Warren-Boulton Dir. ¶ 161;
see also Warren-Boulton Dir. ¶¶ 171-78.
- Although Dean Schmalensee testified that Microsoft's consistency
concern is "plausible" (Schmalensee, 1/20/99am, at 22:13 - 23:6) he
investigated neither the actual decisionmaking process that led to the
"Windows Experience" nor Microsoft's pattern of granting and denying
exceptions. Schmalensee, 1/20/99pm, at 26:17 - 27:3, 35:3-14.
(2) Microsoft's purported concern with protecting
product quality and goodwill cannot explain its
restrictions
191. Microsoft's argument that the challenged OEM restrictions were designed
to preserve the quality of Windows and protect Microsoft's goodwill (Kempin Dir. ¶ 32,
36-38) cannot withstand scrutiny.
191.1. First, the restrictions are not necessary to induce OEMs to
preserve Windows' quality.
- OEMs, which pay customer support costs and operate in a
competitive market (Norris, 6/7/99pm, at 59:16-24) have no
incentive to supply alternative OEM shells, remove icons, or to
modify the boot-up sequence if doing so makes their PCs less
attractive to users. Warren-Boulton Dir. ¶¶ 83-88, 160, 181.
- Although Dean Schmalensee vaguely alluded to literature on
"channel conflict" (Schmalensee 1/20/99am, at 32:9-12;
Schmalensee, 6/21/99am, at 38:9 - 40:10) the evidence in fact
shows the absence of "the kind of conflict of interest" between
Microsoft and OEMs regarding the quality of Windows that might
lend credibility to Microsoft's testimony. Warren-Boulton,
11/30/98pm, at 57:3 - 58:8; Norris, 6/7/99pm, at 69:3-20 (testifying
that IBM "would be impacted" were it to offer a "confusing or poor"
user experience).
191.2. Second, that Microsoft's quality control justification is pretextual is
also shown by the fact that Microsoft permits OEMs to take any number of actions that
could jeopardize product quality.
- Microsoft permits OEMs to preinstall the third-party software of their
choice (including lengthy tutorials, ISP sign-up and registration
mechanisms) and other features in the start-up sequence, as well
as user-activated features of all sorts (such as alternative OEM
shells). Kempin Dir. ¶¶ 17-22, 46.
- Many of these features require significant technical sophistication
on the part of OEMs to implement, and the operation of any of
them could create technical problems for which the user could hold
Microsoft responsible. Soyring Dir. ¶¶ 20-22.
191.3. Third, other OS vendors -- which lack Microsoft's monopoly
power, and thus its ability and incentive to use licensing restrictions for strategic
purposes, and have even greater incentive to meet consumer demand -- do not
impose the type of restrictions Microsoft mandates.
- IBM permits licensees of its OS/2 operating system to customize its
screen displays extensively. Soyring Dir. ¶¶ 31-33; see also
Soyring, 11/18/98pm, at 80:19 - 81:19 (noting that IBM imposes
some restrictions, but "gives" OEMs "a great deal of freedom in
choosing which functions are either installed or used and what the
appearance of the screen would be"). In contrast to a Windows
licensee, an OEM licensing OS/2 "could override the entire desktop
in favor of a customized desktop or could set an application to start
automatically when the machine turns on." Soyring Dir. ¶ 31. IBM
has experienced "no erosion of . . . goodwill" or other ill-effects from
permitting OEMs such flexibility. Soyring Dir. ¶ 32.
- Apple similarly permits value added resellers to remove
applications and reconfigure the Apple desktop. Tevanian Dir. ¶
26; see also Limp Dep., (played 12/16/98am), at 32:24 - 33:21.
- Smaller operating system vendors often ignore restrictions on
OEMs set forth in their contracts because they recognize that it is
"in the[ir] best interests . . . to give [the OEMs] as much freedom
as possible." Warren-Boulton, 11/30/98pm, at 58:4 - 59:1.
191.4. Fourth, Mr. Kempin's testimony about preserving product quality
was contradicted by other, more reliable evidence, as well.
- His own videotape shows several perfectly well functioning
alternative OEM shells. DX 2163.
- Evidence showing that Microsoft imposed the Windows Experience
not because (as Kempin testified) it discovered that OEM shells
were poor (some were, but many were not) but rather because it
discovered, in Bill Gates' words, that OEMs were bundling "non-Microsoft browsers and coming up with offerings together with
Internet Service providers that get displayed on their machines in a
FAR more prominent way than MSN or our Internet Browser." GX
295. See supra Part V.C.1.a; ¶ 177.2.
- Evidence that Microsoft does not prohibit OEMs from taking
numerous other actions, including preinstalling inferior software,
that could reduce product quality. Soyring Dir. ¶¶ 20-22.
191.5. Fifth, Mr. Kempin's testimony itself was evasive, inconsistent, and
not believable.
- Although Kempin testified that the Windows Experience was
intended to prevent OEMs from "tampering" with Windows
(Kempin, 2/24/99pm, at 60:10 - 61:21; Kempin, 2/25/99am, at 25:4-9), he later admitted that he meant by these terms doing anything
Microsoft had not approved. Kempin, 2/25/99am, at 57:21 - 61:20;
Kempin, 2/24/99pm, at 67:18 - 68:3.
- Kempin testified that Windows would not function if Internet
Explorer were removed but, when pressed, conceded that he
lacked any pertinent technical knowledge. Compare Kempin Dir.
¶¶ 4-6, 66-68 with Kempin, 2/24/99pm, at 90:3 - 93:18 (Kempin
does not know how add/remove works); Kempin, 2/25/99am, at
66:21 - 68:25 (Kempin does not know how technically integrated
Internet Explorer 1 and 2 were with Windows); Kempin, 2/25/99pm,
at 5:2 - 10:2 (Kempin does not known the degree of technical
integration of Internet Explorer technologies). See also Kempin,
2/25/99pm, at 69:16 - 70:18 (Kempin claims that Internet Explorer
has to launch in certain instances in Windows 98 for technical
reasons, but Kempin cannot recall those reasons).
- Kempin testified that Microsoft's Windows 95 licenses always
prohibited OEMs' configuring programs to run automatically at the
end of the start-up sequence, including booting directly into
alternative OEM shells (and, therefore, that this restriction was not
first imposed by the Windows Experience in 1996-97). Kempin,
2/25/99am, at 24:9 - 26:24. But the very document he cited for
that proposition says precisely the opposite. DX 2395; see also
Kempin, 2/25/99pm, at 37:13 -38:13 (conceding Microsoft decided
"'not to allow OEM shells to interrupt the Windows 95 and NT boot
cycels'" in the spring of 1996 (quoting GX 1883)). Nor could
Kempin credibly explain why, if Microsoft never granted OEMs that
right, it used MDA provisions to induce OEMs to keep the standard
Windows desktop instead of simply enforcing its licenses. Kempin,
2/25/99am, at 24:3 - 27:22.
- Kempin denied that Gates, in writing that "[a]pparently a lot of
OEMs are bundling non-Microsoft browsers and coming up with
offerings together with Internet Service providers that get displayed
on their machines in a FAR more prominent way than MSN or our
Internet Browser"(GX 295), was "necessarily" "concerned with
Netscape" (Kempin, 2/25/99am, at 19:5-10), a reading that is
wholly incredible.
- Kempin's testimony concerning whether Microsoft's letters
authorizing OEMs to include an ISP sign-up process in the boot-up
sequence prohibited offering Netscape in the start-up sequence
was contradictory. See supra ¶ 187.2.
- Kempin testified that Microsoft's restrictions reflect terms "quite
common in the software industry (with the exception of UNIX
software vendors)," Kempin Dir. ¶ 25; but the evidence shows the
opposite. See supra ¶ 191.3.
192. Mr. Kempin testified that lifting the challenged restrictions would "degrade"
Windows and "limit[] end user choice" (Kempin Dir. ¶¶ 39, 13), but that testimony
cannot be credited because it is inconsistent with the evidence.
- Microsoft required IBM to scrap its introductory screens, including its
Welcome Center feature, despite the fact that Microsoft thought it
"enhanced the user experience" and never raised any concerns regrading
its quality. Norris, 6/7/99pm, at 62:25 - 63:25. What degraded Windows,
therefore, was not OEMs' conduct, but Microsoft's own restrictions.
- The evidence shows that permitting OEMs to remove the Internet Explorer
icon would increase choice by reducing the costs to OEMs of preinstalling
rival browsers that their customers might demand, see supra ¶¶ 179-180,
and would thereby increase the quality of Windows to those customers.
Kies Dep., 1/13/99, at 7:16-20 (explaining that the ability to remove the
Internet Explorer icon pursuant the Stipulation entered by this Court on
January 18, 1998, enhanced the value of NEC's notebook line of
computers to corporate customers, who generally "prefer[] to receive only
the base OS and drivers and not have any of the other third-party
applications pre-installed").
- Microsoft itself recognized this
- redacted - GX 227 (sealed). Dell executive Joseph Kanicki testified that Dell
wanted this freedom so that it could offer higher value to end users by
permitting customers standardizing on a non-Microsoft browser to remove
the Internet Explorer icon. Kanicki Dep., 1/13/99, at 339:13 - 342:1.
193. Microsoft witnesses' related argument -- that consumers would be
disappointed not to find on their PCs features Microsoft promotes (Kempin Dir. ¶ 42) --
is also inconsistent with the evidence.
- Microsoft permits OEMs to ship PCs with the Active Desktop and its
associated "channels" inactive (and without any warning to customers).
GX 231.
- Consumer expectations could in any event be met by the significantly less
restrictive alternative of a labeling requirement, pursuant to which OEMs
could be required to tell their customers when particular PCs include non-standard features. Warren-Boulton, 12/1/98, at 39:6 - 40:10; see also
Warren-Boulton Dir. ¶¶ 177, 181.
(3) Microsoft's restrictions are unrelated to its
purported concern of preventing fragmentation of
the Windows platform
194. Microsoft's witnesses said that lifting the restrictions challenged by the
plaintiffs would result in different "flavors" of Windows, thereby "destroy[ing]" "one of the
central reasons for the appeal of Windows among customers and developers -- that it
provides a stable, coherent platform for software development" (Kempin Dir. ¶ 30; see
also Warren-Boulton Dir. ¶ 31; Maritz Dir. ¶¶ 167-70); but none of these witnesses
explained how any of the OEM desktop or start-up restrictions has anything to do with
protecting the stability or coherence of the Windows platform.
194.1. None of the OEM restrictions was needed to protect the platform
because the stability of the platform depends only on the stability of the APIs and the
OEM restrictions were aimed at conduct that did not involve removal or alteration of
APIs.
- Neither removing the Internet Explorer icon nor permitting
programs to run at the end of the start-up sequence nor modifying
the start-up sequence need involve removing or altering APIs.
Felten Dir. App. B, at 14 (removing the Internet Explorer icon
requires simply changing a few registry entries).
- There is, moreover, obviously less restrictive alternative that
"restricts or prevents only those modifications . . . that impair the
ability of ISVs to access the APIs provided by the Windows."
Warren-Boulton Dir. ¶ 180.
194.2. Even if OEMs were contractually free to remove or alter APIs,
Microsoft's concern about platform fragmentation would still provide no justification for
the OEM restrictions.
- In the first place, OEMs have no incentives to take actions that may
make Windows less valuable, such as removing APIs that
customers want. Warren-Boulton Dir. ¶¶ 179, 181.
- Moreover, Microsoft itself routinely undermines the stability of its
platform with its frequent Internet Explorer updates. Fisher,
6/3/99am, at 21:24 - 22:14.
- In any event, most ISVs find it necessary to redistribute the Internet
Explorer APIs themselves (thus ensuring a stable platform)
because of existing nonuniformity in the Windows installed base
and Microsoft's frequent Internet Explorer updates. "Microsoft's
API's are not, in fact, stable. They change. And ISV's have to
keep embedding pieces of the appropriate API's into their own
software in shipping it out." Fisher, 6/3/99am, at 22:8-14; supra
Part V.B.3.d(2); ¶ 164.
2. Microsoft used its monopoly power to force OEMs into taking
actions to hinder products or industry developments that
threatened its operating system monopoly
195. Microsoft used its monopoly power both to coerce OEMs into taking actions
that furthered its exclusionary strategy and to punish OEMs that refused.
a. Microsoft used threats and bribes to induce OEMs to
help entrench its operating system monopoly
196. Microsoft used its monopoly power, both through threats and bribes to
OEMs, to entrench its operating system monopoly.
(1) Microsoft used its monopoly power to secure
Compaq's assistance in its exclusionary strategy
197. Compaq, the largest OEM in the United States, has been aware for years
that Microsoft has the power to damage its personal computer business because
Compaq lacks any practical alternative to Windows.
- John Rose's testimony emphasizes the importance of Microsoft's
operating system to Compaq's business. See Rose Dir. ¶ 17 (Compaq
loads only Windows on consumer PCs because it believes that's what
consumers want); Rose, 2/17/99pm, at 41:4-11 (despite increase in price
of Microsoft operating systems in 1998, Compaq has not evaluated any
other operating systems for preinstallation on desktop PCs); Rose,
2/17/99pm, at 64:19-22 (cancellation of Windows license would be "of
great concern" to Compaq), see also Dunn Dep., 10/23/98, at 37:24 -39:15 (DX 2566) (Celeste Dunn, a former Compaq executive, explained
there were no alternatives to Windows during her tenure at Compaq).
- A Compaq presentation entitled "Microsoft Meeting Preparation --
Portable and Software Marketing PC Division,"
- redacted -
"Judgment: How retaliatory would they get?: Pricing advantage --
Revenue from updates -- Access to early SDKs -- Field sales
activities (Microsoft has ~900 field sales people) -- Support and
training -- Inclusion in advertising -- Tone toward Compaq in press
and with customers -- Selection and elevation of other OEMs as
leaders -- Make integration relations even more strained than they
are today -- Access to source code, modification ownership --
Microsoft directional information and plans -- Customers." GX 433,
at slide 8 (sealed).
- - redacted -
- See also supra Part II.A; ¶ 15.1.2 (detailing evidence that Compaq has no
commercially viable alternative to Windows).
198. Compaq recognizes that it has a - redacted - with Microsoft.
- redacted -
- - redacted -
- - redacted - Rose conceded at trial that Microsoft's relationship with Compaq is
different from its relationship with other OEMs. Rose, 2/18/99am, at 10:1-18.
-
- redacted -
199. Microsoft has frequently granted Compaq more favorable terms than other
OEMs as a result of Compaq's acquiescence in Microsoft's exclusionary strategies and
because of the
- redacted -
199.1. In 1992, Microsoft granted Compaq a five-year "Frontline
Partnership" license agreement, covering the years 1993 through 1998,
- As Professor Fisher summarized:
- redacted -
Fisher, 1/11/99am, at 20:12-18 (sealed session).
-
- redacted -
GX 451 (sealed); GX 449, at MSV 0002626-29 (1992 Frontline
Partnership agreement) (sealed); Rose, 2/18/99pm, at 92:24 -
93:10 (sealed session).
-
- redacted -
- Compaq executive Bob Jackson noted in a December 1994 e-mail
to other Compaq executives: "A major piece of value in the agrmt
[sic] is they cannot raise the price on us, which I believe they can
do in all other agreements." GX 135; Rose, 2/19/99am, at 46:18-22 (from 1992 through 1998, Compaq's Windows royalty "didn't go
up at all")
199.2. Microsoft granted Compaq the significant competitive advantage
of being able to verify that its prices were lower than those charged its competitors, and
Compaq received confidential information about other OEMs' prices.
- An internal Compaq memorandum to its CEO, Eckhard Pfeiffer, in
January 1995 states: "Jan Claesson is Microsoft's new OEM Group
Manager for the Compaq account and reports to Joachim Kempin .
. . ." GX 230, at 5812. "Jan provided very confidential information
about Windows 95 royalties in regards to other OEMs. The
bottom-line was that Compaq still has a significant price
advantage." GX 230, at 5816 (sealed).
- Kempin conceded that GX 230 shows that Claesson provided
Compaq confidential information. Kempin testified that if he would
have "known this at that point in time, I think Mr. Claesson would
have gotten fired." Kempin, 2/25/99pm, at 107:17 - 108:19.
200. When Compaq acted contrary to Microsoft's strategy to gain browser
usage share in 1995, Microsoft wielded its monopoly power to coerce Compaq into both
distributing and highlighting Microsoft's products and excluding Microsoft's rivals.
200.1. As early as late 1994, Compaq executives recognized that it was
in Compaq's interest to be able to choose among competing online services and other
Internet software on the merits.
- In a December 1994, internal e-mails discussing Compaq's
negotiations for Windows 95, Compaq executive Vaughn Rhodes
wrote: "I strongly recommend that we take he position of
negotiating distinctly for the MS network . . . . Compaq should
have each online service (America Online, MS Network, Prodigy,
etc.) prepare a proposal which indicates what they can offer
Compaq . . . . Compaq can then evaluate each proposal on its own
merits . . . ." GX 135.
200.2. In connection with the release of its new consumer line of PCs in
late 1995/early 1996, Compaq removed the Internet Explorer and MSN icons from the
Windows 95 desktop screen in order to feature Netscape and certain AOL software.
- As Rose acknowledged: "I understand that, in early 1996, Compaq
did remove, on some consumer products, the Internet Explorer icon
(as opposed to Internet Explorer software) from the Windows 95
default desktop on its Presario line of personal computers." Rose
Dir. ¶ 25.
- Rose also acknowledged that Compaq had a strategy to feature
Netscape along with AOL. Rose, 2/19/99am, at 64:14-23.
- Celeste Dunn, Compaq's former Vice-President of the Consumer
Software Business Unit with responsibility for software decisions on
the Presario product line, testified that she believed at the time that
featuring Netscape, which had "brand name recognition that the
consumer could equate value to," would benefit Compaq, but that
Compaq "probably would not want to feature the Netscape
browser" if Compaq had to give favored treatment to Internet
Explorer and MSN. Dunn Dep., 10/23/98, at 92:1 - 94:5 (DX 2566).
- Dunn had contemporaneously emphasized the importance of
highlighting Netscape Navigator in an internal May 1996 e-mail
raising concerns about agreeing with Microsoft to display the
Internet Explorer and MSN icons and make Internet Explorer the
default browser: "In regard to browsers, our goal is to feature the
brand leader Netscape . . . ." GX 299.
- Stephen Decker, Compaq's Director of Software Procurement, also
testified that Compaq's removal of the Internet Explorer icon
reflected Commpaq's commitment to Netscape. He testified that
Compaq wanted to remove the Internet Explorer icon because "at
the time, we had a relationship with Netscape and we had been
shipping their product for a while. And, therefore, Netscape was
actually the browser partner and we wanted to give that position on
the Compaq Presario desktop." Decker Dep., 2/18/99am, at 47:19-25.
- Microsoft's account managers responsible for dealing with Compaq
also testified that Compaq removed the icons in order to highlight
AOL and Netscape software. Williams Dep., 1/13/99, at 792:4-21
(Compaq executives told Microsoft that they wanted to remove the
MSN and Internet Explorer icons because it would make it easier to
promote their other partners.); Hardwick Dep., 8/11/98, at 38:4 -
40:7 (DX 2570) (Microsoft OEM representative understood from
Compaq employees that Compaq removed the Internet Explorer
icon so that Compaq could earn a bounty from Netscape for being
the only browser).
200.3. In May 1996, Microsoft responded to Compaq's removal of the
Internet Explorer and MSN icons from Compaq's PCs by threatening to terminate
Compaq's OEM license unless Compaq restored the Internet Explorer and MSN icons
to the desktop.
- On May 31, 1996, Peter Miller, an in-house attorney for Microsoft,
sent a letter to David Cabello, Compaq's then General Counsel,
notifying Compaq of Microsoft's intent to terminate Compaq's
operating system license: "It has come to Microsoft's attention that
Microsoft Windows 95 is being distributed by Compaq with Compaq
Presario systems in a form which has been modified by Compaq.
The most notable example is that Compaq has modified Windows
95 by removing the Microsoft Network icon from the Windows 95
'desktop' screen. . . . The Agreement does not grant Compaq any
rights to modify Product software . . . or to delete items from
Product software. . . . As per the terms of Section 10 'Default and
Termination' of the Agreement, this letter is written notice of
Microsoft's termination of the Agreement. As per Section 10,
termination will be effective thirty (30) days after this notice unless
Compaq cures the above violation within the thirty (30) day period."
GX 649.
- On June 6, 1996, Microsoft's Don Hardwick followed up on the
notice of intent to terminate Compaq's Windows 95 license
agreement by sending Compaq a letter offering to withdraw the
notice if Compaq restored the MSN and Internet Explorer icons:
"Microsoft is requesting that Compaq replace the Microsoft Network
and Internet Explorer icons on the Wndows 95 desktop on all
Compaq Presario machines. Specifically we are asking that these
icons be put back on the Windows 95 desktop so they look and
function exactly the same as how they were originally provided by
Microsoft and/or Authorized Replicators. This means the icons
should not be just Windows 95 shortcuts, since the functionality is
different. In addition, the Microsoft Network and Internet Explorer
icons and Internet Setup Wizard icon should also be put back into
their original locations and functionality under the 'Start' button on
Windows 95. If you are willing to give Microsoft a clear written
assurance that the above will be implemented on all Compaq
Presario machines within sixty (60) days of the date of this letter,
Microsoft will withdraw its Notice of Intent to Terminate letter
addressed to David Cabello and dated May 30, 1996 once such
written assurance is received by Microsoft." GX 650.
200.4. Contrary to the suggestion of Microsoft's witnesses that it was
concerned simply about upholding the integrity of the Windows 95 product and its
license agreements (Rose Dir. ¶ 30; Hardwick Dep., 1/13/99, at 521:10 - 522:1),
Microsoft's actual purpose was to put a stop to Compaq's promotion of rival products
and services, including Netscape Navigator, instead of Internet Explorer and MSN.
200.4.1. When Compaq informed Microsoft in advance of its
intention to remove the Internet Explorer and MSN icons, Microsoft executives
acknowledged that removal of the icons did not affect Windows 95's functioning.
- Celeste Dunn of Compaq testified that Microsoft tested
Compaq's Windows configuration, and had not detected any
technical problems. Dunn Dep., 10/23/98, at 187:12-25 (DX
2566).
- Microsoft's Donald Hardwick testified that he never heard
from anyone that Compaq's removal of the Internet Explorer
icon had any adverse effects on Windows' functioning.
Hardwick Dep., 1/13/99, at 522:2 - 523:4.
- Bengt Akerlind, who was responsible for overseeing
Microsoft's relationship with Compaq, also testified he was
not aware at the time that Compaq's removal of the MSN
and Internet Explorer icons would cause any problems for
the rest of Windows. Akerlind Dep., 8/26/98, at 116:4-7 (DX
2603S).
200.4.2. Microsoft's true concern -- that Compaq was assisting
Netscape -- is evidenced by its strong negative reaction to indications by Compaq that it
might support Netscape in other ways. For example, in early 1996, when Compaq
announced its intention to work with Netscape for its internal Internet needs and on
Internet server initiatives, Microsoft retaliated by initiating cooperative activities with
Compaq's competitors and by insisting that Compaq support Microsoft's Internet
initiatives throughout its business.
- In May 1996, an internal Compaq e-mail by Lori Day
reported, with regard to the February announcement of
Compaq's partnership with Netscape, that "Microsoft was
upset with our announcement and our internal use of
Netscape and initiated a number of activities with DEC and
HP, reducing their emphasis on the Compaq partnership."
GX 298.
- Day went on in the e-mail to report that Compaq was
negotiating an extension of the Frontline Partnership to the
Internet/Intranet with Microsoft in order to get "realigned"
with them. Microsoft insisted, against Compaq's wishes,
that the Internet/Intranet partnership should extend
throughout Compaq's business lines. Microsoft wanted:
- "Compaq to ship new versions of Internet Explorer as
the default browser on all desktop and server
platforms within 8 weeks from release." GX 298.
- "Compaq to display MSN icon on Desktop screen on
all Windows 95 PCs. Compaq and Microsoft to
explore other areas of joint activities such as:
Promote Internet Explorer activites on Compaq
Desktop PCs." GX 298.
- Another internal Compaq e-mail,
- redacted -
- Dunn testified that after receiving the notice of Microsoft's
intent to terminate Compaq's operating system license, other
senior Compaq executives expressed concern that Microsoft
"was not willing to go forth with the finalization of an NT
license agreement without having a resolution -- a complete
resolution on our -- on our icon issues on the PC platform."
Dunn Dep., 10/23/98, at 107:5 - 108:23 (DX 2566).
200.5. Because Compaq recognized that termination of its Microsoft
license would mean the end of its PC business, it complied with Microsoft's demands
and restored the MSN and Internet Explorer icons to the Windows desktop screen on
its Presario PCs.
- On June 21, 1996, Compaq gave in to Microsoft's demands. Dunn
sent a letter to Hardwick, indicating that Compaq has "made the
changes you requested to the Windows 95 desktop of the current
release of the Compaq Presario systems. We have replaced the
Microsoft Network and Internet Explorer icons on the Windows 95
desktop as executable icons so they look and function exactly the
same as how we originally received them from Microsoft and have
placed Microsoft Network, Internet Explorer icons and Internet
Setup Wizard icons in their original locations under the Start button
on the Windows 95 desktop." The letter also pointed out that icons
for AOL and for Netscape were on the Windows 95 desktop for
Presario systems. GX 645
- On June 25, 1996, Microsoft sent Compaq a letter withdrawing the
Notice of Intent to Terminate Compaq's Windows operating system
license agreement based on Compaq's representations. GX 301.
200.6. Microsoft's coercion of Compaq to promote Microsoft's Internet
software exclusively and to restore the Internet Explorer icon raised the cost to Compaq
of featuring Netscape and caused Compaq to remove Netscape Navigator from its
Presario PCs.
- See supra Part V.B.4.a.1-3; ¶ 167; Part V.C.1.b; ¶ 179.
200.7. John Rose's testimony concerning the events surrounding the
removal of the MSN and Internet Explorer icons is inconsistent with both the first-hand
accounts of other Compaq executives and the contemporaneous evidence.
200.7.1. Mr. Rose's assertion that Compaq had agreed in August
1995 specifically not to remove icons from Windows 95 is wrong. It is inconsistent with
contemporaneous Compaq documents and the testimony of Compaq executives who
were personally involved.
- Compaq and Microsoft were still discussing whether
Compaq would agree to display the icons in May 1996. See
supra ¶ 200.3; GX 298 (Lori Day, of Compaq, e-mail dated
5/8/96, stating "we have not agreed" on the proposal that
"Compaq to display MSN icon on the desktop on all
Windows 95 PCs"); Rose, 2/17/99pm, at 79:15 - 80:10
(Rose was not aware if other top executives at Compaq
raised the issue of conflict between GX 298 and any alleged
earlier agreement with Microsoft, and did not raise any such
issue himself).
- Rose could not explain how his allegation of an August 1995
commitment by Compaq not to remove icons from the
Windows desktop was consistent with the behavior of senior
Compaq executives at the time.
- Rose acknowledged that Compaq signed an
agreement with AOL in late August 1995 under which
Compaq agreed to "position AOL services above all
other Online Services within the user interface of its
products."
- Rose further acknowledged that the Compaq/AOL
agreement would have been in conflict with what he
alleges was an early August 1995 oral agreement by
Compaq not to remove the MSN and Internet
Explorer icons.
- And yet, Rose acknowledged, Steve Flannigan, who
was Compaq's executive partner in charge of the
Microsoft relationship, did not indicate at the late-August time of the Compaq/AOL deal that the
agreement would violate Compaq policy or any pre-existing agreement.
Rose, 2/17/99pm, at 67:3-19; DX 2261, at 1.1.1
(Compaq/AOL agreement, 8/23/95).
- Rose admitted that he was not involved in any decisions about the
removal of the Windows icons from Presario desktops. Rose
testified: "I was not personally involved there, so I couldn't tell you
what happened" (Rose, 2/19/99am, at 62:19 - 64:12), but that
Celeste Dunn was involved. Rose conceded that "ultimately
Celeste was responsible for the software" (Rose, 2/18/99am, at
37:7-11), and that he has "no reason to believe" Dunn's testimony
is inaccurate on a range of points that are inconsistent with his
allegations:
- Mike Heil, whom Rose acknowledges made final decisions
on Presario, supported Compaq's decision to remove the
icons;
- Don Hardwick and others at Microsoft were informed of the
plan to remove the icons in advance;
- No one from Microsoft indicated that removing the icons
would violate a summer 1995 agreement.
Rose, 2/18/99am, at 37:16 - 42:1.
200.7.2. Mr. Rose's testimony concerning Compaq's removal of
the Internet Explorer and MSN icons in 1996 is unreliable in other ways as well.
200.7.2.1. Mr. Rose was not personally involved in the
removal of the Internet Explorer and MSN icons.
- Rose did not learn of the removal of the icons and the
issue of a possible violation of an agreement with
Microsoft until May 1996 -- months after their
removal, and nine months after the alleged
agreement with Microsoft and the known agreement
with AOL that would have violated it. Rose,
2/17/99pm, at 64:8-18.
- See also supra ¶ 200.7.
200.7.2.2. Mr. Rose repeatedly altered his testimony about
events surrounding the removal and restoration of the Internet Explorer and MSN icons,
and deliberately played word games in order to advance Microsoft's agenda.
- Mr. Rose either had not read his direct testimony or
deliberately tried to obfuscate the proceedings by
asserting Compaq did not "remove" the Internet
Explorer and MSN icons.
- Compare Rose Dir. ¶ 25 (Rose expressly
stated that Compaq removed the Internet
Explorer icon from Windows: "I understand
that, in early 1996, Compaq did remove, on
some consumer products, the Internet Explorer
icon (as opposed to the Internet Explorer
software) from the Windows 95 default desktop
on its Presario line of personal computers").
- With Rose, 2/18/99am, at 34:25 - 35:9 (Rose
asserted that in fact Compaq did not remove
the icons: "Compaq never removed the
Internet Explorer or MSN icon from the
desktop. What we did was we failed to comply
with the OPK rules that we had agreed to. So
we never put the icon for the Internet Explorer
or MSN on--displayed on the Presario screen
as an icon. So we didn't remove it. We just
never put it up there."); Rose, 2/18/99am, at
51:7-15 ("A: It may be semantics, but that's -
.Q: Your semantics. A: Yes, my semantics.").
- Mr. Rose changed his testimony about the
circumstances under which Compaq stopped
shipping Netscape Navigator.
- Compare Rose, 2/18/99am, at 65:12-25 ("Q: In
1996, did Compaq stop shipping Netscape
Navigator with its PCs? A: I don't know
specifically because there were some
compatibility issues in that timeframe with
Netscape Navigator in our overall system, so
we went through an iterative process with
Netscape on getting the compatibility issues
resolved. So, at that point in time, the
Netscape Navigator may have been off, then
back on. But the issue was compatibility. Q:
When I asked you that question before, you
said the issue was AOL. Do you remember
that? A: You asked me a different question.
The issue overall was AOL -- AOL's desire that
we only feature AOL and GNN.").
- With Rose, 2/18/99am, at 62:20 - 64:6 (the
earlier testimony, testifying that his
understanding at that time was that Compaq
would stop shipping Navigator on portable
products not because Internet Explorer already
filled the product category but because "at that
time the issue was AOL, and AOL did not want
us to feature any browser in there. . . .").
200.7.2.3. Mr. Rose's testimony is factually incorrect in
important respects.
- Rose's testimony that Compaq did not remove
Internet Explorer from the Windows 95 Start menu is
contradicted by contemporaneous documents. Rose
Dir. ¶ 25. Compare GX 650 and DX 2265 (June 6,
1996 Hardwick letter to Dunn addressing need for
Compaq to return Internet Explorer to its "original
locations and functionality under the 'Start' button on
Windows 95"); DX 2266 (June 26, 1996 Dunn letter to
Microsoft on same topic).
- Rose's testimony that Compaq merely removed, or
did not install, the Internet Explorer and MSN icons
and "replaced it with AOL and featured AOL, which
had the Netscape Navigator in it" is incorrect. Rose,
2/17/99pm, at 69:4-10; Rose, 2/18/99am, at 48:24 -
49:10. Compaq in fact separately preinstalled
Netscape Navigator, independently of AOL, as Rose
himself recognized in his written testimony. Rose Dir.
¶ 26 (testifying Compaq featured Navigator on some
models before Compaq's AOL agreement).
- Rose's repeated testimony that GNN was AOL's
"proprietary browser" is incorrect. Rose, 2/18/99am,
at 63:20 - 64:6; Rose, 2/19/99am, at 35:1-7. GNN
was an ISP offering distinct from AOL's flagship
"AOL" service, which included (but was not limited to)
a browser. As Compaq's agreement with AOL made
clear, Compaq's obligation was to feature the AOL
service and GNN, defined in the agreement as a
"direct Internet service provider." DX 2261.
200.7.2.4. Mr. Rose's testimony that Compaq removed the
Internet Explorer and MSN icons solely because of its contract with AOL) is inconsistent
with contemporaneous documents prepared by Compaq executives -- Celeste Dunn
and Stephen Decker -- who were involved in the Consumer Division's software
decisionmaking. Rose Dir. ¶ 26; Rose, 2/18/99am, at 48:21 ("The issue here was AOL
and the AOL browser").
- According to Decker, Compaq removed the Internet
Explorer icon from the Compaq desktop because, "at
the time, we had a relationship with Netscape and we
had been shipping their product for a while. And,
therefore, Netscape was actually the browser partner
and we wanted to give [them] that position on the
Compaq Presario desktop." Decker Dep. (read
2/18/99am), at 47:21-25.
201. In response to Microsoft's threat to terminate its license agreement, and in
recognition of the fact that antagonizing Microsoft through dealings with Netscape or
otherwise would jeopardize the favorable terms it enjoyed, Compaq reverted to a policy
of not taking competitive actions that would invite Microsoft's retaliation.
201.1. Compaq agreed to distribute and promote Internet Explorer as its
default browser on all desktop and server PCs in order to heal the conflict with Microsoft
created by Compaq's internal use of Netscape products and to avoid further
antagonizing Microsoft.
- In response to the email in which Dunn explained Microsoft was
upset with Compaq's Netscape partnership and internal usage, a
senior Compaq executive wrote to others within the company:
"Please ensure that we can comply. Please work with Lori, Gene
and Steve Flannigan as soon as possible to ensure success" in
negotiating a palliative extension of the Compaq/Microsoft Frontline
Partnership. GX 298.
- Compaq extended its Frontline Partnership with Microsoft to the
Internet/Intranet in a May 13th, 1996 agreement. Microsoft's cover
letter made clear that the agreement included Compaq's
"company-wide commitment to display the MSN icon on the
desktop screen of all Windows 95 and Windows NT Workstation
PCs and to ship Internet Explorer as the default Web browser on all
desktop and server systems." The agreement also required
Microsoft to ship new versions of Internet Explorer as the default
browser on new Compaq products and to focus the "majority of
Compaq's key Internet/Intranet announcements and marketing
activities . . . on Microsoft's technologies and strategy." GX 1133.
- At the time senior executives at Compaq signed the agreement,
Dunn was concerned that its terms required Compaq's Consumer
Division to "significantly alter the current Predator product, revise its
business model, and jeopardize two profitable, revenue generating
contracts" with AOL and Netscape. She also warned that the
agreement conflicted with "our goal to feature the brand leader
Netscape" and that it could raise Compaq's support costs, since
"Both Netscape and AOL have agreed to provide end user support
for their products and the On-line/Internet services. Traditionally
Compaq has provided end user support for MS products and
nothing in the MOU states differently. The cost of such support is
estimated to be quite high." GX 299.
- Compaq also entered into other agreements relating to distribution
of Internet Explorer with Microsoft. GX 1130 (Amendment 1 to the
September 10, 1996 agreement requiring Compaq to offer Internet
Explorer as the preferred browser for listed Compaq Internet
Products and to use two or more advanced features of Internet
Explorer HTML extensions in Compaq's home page for each
product); GX 1137 (July 1, 1996 agreement requiring Compaq to
offer Internet Explorer as preferred browser for Support Software
CD for desktop products); GX 1155 (sealed).
- Compaq executed an agreement under which
Compaq promoted Internet Explorer exclusively. Decker Dep.,
2/18/99pm, at 15:3-12; GX 1155 (sealed). Rose acknowledged he
had no reason to doubt Decker's testimony. Rose, 2/18/99pm, at
14:15 - 15:16.
201.2. Compaq complied with certain provisions of the "Windows
Experience" screen and startup sequence restrictions
- - redacted -
GX 1023 (sealed). Despite that concern,
Rose testified, Compaq executed an agreement in June 1996 to
amend the Frontline Partnership to commit Compaq to "not replace
or modify the OPK [OEM Preinstallation Kit] install process in any
way." Rose Dir. ¶ 29; DX 2264.
- See also supra V.C.2.a.(1); ¶ 197.
201.3. In a further effort not to antagonize Microsoft and not to risk
Microsoft's retaliation, Compaq decided not to preinstall Apple's QuickTime multimedia
software.
- Stephen Decker of Compaq told Phil Schiller of Apple that Compaq
was reluctant to preinstall QuickTime for fear of upsetting Microsoft.
Schiller Dep., 1/13/99, at 238:18 - 239:10.
- As a March 1998 e-mail from David Obelcz, who worked under
Decker at Compaq, to Phil Schiller of Apple explained: "I wanted to
thank you for your visit to Compaq and all the effort you have put in
for QuickTime 3.0. I understand the path Compaq has taken and I
know it was not the idea direction you had hoped for from the
Consumer Division. . . . I think Apple has a lot to offer and I have
been evangelizing QuickTime 3.0 and QuickTime 3.0 Pro as an
excellent alternative to DirectShow for DVD title development. The
folks in Redmond beat me up for it but also quietly tell me they are
impressed. You have a great product Phil and I am sure we can
find a home for it. . . ." GX 269.
201.4. Having experienced Microsoft's use of its monopoly power to
secure the prominent positioning of Internet Explorer and MSN on the desktop of all
Compaq PCs, Compaq recognizes
- An internal Compaq presentation on the "Microsoft Relationship" in
January 1998 described
- redacted -
- An August 12, 1997 review of the "Microsoft Partnership"
- redacted -
201.5. Compaq also supported Microsoft in this litigation, including
by offering the testimony of John Rose, which was largely speculative and inconsistent
with other evidence.
- Gates thanked Rose for his testimony, and Rose was not
forthcoming about this discussion. Rose, 2/18/99pm, at 26:1-
31:14,
- redacted -
201.5.1. Little of Mr. Rose's testimony is based on personal
knowledge about the issues it addresses.
- Rose acknowledged that he has not been involved in the
Compaq's PC business for over two years. He testified,
"The last time I had profit-and-loss accountability -- and I will
use that as a measurement of accountability -- for the PC
Products Group was June of 1996; and for the Consumer
Products Group, it was in September, actually August -- it
ended in August, the beginning of September -- of 1995."
Rose, 2/17/99pm, at 10:12 - 11:4. Accordingly, he could not
even "fathom a guess" about what Compaq's support costs
relating to Windows are. Rose, 2/18/99pm, at 44:25 - 45:24.
- Rose has not been involved in the negotiation of Compaq's
license agreements with Microsoft, in contrast to others from
Compaq who have testified about browser and operating
system issues. Rose, 2/18/99pm at 18:23 - 22:13
(acknowledging Decker and Flannigan, but not Rose
himself, negotiated Compaq's current license with
Microsoft).
- Rose was not involved in any decisions about the removal of
the MSN and Internet Explorer icons from the Windows
desktop, and indeed did not learn of the removal for some
time. Rose, 2/19/99am, at 62:19 - 64:12; Rose, 2/17/99pm,
64:8-18.
201.5.2. Mr. Rose's testimony about demand for Windows and
Internet Explorer, and the benefits of their linkage, is uncorroborated and inconsistent
with the documentary evidence.
- Rose's written testimony that "Windows 98 has features and
functions that our customers want and need" (Rose Dir. ¶
16) is inconsistent with internal Compaq documents.
- redacted -
- Rose conceded that he has no foundation for distinguishing
between operating system and other (application) software,
and has described both browsers and word processing
software as "features" of personal computers generally.
Rose Dep., 2/18/99pm, at 49:1 - 50:3.
- redacted -
- Rose admitted he was unaware that the retail version of
Windows 95 did not have a built-in browser. Rose,
2/18/99pm, at 53:4-7. Given this ignorance, he could not
meaningfully testify about why large businesses frequently
choose, according to a 1998 Compaq research study, to
replace the current version of Windows they buy with the
retail version without a browser. GX 1242 at 7. Rose
conceded that "about 80 percent of companies wipe or
reformat the hard drives of new desktops. . . .The operating
system (sic) reinstalled most often are OSR2 and the retail
version of Windows 95. Large businesses lean more toward
the retail version of Windows 95." Rose, 2/18/99pm, at 51:13
- 52:9 (referring to GX 1242). He went on to testify, "I'm not
sure why they would want the retail version of Windows 95."
Rose, 2/18/99pm, at 52:16-18.
202. In light of Compaq's renewed commitment to support Microsoft's Internet
strategy and other accommodations, Microsoft has continued to bestow privileges on
Compaq that were not granted to other OEMs.
202.1. Microsoft continued to
- redacted -
- See supra Part II.C.3; ¶ 38.
-
- redacted -
-
- redacted -
-
- redacted -
-
- redacted -
202.2.
-
- redacted -
202.3.
-
- redacted -
-
- redacted -
-
Rose acknowledged that
when Compaq signed the two MDAs, Compaq knew that the
second agreement immediately superseded the first agreement;
Rose agreed that, "when Government Exhibit 464 was signed, no
one thought at that time that these were going to be the real terms.
You knew you were executing at the very same time Exhibit 1438
that would supersede it." Rose, 2/18/99pm, 61:25 - 62:12 (referring
to Microsoft and Compaq terms);
- The special terms of the superseding MDA remain in force, at
Compaq's discretion, for the full two-year term of Compaq's
Windows license agreement, effective April 1, 1998. GX 1438
(sealed); GX 1190, at MS98 0008920 (sealed). The license itself
acknowledges that Compaq has earned a discount for the year
April 1998 through March 1999 and that, if Compaq exercises its
right to extend the contract for another year, the MDA discount will
also be effective for that year. GX 1190, at MS98 0008930
(sealed).
(2) Microsoft used MDA discounts to induce
other OEMs to take exclusionary actions
203. Microsoft's Market Development Agreements ("MDAs") provide discounts
off the Windows license fee to OEMs that undertake actions that benefit Microsoft in
various ways.
- Norris of IBM testified that the MDA is "a vehicle that Microsoft used in
order for us to perform activities that benefited them in many ways. It was
a vehicle that also gave royalty reductions that imposed costs on the P.C.
manufacturer in order to attain the royalty reductions." Norris, 6/7/99am,
at 17:16-24. Norris also testified that it was within Microsoft's sole
discretion whether or not an OEM was deemed to have "met" the MDA
milestones and thereby earned the reduction against its Windows royalty.
Norris, 6/7/99am, at 16:23 - 17:14.
-
- redacted -
(a) Microsoft offered discounts for making
Internet Explorer the default browser
203.1. Microsoft offered certain OEMs MDA discounts for making Internet
Explorer the default browser on the PCs those OEMs shipped to their customers.
- See infra Part V.G; ¶ 297.4.3.2.
-
- redacted -
- Microsoft offered IBM a discount for making Internet Explorer the
exclusive browser on IBM systems, but IBM refused it because it
believed that IBM must maintain browser neutrality. See infra ¶¶
205.1 - .3; Norris, 6/8/99am, at 30:2-14; GX 2203.
(b) Microsoft offered discounts for preserving
the Microsoft-dictated Windows interface
203.2. Microsoft offered OEMs discounts for preserving the standard,
Microsoft-dictated Windows desktop and boot-up sequence.
-
- redacted -
GX 1503, at MSV 0003035 (IBM) (sealed); GX 1506, at
MSV 0005932 (HP) (sealed); GX 1511, at MSV 0004213 (Packard
Bell) (sealed); GX 1498, at GW 019843 (Gateway) (sealed); GX
1509, at MSV 0006946 (Hitachi) (sealed); GX 1493, at MSV
0006225 (AST) (sealed).
(c) Microsoft offered discounts to OEMs that
designed PCs in accordance with the
Microsoft Hardware Design Guide ("HDG")
and subject to validation testing at
Microsoft's Windows Hardware Quality
Labs ("WHQL")
203.3.
-
- redacted -
-
- redacted -
-
- redacted -
-
- redacted -
203.3.1. Certification under the HDG is very competitively
significant for OEMs.
- Mr. Norris testified that customers check to see if Microsoft
has certified that its systems meet Microsoft's standards;
accordingly, he explained, IBM believes that it is
competitively important to meet the WHQL requirement and
get Microsoft's certification. Norris, 6/8/99am, at 9:4 - 10:4.
-
- redacted -
203.3.2. An OEM that wanted to offer a simplified or lower-cost PC
to its customers would risk losing millions of dollars in price reductions for all its PCs if
the lower-cost PC did not satisfy these MDA provisions.
- Garry Norris testified that if IBM "wanted to offer any
configuration of systems that were outside of the logo
requirements or that may have been outside the
requirements of the hardware design guide, then it placed at
risk the opportunity for us to gain the royalty of $2 in number
1 and $2 in number 2 for all of the systems that we were
shipping. And if you add that up. . . it's. . . Roughly $20
million in cost savings to IBM." Norris, 6/8/99am, at 10:5-17.
-
- redacted -
203.3.3. As a result, OEMs generally satisfy HDG requirements.
IBM, for example, complied with the HDG/WHQL milestones in its 1996 MDA because
those milestones resulted in approximately $20 million in discounts to IBM.
- IBM complied with the WHQL requirements, and with the
"next generation technology" requirement, in its MDA-96
because "$20 million" -- the amount of the discount IBM
could have lost if it did not comply -- "was a lot of cost
savings in the business." Norris, 6/8/99am, 10:18-23.
- In fact, Dean Schmalensee's chart shows that
- redacted -
b. Microsoft used its monopoly power to punish OEMs that
refused to facilitate its exclusion of rivals
204. Microsoft used its monopoly power to impose -- or threatened to impose --
penalties on OEMs that refused to aid its exclusionary strategy.
(1) Microsoft threatened "MDA repercussions" if IBM
continued to bundle Netscape
205. In early 1997, Microsoft began trying to convince IBM to promote and
distribute the upcoming release of its new browser, Internet Explorer 4.0.
205.1. At a March 6, 1997, meeting with IBM, representatives of
Microsoft threatened that, if IBM did not pre-load and promote Internet Explorer 4.0
exclusively on its PCs (in other words, to the exclusion of Netscape Navigator), it would
suffer "MDA repercussions."
- Norris, 6/8/99am, at 29:11 - 32:5. Norris' handwritten notes taken
during that meeting record that Microsoft representatives told IBM
"No Netscape and receive more MDA $ across the PC Co.,"
meaning that IBM would receive greater discounts against its
Windows royalty rate, if IBM did not promote Netscape. GX 2164,
at 80283. Microsoft also told IBM that it must "promote IE 4.0
exclusively -- if not MDA repercussions," in other words, fewer
discounts. GX 2164, at 80284.
- Microsoft's threat of "MDA repercussions" meant both that
previously offered MDA dollars for exclusive Internet Explorer
distribution would not be made available and that "the MDA we
[IBM] currently had on the table, it was up to their [Microsoft's] sole
discretion as to whether we met them, and they may decide we
didn't meet them."
- Microsoft's OEM account representative, Bengt Akerlind, made
clear the meaning of Microsoft's requirement that IBM distribute
and promote Internet Explorer "exclusively:" "Bengt was very
specific. He said, "No Netscape." Norris, 6/8/99am, at 29:11-14.
IBM "would not be able to load Netscape . . . . It would have to be
Internet Explorer only." Norris, 6/8/99am, at 30:2-14.
- Akerlind also stated that in return for IBM shipping a so-called
"neutral system," by which Microsoft meant PCs with "IE 4.0,
without any competitive software," IBM not only would receive "soft
dollars" but would also receive "new access to the Windows 95 and
BackOffice source code," and the ability to self-certify for
Microsoft's WHQL provisions, which would have helped the IBM
PC Company avoid losing valuable time to market. Norris
6/8/99am, at 27:10 - 29:10; GX 2164 at 80283.
205.2. In a follow-up meeting with IBM on March 27, 1997, Microsoft
representatives again insisted that IBM distribute and promote Internet Explorer
exclusively and offered soft dollar marketing incentives and possible MDA reductions in
return. In what Microsoft called a "private" or "secret" portion of this meeting,
Microsoft's Mr. Akerlind specifically stated that "We have a problem if you load
Netscape," referring to the Navigator browser.
- Ted Hannum, Microsoft's OEM account representative for IBM, told
Norris in a March 21, 1997 telephone conversation planning the
March 27 meeting that, following the main meeting, Microsoft
wanted to have a "private meeting. Ted called secret discussions,"
because Microsoft was "ready to put a proposal on the table for
Internet Explorer." GX 2166, at 80292; Norris, 6/8/99am, at 36:4 -
37:1.
- During the "secret discussions" on March 27, Norris'
contemporaneous notes record that "After we exchanged niceties,
the first thing Bengt said was, 'we have a problem if you load
Netscape," referring to Netscape's Navigator. GX, 2168 at 80298;
Norris, 6/8/99am, at 48:20 - 50:18. Norris, confirmed that this
statement was the same as Akerlind had made in the previous
meeting, and meant that "if we [IBM] were not willing to do some of
the exclusive preloads that they were offering, that there would be
MDA repercussions." Norris, 6/8/99am, at 50:9-18.
- Akerlind reiterated that by "neutral system"Microsoft meant the
Internet Explorer 4.0 browser and other Microsoft software "would
be installed on that neutral system and that there would be no
competing software or applications installed." GX 2168, at 80299;
Norris 6/8/99am, at 47:11 -48:9.
- Norris' planning memo for the March 27 meeting, based on his prior
conversations with Hannum and Akerlind, corroborates that one of
Microsoft's objectives was "Demonstrate a prototype of I.E. 4.0 and
gain IBM's commitment to 'exclusively' and jointly promote I.E. 4.0
as the navigator [sic] of choice. In return Microsoft will offer IBM
soft dollar marketing incentives," and, Norris testified, "possible
MDA reductions." The memo also confirms that, "if IBM
'neutralizes its desktops" and other PCs, Microsoft would allow it to
self certify Windows 95 for WHQL and get access to Windows 95
source code. GX 2167; Norris, 6/8/99am, at 41:1 - 44:23.
- See GX 2203 at 93800 (a March 12, 1997 email to the general
manager of IBM's Internet Division corroborating that "Last week
Microsoft discussed the potential for IBM PC Co. to be part of the
IE 4.0 launch and rollout. In order for us to be part of it, they want
us to use IE 4.0 exclusively."); Norris, 6/9/99pm, at 90:16 - 91:23.
205.3. IBM ultimately refused Microsoft's proposal that it exclusively ship
Internet Explorer 4.0 and stop shipping Netscape because of the importance to IBM of
maintaining browser neturality.
- GX 2203, at 93798 (concluding that accepting Microsoft's proposal
and shipping Internet Explorer 4.0 exclusively is unacceptable
because "IBM has to maintain a browser neutral stance").
(2) Microsoft threatened to harm Gateway if it
supported or bundled Netscape
206. Microsoft also told Gateway that its featuring of Netscape would harm its
relationship with Microsoft.
- A Gateway employee reported that Microsoft's OEM account manager
told Gateway that its's use of Netscape on Gateway's corporate intranet
was "a HUGE issue with MS" and that "MS wants to get back to doing co-marketing and sales campaigns with GW, but they won't if they see GW is
anything but pro Microsoft." GX 308; GX 652 (Gateway CID response)
(sealed); Von Holle Dep., 1/13/99, at 312:23 - 314:8.
- The same Gateway employee also noted that "Dell turned Netscape down
because they did not want to turn their relationship with Microsoft.
Therefore, they get special things because of it." GX 308.
(3) Microsoft repeatedly penalized IBM for competing
against Microsoft
207. In addition to penalizing OEMs that featured Netscape, Microsoft more
generally discriminated against IBM for featuring competing products. These products
included both IBM's OS/2 operating system, which competed against Windows, and
various application programs. IBM's bundling of non-Microsoft applications deprived
Microsoft of complementary revenues from the sale of its own applications -- a principal
source of Microsoft's monopoly profits.
(a) Microsoft withheld a Windows 95 license
from IBM until 15 minutes before the
product's launch because of IBM's
preloading of competing products
208. Microsoft delayed granting an essential Windows 95 license to IBM until 15
minutes before the product's launch because of IBM's preloading of competing
products.
208.1. The IBM PC Company began negotiations with Microsoft for a
Windows 95 license in approximately March 1995.
- Norris, 6/7/99am, at 24:5-14.
- Garry Norris, who was Program Director for Software Strategy and
Strategic Relations for the IBM PC Company from approximately
March 1995 to April 1997, testified about his extensive first-hand
experience as the IBM representative primarily responsible for
leading IBM's Windows negotiations with Microsoft during this time.
Norris, 6/7/99am, at 6:1-6; Norris, 6/7/99am, at 7:25 - 8:8. Norris
personally served as the lead IBM negotiator for the Windows 95
license, negotiating with Microsoft's OEM account representative
for IBM, Mark Baber, and with Joachim Kempin. Norris, 6/7/99am,
at 24:15 - 25:1.
208.2. From March until approximately early June 1995, IBM's Windows
95 license negotiations with Microsoft proceeded relatively smoothly. Then, beginning
in June, Microsoft slowed the pace of the license negotiations significantly.
- Norris, 6/7/99am, at 25:2 - 26:5
- By contrast to the first two months of the negotiations, Microsoft
suddenly stopped returning IBM's phone calls, and halted or
seriously delayed returning to IBM marked-up drafts of license
language to IBM. Norris, 6/7/99am, at 26:6-20.
- Norris contemporaneously memorialized Microsoft's various
delaying tactics in a July 19, 1995 e-mail to his superiors. Among
other Microsoft actions, he recorded that "There has been very little
progress over the last two weeks . . . ; MCSFT team has been non
responsive; even on the smallest of items we have made no
progress; no sense or urgency; cancellation of meetings; not
delivering faxes when promised; they have been acting on a part
time basis and disjointed manner; taking days to get simple issues
resolved . . . ." GX 2199; Norris, 6/7/99am, at 28:7 - 30:19; GX
2197.
208.3. Microsoft's unresponsiveness and its actions to delay the
Windows 95 license negotiations with IBM followed soon after IBM announced that it
was seeking to acquire Lotus Development Corp., a direct competitor of Microsoft in
messaging and office suite software, among other products.
- IBM announced on June 5, 1995 a hostile takeover attempt of
Lotus. On June 11, IBM and Lotus reached an agreement for IBM
to acquire Lotus. IBM completed the acquisition on approximately
July 5, 1995. Norris, 6/7/99am, at 35:9 - 36:12.
- Among other products, Lotus offered Lotus Notes, an e-mail
product, and Lotus SmartSuite, an office productivity suite, that
competed with Microsoft's software. Norris, 6/7/99am, at 35:21 -
36:5.
208.4. Following IBM's June 5, 1995, announcement but before Microsoft
began to slow the pace of the Windows 95 negotiations, Microsoft repeatedly
questioned IBM representatives about IBM's plans for Lotus' products that competed
with Microsoft's products and expressed serious concern that such plans might place
the two companies on a "collision course."
- On several occasions after June 5, Mark Baber of Microsoft asked
Mr. Norris "what are your plans for Lotus? What are IBM's plans?
Do you plan on pre-loading SmartSuite? Are you going to drop
SmartSuite in the boxes of your PC systems? Exactly what do you
plan on doing with SmartSuite?" Norris, 6/7/99am, at 36:13-25.
- On June 27, 1995, Joachim Kempin met with IBM executive Tony
Santelli and Roy Clauson. Santelli's contemporaneous report of
that meeting reflects that "Joachim expressed concern that our
companies are headed on a collision course." Among the specific
issues raised by Mr. Kempin were: 2) Microsoft needs to better
understand what's behind the Lotus deal. LVG [Louis V. Gerstner]
should have called Gates to explain. 3) He [Kempin] heard rumors
in South America (he just returned) that IBM was planning to
preload 'Lotus Smart Suite' on all IBM PC's and sell it to 'OEM's for
$5/copy.'" Kempin felt strongly that a meeting between Gates and
Gerstner was "crucial," and that one of the agenda items should be
"Better understanding of IBM Lotus strategy." GX 2153; Norris,
6/7/99am, at 38:15 - 40:20.
- The other IBM participant in the meeting confirms Mr. Santelli's
report. A June 28, 1995, Clauson e-mail states that "Joachim
thinks Lou and Bill need to talk, at times, as a professional
courtesy, when significant events are about to take place." Kempin
also stated that "MS is definitely worried about SmartSuite being
given away and eating into their 'office heartland'" and that "There
are lots of 'combative' people in MS ready to go to war with IBM."
GX 2204.
208.5. On approximately July 17, 1995, after IBM's acquisition of Lotus,
IBM announced that it was going to make Lotus SmartSuite "the primary desktop
offering from IBM" in the United States.
- Norris, 6/7/99am, at 42:17 - 43:2.
208.6. On July 20, 1995, just three days after IBM's announcement that it
would feature SmartSuite on its PCs, Microsoft informed IBM that it was cutting off
altogether further Windows 95 negotiations with IBM. Microsoft's purported reason for
halting negotiations was that it wanted first to resolve an ongoing, unrelated audit of
IBM's past royalty payments. Microsoft also cut off IBM's access to Windows 95 code
that IBM needed for its PC product planning and development.
- Mark Baber of Microsoft called Norris on July 20 and told him that
he had been instructed by senior Microsoft executives, specifically
Bill Gates, Steve Ballmer, and Joachim Kempin, to halt further
negotiations with IBM. Norris, 6/7/99am, at 30:20 - 31:6. This call
came just three days after IBM's announcement of its plans for
SmartSuite. Norris, 6/7/99am, at 43:3-8.
- Microsoft "had cut off access to the code. Without the code, we
couldn't continue development of our products." Norris, 6/7/99am,
at 50:9-21.
208.6.1. Prior to this July 20 call, there had never been any
connection or linkage between the ongoing audit and IBM's negotiations for a Windows
95 license.
- "The audit was never linked to the Windows 95 license
agreement in the past. We had never discussed the audit
being a part of the Windows 95 license agreement. They
had never mentioned that the audit was related to the
Windows 95 license agreement." Norris, 6/7/99am, at 32:4-18.
- Because there had never been any connection between the
two issues before July 20, Norris was "quite surprised" by
Baber's call telling him that Gates and others were now
linking the two issues and halting all further negotiations.
Before July 20, Norris, the lead negotiator for the Windows
license, had no involvement "at all" in the audit. Norris,
6/7/99am, at 31:20 - 32:12.
208.6.2. The decision to halt all further Windows 95 license
negotiations, ostensibly until the audit had been resolved, was made entirely by
Microsoft. IBM strongly opposed any linkage that would impede its negotiation and
receipt of a Windows 95 license and actively sought to resume negotiations as quickly
as possible.
208.6.2.1. Contemporaneous documents confirm that it was
Microsoft, not IBM, that linked the audit to the negotiations and used it as a pretext to
cut off all further discussions in those negotiations.
- Even before Microsoft cut off the Windows 95
negotiations, IBM informed Microsoft that those
negotiations were completely unrelated to the
ongoing audit. In a July 18, 1995, letter concerning
the audit, an IBM executive involved with the audit
stated unequivocally that: "With respect to your
comments about the signing of the Windows 95
license agreement, we consider the Windows 95
contract negotiations to be completely separate and
unrelated to the audit and are actively negotiating
with Mark Baber to close a Windows 95 agreement as
soon as possible." GX 2371.
- On July 20, 1995, the day IBM learned that Microsoft
was halting further Windows license negotiations, IBM
immediately protested in a letter from Rick Thoman to
Bill Gates, which Norris helped write. IBM told Gates
that IBM had been frustrated with the previous pace
of negotiations but that "Today, Microsoft introduced
a new issue, the pace of an existing contract audit,
the settlement of which your team wants as a
condition of finishing the contract and shipping
product. This is a complete reversal of Microsoft's
prior position. There are no conditions under which
this complex audit can be closed immediately." GX
2197 (emphasis added); Norris, 6/7/99am, at 34:22 -
35:8; Norris, 6/7/99am, at 46:6-20 (because the
several audits were being done consecutively, there
was "no way" that the audit could have been
completely resolved by the August 24, 1995,
Windows 95 release date).
- In a July 24, 1995, telephone conversation with
Thoman, Bill Gates stated that it was the decision of
"senior management" at Microsoft, specifically Gates
and Mr. Ballmer, to cut off the Windows negotiations
and link them to completion of the audit. Norris,
6/7/99am, at 45:7-11.
- Thoman advised Kempin in an August 3, 1995, letter
that "IBM is very serious about pursuing a Windows
95 license agreement. . . . I would hope IBM's intent
is clear. The PC Company has already stated
publicly that it will ship and support Windows 95."
Thoman also told Kempin that IBM's OS/2 group
would continue to compete with Windows 95 but that
the competition "should not be an issue on how you
work together with the PC Company." GX 2196;
Norris, 6/7/99am, at 47:19 - 49:25.
- IBM made vigorous attempts to resolve the audit as
quickly as possible so that negotiations for its
Windows 95 license could resume. First, as an
indication of its good faith, IBM offered a $10 million
prepayment to Microsoft to clear up any
discrepancies that the audit might ultimately have
revealed. Norris, 6/7/99am, at 45:12-21. Second,
IBM agreed that in any Windows 95 license
agreement it would pay penalties and interest if any
future audit disclosed underreporting by IBM. GX
2196; Norris, 6/7/99am, at 46:21 - 47:7; GX 2197
(IBM's Thoman told Gates on July 20 that the
previous pace of the audit was due to the
complexities of the issues and that "the auditors have
told IBM that we have cooperated fully.").
- Tony Santelli told Joachim Kempin in an August 21,
1995 letter that, given that all of the issues
concerning the Windows 95 license were resolved, "I
do not understand your reasons for continuing to link
delivery of the Windows 95 code to resolution of the
audit. . . . I think we should unlink the audit settlement
from the Windows 95 code delivery." GX 2139.
208.6.2.2. Microsoft's attempts at trial to suggest that it was
actually IBM, not Microsoft, that wanted to link the audit to the Windows license
negotiations are erroneous and disingenuous.
- Microsoft questioned Norris about DX 2638, a
July 14, 1995, letter from Nell Miller at Microsoft to
James Miller at IBM concerning the audit. In that
letter, Microsoft says that IBM's Miller had said that
IBM might need to stop the audit and was "not sure
you can sign the license for Windows 95 at this time."
Norris, 6/9/99am, 54:21 - 55:16.
- Microsoft did not introduce or mention, however,
IBM's response, dated just four days later, in which
the version of events Microsoft now advances is flatly
rejected. On July 18, Jim Miller of IBM wrote back to
Microsoft concerning the audit issues: "With respect
to your comments about the signing of the Windows
95 license agreement, we consider the Windows 95
contract negotiations to be completely separate and
unrelated to the audit and are actively negotiating
with Mark Baber to close a Windows 95 agreement as
soon as possible." GX 2371 (emphasis added).
208.6.3. Microsoft was willing to resume Windows 95 license
negotiations with IBM, by more readily resolving the audit dispute, on the condition that
IBM would agree not to ship the competing SmartSuite product on its PCs for a
minimum of six to twelve months. In fact, Joachim Kempin expressly told IBM
executives that Microsoft would accept a lower payment to settle the audit if IBM
agreed not to ship SmartSuite.
- On August 9, 1995, IBM's
Santelli met with Kempin and
Baber of Microsoft. Santelli's
memo of the meeting, prepared
just a few days later, records
Kempin's statements. Kempin
discussed a payment to settle the
audit, and then "suggested this
payment may be reduced
through an offer of good faith
from IBM . . . something he could
show Gates. He sugested IBM
not bundle Lotus SmartSuite on
our systems for a minimum of six
months to one year." GX 2195;
Norris, 6/7/99am, at 51:9 - 54:13
- Santelli also recorded in his memo Kempin's explanation for
Microsoft attempting to keep IBM from shipping software
products that competed with Microsoft's products. Santelli
wrote that Kempin "is concerned on the impact to Microsoft
profit from Office if they begin offering to OEM's as a
preload. They view the 'threat of bundling' as a 'core issue'
in the relationship." GX 2195. Norris confirmed that this is
consistent with his understanding of the meeting at the time
it occurred (Norris, 6/7/99am, at 53:17 - 54:13) and also
consistent with the concerns Kempin had previously
expressed about IBM possibly distributing SmartSuite with
its PCs. Norris, 6/7/99am, at 54:24 - 55:17; GX 2204.
- Kempin confirmed his August 9 proposal to help settle the
audit if IBM agreed not to ship SmartSuite in a August 15
letter to Santelli. Kempin noted that Santelli had told him
that IBM wanted to resolve the issues and complete its
Windows 95 license quickly, but Kempin also warned that
resolution of the audit "could lead to a delay in our ability to
conclude the Windows 95 agreement." Kempin concluded
by saying that "let me come back to one of my key points in
our discussion. If you believe that the amount I am asking
for is too much, I would be willing to trade certain
relationship improving measures for the settlement charges
and/or convert some of the amounts into marketing funds if
IBM too agrees to promote Microsoft's software products
together with their hardware offerings." GX 2138.
Microsoft's willingness to "trade certain relationship-improving measures" referred to trying to gain IBM's
commitment not to compete with Microsoft with SmartSuite.
Norris, 6/7/99am, at 60:10 - 61:5.
208.7. Microsoft ultimately permitted IBM to sign a Windows 95 license
agreement, but only 15 minutes before the official launch of Windows 95 occurred.
- Norris, 6/7/99am, at 63:14-23.
(b) Microsoft conditioned access to critical
marketing support, and other terms and
conditions for Windows provided to other
OEMs, on IBM's not preloading competing
products with the PCs it shipped
209. Microsoft also sought to induce IBM not to ship rival software by
conditioning lower prices and other competitively important resources on IBM's
agreement not to ship software that competed with Microsoft products.
209.1. First, in late 1994, Microsoft attempted to induce IBM to reduce or
eliminate competition from IBM's rival operating system product, OS/2.
- Norris testified that, in the second half of 1994, Microsoft proposed
that IBM enter into a "Frontline Partnership" that would have
required IBM to reduce or eliminate its shipments of OS/2, IBM's
operating system that competed with Windows. Norris, 6/7/99am,
at 13:10 - 14:13.
209.1.1. IBM rejected Microsoft's proposal to reduce operating
system competition with it and decided, instead, to compete vigorously with Microsoft
through OS/2.
- Norris, 6/7/99am, at 13:24 - 14:13.
209.1.2. Microsoft penalized the IBM PC Company in several
significant ways for its refusal to accept Microsoft's proposal not to compete and its
decision to continue its operating system competition.
- First, as a consequence of IBM's refusal to reduce its
competition with Microsoft, Microsoft reduced from three to
one the number of Microsoft OEM account managers that
handled the IBM relationship. This reduced support
adversely affected IBM's ability to test, manufacture, and
ship its PCs to its customers. As a result, IBM lost time to
market for its PC products. Norris, 6/7/99am, at 15:13 -
16:3.
- Second, Microsoft informed IBM that it would treat IBM, not
like other large OEMs such as Compaq, Dell and HP, but
rather like any of the other hundreds of much smaller OEMs.
Norris, 6/7/99am, at 14:14 - 15:9.
- Third, IBM's beginning price for Windows 95 was $75 per
copy, a dramatic increase from the $9 royalty IBM was then
paying for Windows 3.1. Norris, 6/7/99am, at 20:1-12; GX
2132.
209.1.3. Microsoft's representatives expressly stated that it was
penalizing IBM because IBM continued to compete with Microsoft.
- See infra ¶ 209.2.1.
209.2. Second, in the fall of 1995, Microsoft again advised IBM that it
could have a "Frontline Partnership" with Microsoft if it reduced or eliminated its
shipments of certain software products that competed with Microsoft's products.
209.2.1. Microsoft told IBM that, as long as the PC Company
shipped competing software products on its PCs, IBM would get less attractive
Windows prices, marketing support, access to Microsoft account representatives and
technical personnel, and access to Microsoft enabling programs than if it did not ship
competing products.
- Mark Baber told Norris that, because the IBM PC Company
competed with Microsoft, IBM would not enjoy the benefits
of the tier 1 OEMs, but rather would be treated like any other
"tier 3" OEM. Norris, 6/7/99am, at 74:21 - 76:8.
- IBM received this tier 3 treatment even though it was
shipping approximately 4 to 5 million PCs during this time
period. The royalties paid by the IBM PC Co. to Microsoft
increased from approximately $40 million in 1995 to $220
million in 1996, to $330 million in 1997, and to approximately
$440 million in 1998. Norris, 6/7/99am, at 76:14 - 77:5.
- In a January 5, 1996, letter to Santelli, Microsoft's Joachim
Kempin stated that "I firmly believe that the best solutions to
customers around the world will get delivered by partners
who closely cooperate and share common goals. As long
as IBM is working first on their competitive offerings and
prefers to fiercely compete with us in critical areas, we
should just be honest with each other and admit that such
priorities will not lead to a most exciting relationship and
might not even make IBM feel good when selling solutions
based on Microsoft's products." Kempin added: "You are a
valued OEM customer of Microsoft, with whom we will
cooperate as much as your self-imposed restraints allow us
to do." GX 2142; Norris, 6/7/99am, at 82:15 - 83:24.
209.2.2. The IBM PC Company was denied access to Microsoft's
so-called "enabling programs," in which IBM's competitors such as Compaq, HP, and
DEC participated, because it continued to ship products that competed with Microsoft's.
- Surveys conducted by Norris indicated that Microsoft's
refusal to allow IBM to participate in the "enabling programs"
cost IBM over $180 million because it resulted in "customer
perception that IBM lacked a good relationship with
Microsoft." Norris, 6/7/99pm, at 30:1-6. In spite of persistent
attempts to gain access to these "very important" programs
the IBM PC Company was not permitted by Microsoft to
participate. Norris, 6/8/99am, at 44:6 - 45:14.
- These enabling programs included the Microsoft Authorized
Support Center, the Microsoft Certified Solution Provider
Program, and the Authorized Technical Education Center
Program. Norris, 6/7/99am, at 77:6 - 78:14.
- Microsoft blocked IBM's participation in the "enabling
programs" despite the benefit to Windows users of IBM's
participation. Norris, 6/7/99pm, at 36:9-15.
209.2.3. Microsoft's Mr. Gates and Mr. Kempin discussed as early
as March 1994 using Microsoft's global relationship with IBM and its OEM relationship
to apply pressure to IBM not to deal with Lotus.
- In a March 20, 1994, e-mail, Gates wrote to Kempin: "This is
one topic I really want to try to get to the bottom of. Why
does IBM help Lotus so much? Is there anything we can do
about this? Should it become an issue in our global
relationship with IBM?" Kempin's response acknowledges
that Microsoft has mentioned the Lotus situation "as an
issue" in recent "partnership" talks with IBM, and later states
that "I am unsure if we need to see this as an organizational
issue or an OEM issue. I am willing to do whatever it takes
to kick them out, but strongly believe we need a worldwide
hit team to attack IBM as a large account, whereby the OEM
relationship should be used to apply some pressure. GX
328 (emphasis added).
209.3. Third, Microsoft used similar tactics in 1996 and 1997 in a further
effort to convince IBM not to ship the competing Lotus SmartSuite and Notes products.
- In a January 31, 1996, call with Santelli, Joachim Kempin "opened
the discussion expressing a strong concern about IBM PC
Company bundling Lotus SmartSuite. He has two issues; first, why
didn't Microsoft get a chance to compete, and second, it makes our
attempt to improve our relationship more difficult because when
PCCO wins, Lotus wins & Microsoft loses." GX 2157; Norris,
6/7/99am, at 85:4 - 86:5. Kempin's concern represented "no
change in the theme. As long as we are competing, the
relationship was going to be difficult." Norris, 6/7/99am, at 86:6-12.
- In a February 19, 1997, meeting with IBM, Bengt Akerlind stated
that Bill Gates was "really upset" that the IBM PC company
continued to compete with Microsoft by shipping SmartSuite and
Notes. Gates asked "Why are we working with IBM when they are
doing these deals? Don't work with them!" GX 2163, at 80278.
Norris 6/8/99am, at 19:22 - 20:25. Akerlind asked IBM "How
religious is our support of SmartSuite," GX 2163, at 80278
(referring to IBM's support of SmartSuite), meaning "what would it
take to get you to not load SmartSuite?" Norris, 6/8/99am, at 21:1-11.
- Similarly, in March 1997, Microsoft representatives conditioned
IBM's access to Windows 95 source code and permission for it to
self-certify its complaince with Microsoft's WHQL requirements,
both of which were of considerable value to IBM (see supra ¶
203.3.3), on IBM's agreement to remove SmartSuite, Notes, and
Netscape from its PCs and ship Microsoft's software applications.
See supra ¶¶ 198.1.iv; 198.2.iii & iv; GX 2164, at 80284 (Norris
notes of March 6 meeting: "Bengt: SmartSuite, WorldBook, Notes.
Remove objectionable apps and make the systmes neutral.");
Norris, 6/8/99am, at 32:19 - 33:10 (Microsoft "wanted us to stop
loading those applications, remove them, and start loading
Microsoft applications. Neutral.").
- Mark Baber of Microsoft told Norris expressly in the April to June
1996 timeframe that Kempin would not meet with Santelli of IBM to
discuss Windows royalties or improving the IBM-Microsoft
relationship because IBM was distributing Lotus SmartSuite with its
PCs. GX 2183, at 90451; Norris, 6/7/99pm, at 19:11 - 20:18.
- Microsoft's Kempin also told IBM representatives that Microsoft
would not provide quotations for IBM to use in press releases for its
PCs in cases where IBM was bundling SmartSuite on its machines
or placing it in the box. GX 2193; Norris, 6/7/99pm, at 24:9 - 27:10.
209.4. Microsoft also threatened to withhold public endorsements for and
statements of cooperation with IBM because of the IBM PC Company's decision to ship
World Book, an electronic encyclopedia, with its PCs rather than shipping Microsoft's
competing encyclopedia, Encarta.
- GX 2158 (e-mail reporting that at a January 30, 1997 meeting,
Microsoft's IBM account representative reported that Mr. Gates
"was really mad about the World Book deal, given that IBM wants a
close relationship in this market with them."); Norris, 6/8/99am, at
11:4 - 15:24.
- Akerlind reported in the February 19 meeting that Gates wanted to
know why the Microsoft OEM team was continuing to work with
IBM in light of the fact that IBM was offering SmartSuite, Notes,
and WorldBook. GX 2163, at 80278.
(c) Microsoft sought to condition substantial
MDA price reductions on IBM's ceasing to
support competing products
210. Microsoft also sought to condition substantial MDA price discounts on
IBM's ceasing to ship with its PCs products that competed with Microsoft's products.
210.1. Microsoft proposed that at least $8, out of the total possible MDA
discount of $27 applicable to IBM's Windows 95 operating system license, be
conditioned on IBM's reducing or eliminating its shipments of and support for its
competing OS/2 operating system. Given the total number of IBM's PC shipments at
the time, Microsoft was in effect offering IBM roughly $40-48 million in Windows price
reductions if it would reduce or eliminate its OS/2 competition.
- The proposed 1995 MDA, sent by Microsoft to IBM on October 21,
1994, contained among other "milestone" activities the following:
"Adopt Windows 95 as the standard operating system for IBM
($3.00); "Windows 95 is the only OS mentioned in advertisement"
(2 milestones for total of $2.00); and shipping Windows 95
preinstalled on at least 50% of IBM PCs within two months after the
release of Windows 95 ($3.00). GX 2132; Norris, 6/7/99am, at
20:23 - 22:9.
- An $8 royalty reduction per PC shipped, multiplied across the 5-6
million PCs IBM was shipping at this point, would have totaled
approximately $40-48 million per year. Norris, 6/7/99am, at 21:20 -
22:1; Norris, 6/7/99am, at 22:12-18.
210.2. Microsoft's proposed Windows Desktop Family agreement,
presented to IBM in early 1996, conditioned IBM's receipt of future Windows 95 MDA
discounts on IBM's agreeing to a substantial increase in its royalty rate for Windows
3.1, the distribution of which Microsoft sought to discourage.
- Norris, 6/7/99pm, at 11:1 - 13:12.
- See also supra Part II.C.2.b; ¶ 37.1.
210.3. In meetings in early 1997, Microsoft expressly conditioned
additional MDA dollars on IBM's agreeing to distribute and promote Internet Explorer
exclusively and not to distribute or promote Netscape Navigator and other competing
software on IBM PCs.
- As recorded in Norris' contemporaneous notes of the meeting,
Akerlind told IBM that it "would not be able to load Netscape on that
system. It would have to be Internet Explorer only. And if we did,
we would receive more MDA dollars, not just on consumer
systems, but also on ThinkPad systems and also on desktop
systems. So every system that we shipped from the P.C.
Company." IBM also would receive "soft dollars" and payments for
joint marketing events. GX 2164, at 80283; Norris, 6/8/99am,
28:22 - 31:23. The quid pro quo Akerlind described, and that
Norris recorded in his notes, was clear: "No Netscape and receive
more MDA dollars across the P.C. Company." GX 2164, at 80283;
Norris, 6/8/99am, 28:22 - 31:23.
c. Microsoft's anticompetitive intent is evidenced by the
clear contrast in its treatment of IBM and Compaq
211. The anticompetitive character of Microsoft's conduct is confirmed by the
fact that Microsoft makes clear it favors OEMs that support its exclusionary strategies,
such as Compaq, and penalizes OEMs, such as IBM and Gateway, that refuse.
211.1. Microsoft representatives told IBM that its adverse treatment was
the result of IBM's competing with Microsoft and that IBM could have the same deal as
Compaq if it stopped competing with Microsoft.
- Norris was told directly by Microsoft representatives that "'as long
as you're competing with Microsoft, you will suffer in the market in
terms of prices, terms and conditions, marketing support programs,
and technical support programs.'" He also was told on several
specific occasions that "IBM can have Compaq's deal when it quits
competing." Norris, 6/7/99am, at 16:4-15; Norris, 6/7/99am, at
74:9-20.
- Norris reported to his superiors that Microsoft had told him directly
"'as long as you're competing with Microsoft, you're going to have
difficulties in this relationship.'" Norris, 6/7/99am, at 48:25 - 49:25.
- During negotiations about the Windows Desktop Family agreement
in early 1996, Mark Baber of Microsoft told Norris that "MS would
match Compaq if we made the same commitment that Compaq
did." GX 2180, at 13371. Microsoft representatives explained to
Norris that, "when IBM stops competing with Microsoft, then we can
have Compaq's deal: prices, terms and conditions." This meant
that IBM needed to cease shipping the competitive offerings that it
loaded on its PCs, in particular SmartSuite and, previously, OS/2.
Norris, 6/7/99pm, at 17:6 - 18:20.
211.2. Despite its representation that it treats all OEMs equally, Microsoft
in fact plainly favors Compaq and cooperating OEMs and penalizes OEMs, such as
IBM, that refuse to assist Microsoft in its exclusion of its rivals.
- Although Microsoft told OEMs that "all OEM's of Windows 95 are
treated equally by" Microsoft "in meeting the same OPK
requirements for shipping Windows 95 PC's to ensure the customer
of a consistent experience" (GX 294), in practice Microsoft favored
Compaq by permitting it to customize the start-up sequence in
Windows 95 in ways it denied to other OEMs (Romano Dep.
(played 12/16/98pm), at 52:10 - 54:6) and by permitting it to include
its own ISP sign-up process before other OEMs were allowed to do
so. Kempin, 2/24/99pm, at 42:22 - 43:9.
- GX 433, a 1993 Compaq "Microsoft Meeting Preparation"
document, includes slide 8 entitled "Judgment: How retaliatory
would they get?" (sealed; cited portion published). Compaq's John
Rose testified that he expected that this slide referred to possible
retaliation by Microsoft. Rose 2/18/99am, at 22:6-9. In contrast to
Compaq, which ultimately received favorable treatment on such
items from Microsoft as a consequence of not competing, the IBM
PC Company suffered precisely the sort of retaliation listed in GX
433 because it continued to ship with its PCs software that
competed with Microsoft's, including Netscape Navigator and OS/2.
Norris 6/8/99am, at 53:5 - 57:21.
- In an October 30, 1997 e-mail, Bill Gates observed that IBM
"continue[s] to use their PCs to distribute things against us." He
then flatly states that "Overall, we will never have the same
relationship with IBM that we have with Compaq, Dell and even
HP, because of their software ambitions." GX 257. Mr. Gates'
statement of the consequences to IBM of its "software ambitions" is
"very consistent" with the treatment IBM received from Microsoft.
Norris 6/8/99am, at 58:14 - 60:4.
- . See supra Part II.C.3; ¶ 38 (Microsoft's price discrimination favors
OEMs who assist Microsoft's exclusionary strategy).
- Cf. GX 2290 (October 23, 1997 internal Microsoft e-mail that
identifies particular software companies as "friend," "enemy," or
"neutral," and determined the level of support Microsoft would
provided based on those characterizations).
D. Microsoft entered into anticompetitive and exclusionary agreements
with OLSs and ISPs
212. As part of its campaign to maintain its operating system monopoly,
Microsoft entered into exclusionary agreements with the most important Internet
Service Providers (ISPs) and Online Services (OLSs).
212.1. ISPs and OLSs comprise one of the two most important channels
for obtaining and retaining browser market share. See infra Part V.D.1; ¶ 213.
212.2. Microsoft thus determined that gaining preferential treatment for
Internet Explorer through ISPs and OLSs, and excluding rivals, was vital to winning the
browser war. See infra Part V.D.1; ¶ 213.3.
212.3. Microsoft entered into exclusionary agreements with the most
important ISPs and OLSs. See infra Part V.D.2; ¶¶ 215-223.
212.3.1. Microsoft believed that, given a free choice, users would
choose non-Microsoft browsers. Microsoft's agreements thus not only required
preferred distribution and promotion of Internet Explorer, but also prohibited in most
circumstances the distribution and promotion of browser rivals.
212.3.2. To induce ISPs and OLSs to agree to the exclusionary
terms, Microsoft offered them a large payment; for the most part, that payment took the
form of barter, consisting of, among other valuable consideration, access to and
distribution through Windows.
212.4. The exclusionary terms in Microsoft's ISP and OLS agreements
lack any procompetitive purpose and can be explained only as part of a predatory
strategy to maintain Microsoft's operating system monopoly. See infra Part V.D.5; ¶
255.
- Microsoft determined that securing distribution for Internet
Explorer, and limiting Netscape's distribution, through leading
access providers was critical to gaining browser usage share
213. The ISP/OLS channel is one of the two most important (along with the
OEM channel) browser distribution channels.
- See infra Part VII.A.2.a; ¶¶ 363, 363.1.
- Cameron Myrhvold, Vice President of Microsoft's Internet Customer Unit
and Strategic Relationships, who oversaw Microsoft's relationship with
ISPs, testified that "'the ISP channel and the OEM channel are the two
most important channels for distribution.'" Myhrvold, 1/19/99pm, at 52:5-7
(quoting Myhrvold's deposition).
- Microsoft's economic expert, Dean Schmalensee, testified that it is
consistent with his understanding that ISPs and OEMs are the two most
important channels for distributing browsers. Schmalensee, 1/19/99pm,
at 52:5-17 (quoting Myhrvold's deposition).
213.1. ISPs and OLSs -- collectively, access providers -- provide access
to the Internet and, to facilitate their customers' ability to navigate the Internet, usually
distribute a browser to their customers.
- To reach the Internet, consumers subscribe, for a fee, to an access
provider's Internet service. GX 93. The access provider, in turn,
provides a communications link between the consumer's PC and
the access provider's server computers. The access provider's
servers, in turn, are part of the network of computers that
comprises the Internet itself. Myhrvold Dir. ¶¶ 16-17.
- Myhrvold testified that "ISPs typically distributed web browsing
software pre-configured for their service to make it easier for the
consumer to connect and use the internet." Myhrvold Dir. ¶ 17; see
also Colburn Dir. ¶ 7 (AOL distributes a browser to its members as
part of its client software).
213.2. Internet users tend to use the browser acquired with their
computers or through their Internet access providers.
- Brad Chase concluded, in April 1997, that "29% of all Internet users
in the US got their browser from their ISP," and, "we can't say it
enough, ISPs are our most important channel." GX 510, at MS7
004129 and MS7 004136; Chase, 2/16/99pm, at 23:23 - 24:23
(over 50% of users obtained their browsers from either OEMs or
ISPs/OLSs).
- Microsoft's Bjorn Hovstadius wrote on September 9, 1996, in
answer to the question "Why are ISPs important to our Internet
mission?" that "[e]very user that wants to get on the Internet needs
a connection, and that "[f]or a new user" ISPs are "probably their
first exposure to the Internet." GX 93. Hovstadius provided "data
that back[s] this up" showing that more "Internet users got their
browser from an ISP or OLS" than through any other channel. GX
93.
- A Microsoft presentation entitled "IE Market Review" written by
Kumar Mehta in April 1997 states that more Internet users acquired
their browser from an ISP during 1996 and 1997 than from any
other source. GX 415, at MSV 10551 - 10552.
- See infra Part VII.A.2.a; ¶¶ 363, 363.1 (detailing the reasons that
the ISP/OLS and OEM channels are more efficient and effective
than alternate browser distribution channels).
213.3. Microsoft thus believed that ISPs and OLSs "drive browser market
share" and that gaining preferential distribution and promotion for Internet Explorer
through ISPs and OLSs was critical to its objective of winning the browser war.
- Microsoft's December 1996 plan on "Working with ISPs in North
America" stated that "ISPs Drive Browser Market share. 35% of
end-user Internet access customers get their browser from an ISP."
GX 200 (emphasis in original); see also Myhrvold, 2/9/99pm, at
49:12-17, 62:2 -62:21 (testifying that what he meant when he wrote
GX 200 was that ISPs were "important for distribution" but later
conceding that distribution through ISPs would "result in usage"
because "it was a good way to access customers coming on to the
Internet.").
- Myhrvold, in February 1996, told Steve Ballmer, his entire sales
force, and others that network operators, including "Internet Access
Providers," are "an important potential asset in the battle for the
Internet." GX 472, at MS6 5003903.
- Brad Chase, Microsoft's Vice Present for Marketing, Personal and
Business System Division, wrote in a confidential April 4, 1996,
planning memorandum, entitled "Winning the Internet platform
battle," that licensing Internet Explorer to "all Internet Access
Providers" was "the best and fastest way to build share with new
users." GX 39.
213.4. Myrhvold's effort at trial to recant his earlier testimony (Myhrvold,
2/9/99pm, at 40:3 - 40:8 (asserting that it is "not necessarily true" that ISPs/OLSs and
OEMs are the two most important distribution channels)) is unpersuasive and
incredible:
- Myhrvold conceded that, when he testified at his deposition that
"the ISP channel and the OEM channel are the two most important
channels for distribution," he was relying on actual studies showing
the importance of the ISP/OLS channel. Myhrvold, 2/10/99am, at
33:2 - 19. As evidenced by what he wrote in his direct testimony,
he continues to believe that ISPs are an "important" channel of
distribution. Myhrvold Dir. ¶ 20.
- By contrast, Myhrvold's professed discovery -- while preparing for
trial -- that he "was wrong" about the relative importance of the
ISP/OLS channel was based on his understanding of Netscape's
marketing announcements about browser downloads, which
"fascinated" him. Myrhvold, 2/9/99pm, at 41:2 - 42:19. But
Myhrvold neither reviewed Barksdale's testimony concerning these
announcements nor compared his interpretation of these
announcements -- that downloading is more important for browser
distribution than access providers -- against the very data Microsoft
sponsored in this case. Myhrvold, 2/9/99pm, at 43:18 - 44:9.
- Had he done so, Myhrvold would have discovered that Microsoft's
own data showed -- contrary to the thrust of his trial testimony --
that the number of users who said they had acquired Netscape
Navigator by downloading did not change between the 1st and 3rd
Quarter of 1998. GX 1845 (Chart illustrating that 6.7 million users
said they had acquired Netscape Navigator by downloading in both
the 1st and 3rd Quarters of 1998); Chase, 2/11/99pm, at 4:12-20.
And he would have realized that the 12 million browsers
supposedly downloaded is more than the total number of browsers
in use, according to statistics that Microsoft itself sponsored at trial
and, therefore, that Netscape's claims about downloading must be
incorrect. Myhrvold, 2/9/99pm, at 41:4 - 44:16.
2. In furtherance of its goal of gaining browser usage share,
Microsoft entered into exclusionary agreements with the most
important ISPs and OLSs
214. Microsoft entered into exclusionary agreements with the major ISPs and
OLSs. These agreements, which covered subscribers accounting for more than
95 percent of the top 80 consumer Internet access providers, secured preferential
distribution for Internet Explorer and severely restricted the distribution and promotion of
non-Microsoft browsers by the most important access providers.
- Fisher Dir. ¶ 216.
a. Microsoft's exclusionary OLS agreements
215. In exchange for, among other inducements, prominent promotion in a
folder located on the Windows desktop (the "Online Services Folder"), AOL, AT&T,
CompuServe and Prodigy -- four of the most important Internet access providers --
agreed to a number of restrictions on their ability to promote and distribute non-Microsoft browsers:(2)
215.1. OLSs are required to distribute and promote Internet Explorer as
the exclusive or default browser.
215.2. OLSs are required to restrict severely their promotion of browsers
other than Internet Explorer. For instance, OLSs cannot express or imply to a customer
that another browser is available.
215.3. OLSs cannot provide a non-Microsoft browser to subscribers
unless a subscriber specifically asks the OLS to do so.
215.4. In no event -- even in response to specific subscriber requests --
can an OLS ship non-Microsoft browsers that in aggregate amount to more than
15 percent of the total number of browsers shipped by that OLS.
216. America Online's agreement with Microsoft is particularly restrictive:
- AOL agreed to "exclusively promote, market and distribute" Internet
Explorer. Microsoft and AOL License and Marketing Agreement. GX 804,
at AOL 0001738 (§ 7.1).
- Brad Chase, Microsoft's executive in charge of the AOL relationship,
wrote that Internet Explorer would be the "standard choice" for all AOL
customers; Chase further characterized the exceptions to AOL's use of
Internet Explorer as "pretty remote." GX 180.
- David Colburn, Senior Vice President of Business Affairs for AOL, testified
that the agreement "provided for virtual exclusivity in favor of Internet
Explorer on AOL" and that AOL was "only permitted to ship another
browser when required by a third party provider, distributor or corporate
account, and only after taking all reasonable efforts to cause the third
party to distribute the third party browser on its own; even then, the
number of third party browsers that AOL could distribute was limited to
less than 15% of AOL's total browser shipments." Colburn Dir. ¶ 29.
- Microsoft's restrictions applied in all channels through which AOL
distributed and promoted browsers; they had an effect, therefore, well
beyond subscribers who learned about AOL's service through the
promotion Microsoft provided AOL through Windows. Colburn Dir. ¶ 28;
GX 804, at AOL 0001738 (Microsoft/AOL contract, sections 7.1 and 7.2).
- Microsoft's agreement with AOL was so restrictive that, when Netscape
agreed to distribute and promote the AOL instant messenging service
("AIM"), AOL was not allowed to promote or distribute Netscape Navigator
through AOL's online service in return. Colburn Dir. ¶ 36; see also GX
826 (an internal Microsoft email described the AIM deal as "nothing major"
because, in the words of one Microsoft executive, Internet Explorer was
"still their default client, new users still get IE, old users still get upgraded
to IE."); GX 831.
b. Microsoft's exclusionary ISP agreements
217. Microsoft entered into similar restrictive agreements with ISPs. As with its
OLS agreements, Microsoft exchanged valuable consideration, including promotion for
ISPs through another folder on the Windows desktop (the Internet Connection Wizard,
or "ICW," which connected to Micosoft's Internet Referral Server). In exchange, the
ISPs typically agreed, in so-called Internet Referral Server ("IRS") agreements, to the
following restrictions on their ability to promote and distribute non-Microsoft browsers:(3)
217.1. ISPs must offer Internet Explorer as the standard or default Web
browser.
217.2. ISPs must restrict severely their ability to promote browsers other
than Internet Explorer, including agreeing not to express or imply that other browsers
are available.
217.3. ISPs may not provide another browser to a subscriber unless
specifically requested to do so by the subscriber.
217.4. Even when a customer specifically requests another browser, the
ISP cannot provide another browser if doing so would cause the total shipments of its
non-Microsoft browsers to exceed a specified percentage, typically 25%, of all browsers
shipped by that ISP.
217.5. Some ISPs entered into agreements with Microsoft that included
even more stringent distribution restrictions.
- Microsoft prevented Brigadoon from shipping other browsers with
more than 10% of total browser shipments. GX 1140.
- Microsoft prevented IDT from shipping other browsers with more
than 15% of total browser shipments. GX 1147.
218. The testimony of MCI's Stephen Von Rump illustrates how Microsoft's ISP
agreements disadvantaged non-Microsoft browsers.
-
MCI could not tell its customers that other browsers, such as Netscape,
were available or provide Netscape to customers unless specifically
requested. Von Rump Dep., 1/13/99, at 323:25 - 324:21. By contrast,
MCI was required to promote Internet Explorer as the "browser
recommended for use" with its ISP service. Von Rump Dep., 1/13/99, at
326:13-18.
- Microsoft required MCI to place an Internet Explorer logo on its Internet
service's home page, along with a link to an Internet Explorer download
site, but prohibited MCI from including similar links or promotion for
Netscape or taking paid advertising to promote Netscape Navigator. Von
Rump Dep., 1/13/99, at 324:23 - 327:4.
- As with its agreements with OLSs, Microsoft's restrictions applied to MCI
subscribers who became subscribers through means other than through
the ICW. GX 1132,at MS6 60008292 (Addendum A to the Microsoft/MCI
agreement, § 2.1) (sealed).
c. Microsoft's Exclusionary "Internet Explorer preferred"
agreements
219. Microsoft also entered into several hundred "IE preferred" agreements with
smaller ISPs.
- A Microsoft study indicated that ISPs representing 95% of Internet access
users had signed "IE Preferred" Agreements. GX 350.
- Fisher testified that "more than 95 percent of subscribers to ISPs in the
'Top 80' subscribe to ISPs that were contractually required to distribute IE
preferentially". Fisher Dir. ¶ 216 (referring to GX 12).
220. Microsoft granted these ISPs royalty-free rights to customize and distribute
Internet Explorer in return for their agreeing to make Internet Explorer the ISP's
"preferred" browser.
- In his direct, Cameron Myhrvold describes an ISP's ability to customize
Internet Explorer with the IEAK, which was free to ISPs. The IEAK allows
ISPs, when customizing Internet Explorer, to change the default home
page to point to the ISPs' service. Myhrvold Dir. ¶¶ 33, 34.
- In answer to the question "What are we offering ISPs?", Microsoft's Bjorn
Hovstadius wrote on September 9, 1996, "The basics - A license to
distribute IE for free. If an ISP is willing to make IE the preferred browser
and agree to a few other requirements in our license agreement we offer
to license IE and its add-on components for free. We allow them to
distribute another browser if they wish but it is very important that IE is the
preferred browser. We will not sign deals were that not the case." GX 93
- As Myhrvold conceded, Microsoft considered ISPs to be in breach of
these agreements if they did not make Internet Explorer the "default" or
"preferred" browser. Myhrvold, 2/10/99pm, at 42:3 - 43:7.
d. Microsoft anticipated that its exclusionary agreements
would wrest significant browser share from Netscape
221. Without restrictions on the distribution of rival browsers, Microsoft believed
that ISPs would give users a "side-by-side" choice of browsers and that users would
chose Netscape Navigator over Internet Explorer. Microsoft's restrictions were
specifically designed to wrest potential users away from Netscape.
- Cameron Myhrvold admitted that Microsoft believed the restrictions were
necessary precisely because of consumer demand. Microsoft feared that
offering users a "side by side" choice of browsers would result in users
choosing Netscape Navigator over Internet Explorer. Myhrvold,
2/10/99am, at 62:7-20.
- ISPs/OLSs favor giving users a choice of browsers. Myhrvold, 2/9/99pm,
at 72:12-14. See infra Part V.D.4.c.(2); ¶ 251.
- ISPs/OLSs initially resisted Microsoft's browser distribution restrictions
because they wanted more flexibility to meet consumer demand. See,
e.g., GX 198 (Prodigy believed that its Microsoft agreement contained "a
number of extremely objectionable provisions"); GX 228, at MS98
0113059 (Netcom resisted signing restrictive Internet Explorer 4 referral
server agreement that would impede its ability to satisfy its customers.)
222. Microsoft's exclusionary agreements with ISPs/OLSs covered the most
important access providers.
222.1. The ISPs and OLSs that agreed to Microsoft's OLS and IRS
agreements accounted for a large percentage of Internet access in the United States.
- By June 1997, 14 of the top 15 access providers in North America
were included in the OLS Folder or ICW and shipped Internet
Explorer as their preferred browser. GX 211.
- James Barksdale testified that, although there are thousands of
ISPs, over 75% of the world's Internet users access it from the 8 to
10 largest providers. Barksdale Dir. ¶ 129. Barksdale also referred
to a report on "Consumer Choice in Web Browsers" based on a
June 1998 survey of top ISPs which concluded: "The disturbing
reality is that the four largest retail Internet Service Providers, with
a combined subscriber base of over 20 million customers, distribute
only Internet Explorer to their customers." Barksdale Dir. ¶ 158.
- Colburn testified that AOL believed that with an AT&T deal,
Microsoft would own most of the consumer audience for browsers.
Colburn, 10/29/99pm, at 61:16 - 62:7; see also DX 502 (AOL
believed that, with a CompuServe deal, Microsoft would "own the
consumer franchise for browsers.").
222.2. The AOL deal was particularly significant to Microsoft's effort to
gain browser share because AOL was and remains the largest access provider.
- At the time it signed its agreement with Microsoft, AOL was (and
remains) the largest single access provider. Silverberg, 1/13/99, at
684:4 -685:10; Fisher Dir. ¶ 178. In 1998, AOL had in excess of 13
million subscribers, with its members generating over 1 billion Web
hits daily. Colburn Dir. ¶ 6.
- In an e-mail written to the executive staff in March 1996, Brad
Chase wrote, "This partnership significantly expands the IE
customer base for third party developers and shows how serious
we are about getting browser share." He states that U.S. market
share data indicates that, "Having AOL users in our camp gives our
Internet technologies and platform a powerful market presence."
GX 180.
- In January 1996, Bill Gates wrote that, "What we want from AOL is
that for a period of time - say 2 years - the browser that they give
out to their customers and the one they mention and put on their
pages and the one they exploit is ours and not Netscapes.We want
all the hits that come off of AOL to register on servers as our
browser so people can start seeing us as having measurable
browser share." DX 1545.
- Indeed, Microsoft's Ben Slivka reported to Paul Maritz and others
that having AOL, CompuServe, and MSN offering Microsoft
technology was akin to having "Compaq and IBM in the early MS-DOS days." GX 811.
222.3. The coverage of Microsoft's exclusionary agreements is even
broader when its "IE preferred" agreements are included:
- A Microsoft study entitled, "Netscape Competitive Analysis:
ISP/OLS Channel Revenue Segment" describes ISPs and OLSs by
their contractual relationship with Microsoft and Netscape. These
relationships are broken up into three categories: Internet Explorer
Preferred, Netscape Preferred, and Internet Explorer Parity.
Eighty-two percent of the ISPs surveyed which reported shipping
browsers were categorized by Microsoft as Internet Explorer
Preferred. GX 835, at MS98 0112828; see also GX 1092, at MS98
0112836 (same).
- ISPs that were required to make Internet Explorer their preferred
browser accounted for more than 96% of Internet access
subscribers. GX 835, at MS98 0112826-7.
- According to Microsoft's tracking documents, the vast majority of
the large ISPs in North America, and the four major North American
online services, have all entered into agreements with Microsoft
that require Internet Explorer to be the preferred or virtually
exclusive browser. GX 1833.
- Microsoft estimated that, by September 1996, more than 2,000
ISPs had signed preferential distribution agreements with
Microsoft. GX 93.
- A FY98 Mid-Year Review of the Internet Customer Unit reported
that 45 of the top 50 ISPs are "IE Preferred" and "82% of Breadth
ISPs ship IE as the primary browser." GX 424, at MS7 000588
(sealed).
223. The exclusionary provisions in Microsoft's contracts apply to all
browser distribution and promotion by the covered access providers, not just to
distribution and promotion of browsers to those subscribers acquired through the OLS
Folder or ICW.
- Myhrvold wrote on April 3, 1997: "Remember that ISPs have to swear
allegiance to IE for typically 75% of all the browsers they distribute in
order to get into the referral server." GX 440 (emphasis added).
- The shipments restrictions in Microsoft's contracts with ISPs also
specifically state that the restrictions cover "total shipments of all web
browsers through or by the ISP service". E.g., GX 1141, at MS6 500007
(Earthlink agreement, § 3.1) (sealed); GX 1140 (summary of the
Brigadoon agreement); GX 1147 (summary of the IDT Agreement); GX
1144, at MS6 5001130 (sealed) (SpryNet agreement, § 3.1); GX 1146,at
MS6 5000924 ( Mindspring agreement, § 3.1) (sealed); GX 1213, at MS6
5000386 (AT&T agreement, § 3.3) (sealed); GX 1214, at MS6 5000953
(Netcom Agreement, § 3.1) (sealed).
3. The importance of the exclusionary terms is evidenced
by how much Microsoft paid ISPs and OLSs to enter into
the agreements
224. Microsoft secured its exclusionary agreements by offering ISPs and OLSs
what amounted to a very large bribe. That bribe included both valuable technology
and technical assistance and Microsoft's provision of valuable promotion through
Windows.
- Professor Fisher testified that "Microsoft also made valuable concessions,
directly and indirectly, to the ISPs. These varied across ISPs but included
joint marketing programs, pricing deals, and discounts from referral fees
for users switched from competitive browsers." Fisher Dir. ¶ 182.
- Dr. Warren-Boulton testified that Microsoft made ISPs "an offer they
couldn't refuse" to distribute Internet Explorer to at least 75% of their
subscribers. Warren Boulton 11/30/98pm, at 24:16 - 26:22.
a. Microsoft paid significant value other than promotion
through Windows to induce ISPs and OLSs to agree to
its exclusionary terms
225. Microsoft provided OLSs and ISPs with valuable technology, technical
assistance, and other consideration for free, expecting both distribution for Internet
Explorer and the exclusion of browser rivals in return. This valuable consideration
included, among other things, the following:
225.1. Internet Explorer. Microsoft did not charge OLSs or ISPs for Internet
Explorer, even though Microsoft spent millions developing and improving Internet
Explorer, including a "componentized" version that ISPs and OLSs could use to
promote their own products and services.
- See infra Part V.G.
- Brad Chase described how Microsoft embarked upon a costly effort
to componentize Internet Explorer. Chase Dir. ¶¶ 19-21.
- Microsoft believed that giving away its technology would facilitate
its objective of persuading ISPs to agree to restrict their distribution
of rival browsers. See, e.g., GX 39, at MS6 5005720 ("You should
be able to break most of the Netscape licensing deals and return
them to our advantage because our browsers are free"); GX 472, at
MS6 5003904 (because it is "essential" to increase browser share,
Microsoft will "license at no cost the Internet Explorer for
distribution").
225.2. Technical assistance. Microsoft provided OLSs and ISPs with valuable
technical assistance for which Microsoft did not charge.
- Chase testified about the extensive technical aid Microsoft provided
to AOL, including hiring developers to work exclusively with AOL
and making "significant changes to Internet Explorer to meet some
of AOL's requests" Chase Dir. ¶¶ 51-52.
225.3. Source code. Microsoft gave AOL a free source code license to Internet
Explorer.
- Chase testified: "The willingness to divulge source code to AOL
demonstrated how far Microsoft was willing to go to win AOL's
business" Chase Dir. ¶ 42; see also Chase Dir. ¶ 69 (detailing
Microsoft's grant of source code rights for various versions of
Internet Explorer to AOL); GX 804,at AOL 0001724, -1759 (§ 1.1
and attachment 1) (granting source code license).
225.4. Customization of start page. Microsoft provided free browser
customization software, known as the Internet Explorer Administration Kit ("IEAK"), to
OLSs and ISPs. This software, among other things, allows OLSs and ISPs to change
the default home page on the browser to point to the OLSs' or ISPs' Web sites, rather
than to Microsoft's. In effect, Microsoft thus transferred to ISPs and OLSs revenues it
could have obtained from its home page.
- Cameron Myhrvold testified that Microsoft promoted ISPs' services
by, among other things, "granting each ISP the right to customize
Internet Explorer for its service." Further, he explained that
Microsoft licensed the IEAK at "no charge," allowing "each ISP to
preset the default home page so that customers would be taken to
the ISPs' Web site whenever they logged onto the Internet."
Myhrvold Dir. ¶¶ 10, 32-33. See also GX 39, at MS6 5005720
(describing a free customization kit available with Internet Explorer
3.0 which enabled ISPs to use their own brand and logo).
- Intuit's William Harris testified, based on a representation from
William Poole of Microsoft during the break of his cross
examination, that Microsoft will continue to allow ISPs to change
the default home page in Internet Explorer 5. Harris, 1/5/99am, at
42:16 - 43:3.
226. Microsoft in some cases paid cash (or its equivalent) to ISPs and OLSs to
facilitate their acceptance of restrictive terms.
- Microsoft paid off minimum commitments owed by the OLSs' and ISPs' to
Netscape, in order to induce a switch to Internet Explorer. Myhrvold Dir. ¶
29. For example, Brad Silverberg told AT&T that Microsoft would allow
AT&T to use revenue from ICW bounties to pay down the Netscape
minimum commitments if AT&T agreed to ship Internet Explorer on a
preferred basis. GX 179.
- Microsoft created co-marketing funds for ISPs that would offer Internet
Explorer on a standard or default basis. Myhrvold described a co-marketing fund of up to five million dollars which Microsoft created to
enable MCI to switch Web browsing software. Myhrvold Dir. ¶ 29.
- Microsoft offered to pay down AT&T's $17 million minimum commitment to
Netscape in return for AT&T's commitment to abandon Netscape
Navigator and sign a restrictive deal with Microsoft. GX 179.
- Microsoft offered discounts on the referral server bounties owed to it by
ISPs for every Netscape Navigator user (or user of other browser)
converted to Internet Explorer. Myhrvold Dir. ¶ 62; Myhrvold, 2/10/99pm,
at 11:18 - 13:2; see also GX 81 ("Here is what we have proposed for
upgrading Netcom's existing customers to IE. It is essentially a 'reverse
bounty' of $9."); GX 86 (discussing Microsoft's to "offer a [sic] exclusive
discount for all of their Netscape customers who move to IE (and not other
browser's users) in exchange for waiving the fee to get on the referral server.").
- Microsoft paid UUNet $500,000 for UUNet's abandonment of Netscape
Navigator. DX 2260,at 6 (§ 5.1) (UUNet Internet Referral Server
Agreement). Myhrvold's admission that the payment occurred (Myhrvold
Dir. ¶ 117) combined with the contemporaneous documents asking how
much it would take to "get Pipex [UUNet] off of Netscape" (GX 1812),
make Myhrvold's current denial that the payment had anything to do with
browsers dubious at best. Myhrvold, 2/10/99am, at 21:18 - 26:14.
Indeed, Myhrvold e-mailed a Microsoft employee working on the UUNet
account that "I actually think tying the payment to their shipping IE is a
great idea, though I would not do this formally." GX 102. His explanation
of this email at trial -- that he was merely attempting to provide
encouragement to a remote employee via email but immediately called
him to tell him it would not be "appropriate" is incredible. Myhrvold,
2/10/99pm, at 44:22 -45:20 (attempting to explain away the plain
language of GX 102)..
b. Microsoft also bribed ISPs and OLSs by offering what both
access providers and Microsoft viewed as valuable promotion
through Windows
227. One of the most valuable assets that Microsoft used to induce OLSs and
ISPs to enter into restrictive agreements was promotion of their services through
Windows.
- Dr. Warren-Boulton testified that Microsoft used promotion through
Windows to "induce OLSs to enter into agreements that restricted the
distribution and promotion of competing browsers." Warren-Boulton Dir. ¶
102.
- Professor Fisher testified: "Microsoft had power over what is referred to
here as real estate, positions on the desktop and in the system. One of
the things Microsoft could perfectly well have done would be to charge a
high price for that." Fisher, 1/12/99am, at 27:8-12.
(1) Promotion in Windows is valuable to ISPs and
OLSs because Windows is ubiquitous and users
tend to select Internet access providers promoted
through Windows
228. Promotion through Windows, whether in the start-up sequence, on the
Windows desktop, or in a prominent folder on the desktop, is extremely valuable to
ISPs and OLSs.
228.1. Because Windows is shipped on more than 90% of personal
computers, it provides the equivalent of a ubiquitous billboard, one that will be viewed
by millions of computer users. The impact is especially important for Internet-related
services; users, particularly novices, tend to select Internet-related services that are
prominently advertised in the start-up sequence or desktop.
- David Colburn testified that placement in the Windows box was
especially valuable in reaching novice users without an Internet
connection, a set of users Colburn described as "AOL's niche."
Therefore this distribution channel was "infinitely more fertile for
subscriber acquisition" than other distribution methods. Colburn,
10/29/98pm, at 45:5 - 46:2 In general, software already loaded on
a computer is a "uniquely effective method of distribution and
promotion," and placement of an icon "on the desktop would be a
uniquely effective method of promotion and installation." Colburn
Dir. ¶ 18
- MCI believed that the company's inclusion in the Internet
Connection Wizard which appeared in the course of using and
installing Windows was valuable because it would give the ISP
access to a large number of potential users. Von Rump Dep.,
1/13/99, at 322:2 - 323:15.
- Dr. Warren-Boulton testified that "the Windows 95/98 desktop and
boot-up sequence provide an attractive advertising vehicle for
OLSs. Placement on Windows screens is valuable to OLSs
because, among other reasons, it ensures that the OLS reaches
many potential new subscribers at the precise time when those
new subscribers must open an account to secure access to the
Internet. Warren-Boulton Dir. ¶ 101; see also supra Part V.C.1.a; ¶
176.1.
228.2. Promotion through Windows is also valuable because, in contrast
to other means of promoting Internet access services, such as mass mailing, it involves
virtually no costs except purchasing the placement itself from either Microsoft or OEMs.
It is a more cost-effective means of reaching potential subscribers than other methods,
such as mass advertising and mailing.
- Steve Case told James Barksdale that placement on the Windows
desktop was "extremely valuable" to ISPs because it meant
"immediate access to Windows users -- who constitute over 90% of
personal computer users -- without incurring the substantial hard-dollar costs associated with other distribution methods." Barksdale
Dir. ¶ 31.
- Cameron Myhrvold testified that inclusion in the Windows Referral
Server is an "economic channel" for ISPs because "ISPs can
acquire a customer via the Windows Referral Server at a lower cost
than they acquire a customer from their own sales and marketing
channels. Myhrvold Dir. ¶ 86 (explaining that the average cost paid
by ISPs through other channels is approximately $42.50 which is
higher than the referral fees that ISPs pay to Microsoft). Myhrvold
also admitted that the Internet Connection Wizard "represents a
good value to ISPs since it is a cheaper way of acquiring
customers than the industry average of acquiring customers. So I
would say for most ISPs it is a good value, yes." Myhrvold,
2/10/99am, at 31:5-9.
- Dean Schmalensee conceded that control over Windows gives
Microsoft the ability to obtain wide promotion and distribution of
products at minimal expense. Schmalensee, 1/19/99pm, at 41:13 -
42:11 (Microsoft considered placement on the Windows desktop an
important distribution channel for Internet Explorer).
- Microsoft also believed that its ability to promote and distribute
MSN through the Windows desktop gave MSN service a "huge
advantage over the competition.... AOL and CompuServe had to
spend $40 to $80 to acquire each new customer. It was very
expensive to offer bounties and ship free disks around the world.
In the meantime, MSN could acquire new customers virtually for
free." GX 1372, at 4-5.
(2) Microsoft created, and gave away, prominent
desktop placement for ISPs and OLSs that agreed
to its exclusionary terms
229. Microsoft created folders on the Windows desktop designed to promote
ISPs and OLSs that agreed to its exclusionary restrictions and used its power over
OEMs to prevent them from removing those folders.
- Microsoft created the Online Services Folder for the Windows 95 and 98
desktop which contained icons representing participating OLSs. A user
clicking on an OLS's icon is invited to register for the particular OLS's
Internet service. Warren-Boulton Dir. ¶ 102; Fisher ¶¶ 174-175.
- Microsoft also included with Windows, beginning with OSR 2.0, the
Internet Connection Wizard. Myhrvold Dir. ¶ 43. In Windows 95, the ICW
consisted of a icon prominently placed on the Windows desktop that,
when a user invokes it, takes the user to Microsoft's Referral Server,
which in turn lists several ISPs. As with the OLS folder, a user that
selects a particular ISP is invited to register for that ISP's service.
Myhrvold Dir. ¶ 44; GX 93 (explaining that the ICW allows users to choose
an ISP and complete the sign-up for an Internet account).
- In Windows 98, Microsoft moved the ICW up to the Windows boot up
sequence and thus made it even more prominent. GX 176A, at MSV
0009137 A.
230. Despite its current assertion that the OLS Folder and ICW were designed
merely to make it easier to connect to the Internet (Myhrvold Dir. ¶ 45), Microsoft
believed at the time it developed those folders that access providers would find them
extremely valuable and that the prospect of obtaining promotion through them would
induce access providers to agree to Microsoft's exclusionary restrictions. Microsoft
created the OLS folder and the ICW with the purpose of trading their value for exclusion
of rival browsers.
- Cameron Myhrvold admitted that "the referral server was partly created in
order to induce ISPs to commit to IE on a preferred basis." Myhrvold,
2/10/99am, at 29:12 - 31:25.
- Myhrvold described the requirements for being in the Online Services
Folder as "high" and explained that ISPs would have to agree to ship
Internet Explorer to at least 85% of their customers. He wrote that the
"Internet/ISP folder... will be the folder used to promote Internet access
along the lines we have discussed (commit to IE on exclusive/preferred
basis, co-branded startpage, bounty, etc.)" GX 185.
- In a 1996 market plan entitled "How to Get to 30% Share in Three
Months", Brad Chase wrote that Microsoft needed to "open up the
Windows box" in order to "remove barriers to browser adoption by Online
Services and Internet Access Providers." GX 334, at MS98 0104682.
231. Microsoft made clear during its negotiations with ISPs and OLSs, and in
the agreements it extracted, that access to the OLS Folder and Internet Referral Server
were linked to access providers' agreement to exclusionary restrictions.
- Microsoft executive Brad Silverberg testified that inclusion in the
"Windows box" provided "potentially great value" to access providers.
Silverberg Dep., 1/13/99, at 689:16-25. In exchange for giving access
providers this value, Silverberg explained, Microsoft would require
"exclusive or very very preferential treatment" for Internet Explorer. GX
183.
- As Silverberg told AT&T during negotiations: "You want to be part of the
Windows box, you're going to have to do something very special for us.
There are very, very few people we allow to be in the Windows box. If
you want that preferential treatment from us, which is extraordinary
treatment, we're going to want something very extraordinary from you."
Silverberg Dep., 1/13/99, at 692:12 - 693:25; see also GX 183.
232. Microsoft's Chairman Bill Gates' decision to use Windows placement to
extract exclusionary terms -- rather than to charge for such placement or use that
placement to advantage Microsoft's Internet access service, MSN -- illustrates the
importance to Microsoft of thwarting the threat that non-Microsoft browsers posed to its
operating system monopoly.
232.1. Mr. Gates initially took the position, in negotiating with AOL, that
distribution with Windows was "sacrosanct" and could not be part of any deal relating to
AOL's distribution of Internet Explorer.
- According to an internal AOL email reporting on a January 18,
1996 meeting between AOL and Microsoft, Gates made it clear that
"the Windows Box itself is 'sacrosanct.' No way AOL could drop the
entire client into Windows for distribution." GX 38.
- Chase testified that "Mr. Gates expressed frustration at Mr. Case's
focus on getting an AOL icon on the Windows desktop. Mr. Gates
said he would not agree to that demand." Chase Dir. ¶ 43. Chase
further explained, in reference to a January 26, 1996 meeting, that
"Mr. Gates strongly resisted the idea of promoting AOL's service
(which competed with MSN) by placing an AOL icon on the
Windows desktop." Chase Dir. ¶ 45.
- As Brad Silverberg testified, Gates was "very, very uncomfortable"
about giving AOL placement on the desktop because Gates "felt it
was putting a bullet through MSN's head." Silverberg Dep.,
1/13/99, at 703:13 - 705:11.
- Gates agreed with the Microsoft executive who ultimately became
responsible for MSN that including AOL on the desktop "gives
away our one unique and valuable asset -- Windows distribution --
at way too low a price. . . . The only real advantage [MSN] has in
this game is Windows distribution." GX 130.
232.2. Mr. Gates ultimately decided, however, that promoting Internet
Explorer was more important than protecting MSN's biggest competitive advantage of
being the only access provider with distribution through Windows.
- As Cusumano and Yoffie report, Gates concluded that bartering
promotion through Windows for exclusionary terms, thereby helping
to protect Microsoft's dominant position in operating systems, was
more valuable than protecting MSN. Gates said: "We have had
three options for how to use the 'Windows box:' First, we can use it
for the browser battle, recognizing that our core assets are at risk.
Second, we could monetize the box, and sell the real estate to the
highest bidder. Or third, we could use the box to sell and promote
internally content assets. I recognize that, by choosing to do the
first, we have leveled the playing field and reduced our
opportunities for competitive advantage with MSN." GX 1372, at
page 5.
232.3. In other words, Mr. Gates realized that securing preferential
distribution for Internet Explorer through AOL, and thus blunting the platform threat
Netscape posed, was worth a very substantial payment.
- David Colburn of AOL was told that Microsoft had "no limitations on what
it would spend to gain market share for Internet Explorer." Colburn Dir. ¶
38.
- AOL executives reported, in a write up detailing the meeting, that Mr.
Gates used "characteristically blunt" words during a meeting to express
this sentiment when he asked AOL "how much do we need to pay you to
screw NS?" GX 38.
- Professor Fisher testified that Microsoft has an incentive to pay AOL a
significant bribe to continue to favor Internet Explorer as an insurance
policy against a paradigm shift. Fisher, 6/1/99pm, at 66:25 - 67:12.
(3) As Microsoft predicted, OLSs and ISPs agreed to
its exclusionary restrictions to obtain valuable
desktop placement
233. As Microsoft anticipated, ISPs and OLSs saw the opportunity to be
included with Windows as very valuable and agreed to Microsoft's exclusionary
restrictions in order to obtain that desktop promotion.
- Prodigy concluded that placement in the Online Services Folder was
"absolutely critical to Prodigy's business" and "essential in order to remain
competitive." Thus, Prodigy "had no choice but to accept an agreement"
with Microsoft which contained "a number of extremely objectionable
provisions." Although Prodigy requested that a section of the agreement
limiting the number of competing browsers Prodigy could ship be deleted,
Microsoft refused. Microsoft was also unwilling to negotiate other terms,
such as the prohibition on including any links on the Prodigy Internet
service to browsers other than Internet Explorer. GX 198.
- Brad Chase understood that AT&T "really, really want[ed] to be in the
Windows box." Chase, 2/16/99am, at 67-68 (quoting GX 179). And
Microsoft executives reported after a meeting with AT&T that they were
still in "good shape on the browser as long as we hold strong on the
preferred status for getting 'in the box.'" GX 183.
- Cameron Myhrvold testified in his deposition that he had no doubt that
AT&T "very badly" wanted placement in the Windows box and that
Microsoft told them that they could not have this placement if AT&T "gave
equal placement to Netscape's Navigator." Myhrvold, 2/10/99am, at 18:12
- 21:3.
- Brad Silverberg was clear in his deposition testimony that distribution
through Windows was of "tremendous value" as a "customer acquisition
facility," especially since OLSs such as AT&T would be shipped with
"every copy of Windows." Silverberg, 1/13/99, at 689:16 - 691:9.
- CompuServe believed that inclusion in the Online Services Folder
represented a very large distribution opportunity -- one which no other
single hardware or software company could match without greater
expense. Warren-Boulton Dir. ¶ 101 (quoting Knott Dep., 2/20/98, at 21-23). In turn, according to Dr. Warren-Boulton, "Microsoft used this asset
to induce OLSs to enter into agreements that restricted the distribution
and promotion of competing browsers." Warren-Boulton Dir. ¶ 102.
- MCI was very interested in being included in the Microsoft Internet referral
server because it represented access to a large market of potential MCI
Internet subscribers. Microsoft stressed to MCI the value of inclusion, on
the reasoning that the referral server would ship with every copy of
Windows. Von Rump Dep. 1/13/99, at 322:2 - 323:15.
(4) AOL viewed promotion through Windows as
particularly valuable and would not have agreed
to Microsoft's exclusionary restrictions absent
placement in the Windows OLS Folder
234. AOL viewed obtaining promotion through Windows as particularly important
because of Microsoft's promotion on the Windows desktop of AOL's principal rival,
MSN, and the relatively low cost of acquiring potential subscribers through distribution
with Windows.
- Brad Chase testified that AOL's Steve Case "was very passionate about
the whole issue of MSN and AOL, and we believed he would be
passionate about trying to get into the Windows box as well." Chase,
2/11/99pm, at 74:24 - 75:23.
- David Colburn testified that, by bundling MSN with Windows, "Microsoft
was able to ensure that every consumer who purchased either a new
computer or a Windows 95 product at retail had MSN easily available."
Colburn argued that this placement gave Microsoft a "potentially decisive
strategic advantage". Colburn Dir. ¶ 15; Colburn, 10/28/98pm, at 52:3-8
(MSN's distribution opportunities on the Windows desktop were a "major
concern" for AOL).
- By contrast, other distribution channels used by AOL -- such as mailing
software directly to individual potential subscribers -- were more costly
and required "more effort by the consumer to access AOL" than was
required for consumers to access MSN, which was included with
Windows. Colburn Dir. ¶ 17. Distribution through Windows, in short, was
"uniquely effective." Colburn Dir. ¶ 18.
235. AOL thus believed that obtaining promotion through Windows would be
extremely valuable, a judgment in which Microsoft shared.
-
AOL executive Miles Gilburne told James Barksdale, at the time AOL
struck its deal with Microsoft, that the Microsoft deal was worth an extra
750,000 to a million subscribers per year and therefore the deal provided
a "very powerrful marketing opportunity" for AOL. Barksdale, 10/21/99am,
at 65:5-18.
- After the deal, Steve Case continued to believe that access to Windows
was valuable. He wrote that AOL should "move heaven and earth" to get
the best version of AOL integrated into Memphis (Windows 98), which
was a "huge" deal. Case directed AOL executives to approach this
project with a "jihad like focus." GX 441.
- During his cross examination, David Colburn calculated that the value of
AOL's placement in the Online Services Folder was, for the 1997-1998
time period alone, "far, far in addition" Colburn, 10/29/98am, at
12:1 - 13:21 (sealed session).
- A Microsoft summary of the operative terms of the AOL/Microsoft contract
notes that Windows distribution "has had a substantial benefit to AOL as a
large % of its subscriber growth has come from this source." GX 1127.
236. AOL would not have made Internet Explorer its standard browser, or
accepted Microsoft's other exclusionary terms, absent placement in the Online Services
Folder.
- Colburn testified that "AOL would not have been willing to negotiate a
browser license with Microsoft had Microsoft not indicated a willingness to
bundle and promote the AOL client software in some form with Windows.
Distribution and promotion on the Windows desktop was one of AOL's
goals - indeed, the most significant one - in negotiating a browser
agreement with Microsoft." Colburn Dir. ¶ 25. The value of distribution
with Windows was the "tell-tale part" of the deal, part (along with free
access to valuable technology) of a "powerful one-two-...three punch" that
AOL was unable to resist in its calculus of whether to do a browser deal
with Microsoft. Colburn believed that the negotiations changed when
Microsoft put the offer of distribution through Windows on the table
because it was a "value that Netscape could not really match". Colburn,
10/28/98pm, at 32:3-18; Colburn, 10/28/98pm, at 76:21 - 77:20
- Steve Case wrote in 1996 that the free valuable technology, coupled with
Microsoft's "distribution (OS) muscle" gave Netscape an "uphill struggle" in
negotiating a browser deal with AOL. DX 1342, at AOL M 0000190.
- James Barksdale testified that, after the Microsoft agreement, Steve Case
and David Colburn told him that AOL would not have entered into the
agreement with Microsoft but for access to the Windows desktop.
Barksdale Dir. ¶ 136.
- Brad Chase confirmed that AOL saw Windows distribution as crucial:
"Steve Case told Bill Gates it was important that AOL be included on the
Windows desktop if there was to be a partnership between the two
companies." Chase Dir. ¶ 43. During cross examination, Mr. Chase
again acknowledged the importance of Windows distribution to AOL by
testifying that "it would have been tougher" to get AOL to do a deal with
Microsoft without giving AOL access to Windows. Chase, 2/11/99pm, at
82:18 - 83:5.
- At the time of the deal, Microsoft recognized that distribution with
Windows was "almost an emotional thing with Case." GX 811.
c. Microsoft unsuccessfully attempted at trial to minimize
the value of distribution and promotion through
Windows
237. Microsoft witnesses argued at trial that value of promotion through
Windows was insubstantial to ISPs and OLSs, including AOL. Their testimony lacks
credibility and is unpersuasive.
237.1. Bill Gates' purported lack of understanding that Windows provided
a unique and valuable advertising vehicle (Gates Dep., played 12/15/98am, at 10:9 -
15:22) is not credible.
- Gates' testimony stands in sharp contrast to his position, at the
time Microsoft negotiated the AOL deal, that giving AOL placement
on the Windows desktop would be "putting a bullet through MSN's
head." Silverberg Dep., 1/13/99, at 703:13 - 704:19.
- Indeed, it was precisely the value of placement on the Windows
desktop that led Gates to insist that AOL receive placement "one
level below" MSN's. See supra Part V.D.3.b(2); ¶ 232.1; GX 346 (a
"slight advantage" for MSN was important to Gates).
- Gates' testimony also stands in sharp contrast to his and others'
contemporaneous writings. For example, in January 1996, Gates
told AOL that the Windows box was "sacrosanct", forcing AOL's
Case to find a solution in which AOL was not promoted "a la MSN".
GX 38. Gates agreed with Microsoft executives who stated that
Windows was a unique and valuable asset and were therefore
"vehemently" against proposals to allow service providers "access
to the Windows box". GX 130.
237.2. Dean Schmalensee's assertion that the "evidence shows the value
that Microsoft provided to AOL through placement in the OLS Folder was quite limited"
(Schmalensee Dir. ¶ 415), is wrong.
237.2.1. First, the evidence in fact shows that both AOL and
Microsoft believed that the placement AOL secured was very valuable, indeed, the
"fulcrum" of the deal.
- See supra Part V.D.3.b.(4); ¶¶ 232-236.
- Colburn testified that AOL's access to Windows 95 was
"more than a key part of the deal. It was the fulcrum of the
deal." Colburn, 10/29/98pm, at 34:8-11.
237.2.2. Second, Dean Schmalensee's calculation designed to
show the limited value of placement in the OLS folder to AOL is flawed.
- Based on the fact that AOL gets additional distribution
through OEMs, and his estimate of the amount AOL
paid for OEM promotion per subscriber obtained
through OEMs, Dean Schmalensee calculated that
the value to AOL of the placement in the Online
Services folder it obtained from Microsoft was, at
most, 8.3 million dollars in 1997 and 18 million dollars
in 1998. Schmalensee Dir. ¶ 426. Accordingly, Dean
Schmalensee reasoned, Microsoft did not incur " a
significant opportunity cost in giving AOL a place in
the OLS folder." Schmalensee Dir. E-19, ¶ 404.
- Dean Schmalensee's calculation fails to take account
of the fact that AOL's agreement with Microsoft
placed it in a stronger position to negotiate deals with
OEMs and, thus, that the amount AOL paid OEMs
would have been much greater but for AOL's
agreement with Microsoft. Colburn, 6/14/99pm, at
87:14 - 88:7 (testifying that AOL's agreements with
OEMs
- redacted -
(sealed session).
- Consistent with AOL's view, Professor Fisher
observed that AOL stopped paying OEMs nearly as
much money after its deal with Microsoft, which
suggests that the folder was valuable." Fisher,
1/7/99pm, at 14:9-14; see also Warren-Boulton,
11/30/98am, at 65:14 - 66:17 (AOL believes it is
valuable to have both placement in the OLS folder
and through OEMs).
- In addition, distribution with Windows allowed AOL to
gain placement on computers from smaller OEMs. In
an internal email, AOL noted that the company was
receiving "good registrations" from the Windows
placement, mostly from "the bottom 25% of the PC
clone market" with which "AOL will never have its own
contracts for distribution". GX 816.
237.2.3. Microsoft's argument that its guaranteed distribution
through Windows must not be all that valuable because AOL has numerous agreements
with OEMs for prominent placement (Chase Dir. ¶ 31; DX 2162) is additionally flawed
because it fails to recognize the enormous power that Microsoft itself has over the
OEMs.
- AOL believed that its contracts with OEMs hold the potential
to become endangered if Microsoft imposes further
restrictions on OEMs. As Colburn observed, "Microsoft had a
lot of power and clout with OEMs," including the ability to
"raise the ante of what would have to be spent" to secure
AOL's relationships with OEMs and the "ability to preclude"
AOL from the desktop altogether. Colburn, 10/28/98am, at
27:6 - 28:7.
d. Microsoft's assertion that its lacks monopoly power over
software distribution is immaterial
238. Microsoft witnesses argued that plaintiffs failed to show that Microsoft had
a monopoly over "software distribution" with which it was able to coerce ISPs and OLSs
to acquiesce to Microsoft's terms (Schmalensee Dir. ¶ 345; Myhrvold Dir. ¶ 82). But
whether Microsoft has a monopoly over "software distribution" does not matter.
238.1. Whether Microsoft had monopoly power over software distribution
has nothing to do with whether the exclusionary restrictions in the ISP and OLS
agreements were anticompetitive. They were anticompetitive because they served no
legitimate purpose and erected barriers to successful distribution of browsers by
Microsoft's rivals.
- Professor Fisher and Dr. Warren-Boulton testified that Microsoft's
agreements were predatory and anticompetitive because the
restrictions Microsoft extracted lack justification. Fisher, 6/1/99am,
at 60:15 - 62:2 (the restrictive provisions in Microsoft's ISPs
contracts are not profitable absent Microsoft's interest in maintaining
its operating system monopoly); Warren-Boulton Dir. ¶¶ 182-183
(testifying that Microsoft's restrictions on the ability of ISPs to
promote and distribute Internet browsers are unrelated to any
efficiency purpose and that any legitimate efficiency purpose could
be accomplished by substantially less restrictive means).
- The immense sums Microsoft bartered and spent to gain
preferential distribution for Internet Explorer can be explained only
as a predatory strategy to protect Microsoft's operating system
monopoly. See infra Part V.G.2; ¶¶ 299.4.
238.2. Microsoft needed, not monopoly power over OLSs and ISPs, but
only the ability to pay valuable consideration (like desktop placement and cash) to
induce the ISPs and OLSs to agree to these anticompetitive terms.
- Professor Fisher testified that whether or not Microsoft has
economic power over software distribution "has very little to do, if
any, with the case." Instead, he explained, the critical issue is
Microsoft's monopoly power in "the area of operating systems for
P.C.'s." Fisher 6/1/99am, at 24:17 - 25:17.
- Instead, Microsoft paid (rather than coerced with monopoly power
over software distribution) the ISPs to agree to exclusionary terms.
Dr. Warren-Boulton testified: Microsoft was aware of the value
OLSs placed on desktop placement (in the online service folder)
and "sought to exchange this valuable asset for exclusionary
restrictions." Warren-Boulton Dir. ¶ 102.
- David Colburn confirmed that Microsoft offered AOL value, including
promotion through Windows, which was "of immense value to AOL."
Colburn Dir ¶ 24.
238.3. To be sure, Microsoft used its operating system monopoly to
prevent OEMs from deleting either the Internet Connection Wizard or Online Services
Folder, or superceding them with auto-loading alternative shells, and thereby enhanced
the value of placement on the Windows desktop.
- See supra Part V.C.2.a(1); ¶¶ 206, 208.1.
e. Microsoft's contention that it simply offered ISPs
and OLSs a better product is erroneous and
misplaced
239. Microsoft witnesses argued that access providers agreed to favor Internet
Explorer and disfavor rivals merely because Microsoft "out competed" Netscape by
offering a better product than Netscape. Chase Dir. ¶ 136 (arguing that the "increasing
popularity of Internet Explorer is largely attributable to Microsoft's improvements in
technology"); Myhrvold Dir. ¶ 122 (testifying that Microsoft "succeeded because of the
work undertaken by our developers which resulted in the technical superiority of Internet
Explorer 3.0"); Myhrvold Dir. ¶ 126 ("ISPs ultimately started embracing Internet Explorer
because it met their needs better than Netscape's Web browsing software"). But this
argument is inconsistent with the evidence and ultimately beside the point.
239.1. First, the evidence shows that Internet Explorer was not superior
to Netscape at the time Microsoft extracted its exclusionary agreements and is not
clearly superior today.
239.1.1. AOL viewed both browsers as "comparable" and
understood, when it entered into the March 1996 browser contract with Microsoft, that
Netscape was prepared to create a browser that would have been "essentially
indistinguishable" from the componentized Internet Explorer.
- Steve Case told Bill Gates, in January 1996, when AOL first
began considering entering into a restrictive agreement with
Microsoft, that he viewed Internet Explorer "technically as
behind Netscape." GX 335
- David Colburn testified that, within AOL, both Netscape
Navigator and Internet Explorer were viewed as
"comparable." Colburn Dir. ¶ 33. Although Internet Explorer
was componentized, Netscape had a "robust browser that
had been tested in the marketplace," and had "more and
better features." Colburn Dir. ¶ 33. At the time of Microsoft
deal, Netscape was viewed as the technical leader by the
industry. Colburn, 10/28/99am, at 59:20 - 61:9.
- AOL understood that, in 1996, Netscape was contractually
committed to create, and prepared to create, a
componentized browser for AOL. Colburn Dir. ¶ 34; James
Barksdale told AOL that Netscape was prepared to do
"whatever they needed to do" to integrate their browser with
the AOL client. Colburn, 10/28/99pm, at 18:3 - 19:11;
Barksdale ,10/26/98am, at 57:9-25. Indeed, AOL believed
that by the time AOL would have been ready to use a
browser in its next software release, Internet Explorer and
Netscape Navigator would probably be "essentially
indistinguishable." Colburn Dir. ¶ 33.
- Netscape had made a commitment to meet AOL's delivery
schedule. Barksdale, 10/26/98am, at 58:2-3. As Mr.
Barksdale testified: "It's not rocket science to do it. We were
willing to do it, and knew how to do it, and offered to do it."
Barksdale, 10/26/98am, at 59:12-13. After AOL entered into
a virtually exclusive agreement with Microsoft, however,
Netscape had little incentive to rapidly develop a
componentized browser for AOL. Barksdale, 10/26/98am, at
67:22 - 68:11. Without a guarantee of at least some
distribution by AOL, quick development of a componentized
browser made little economic sense. DX 1733, at AOL M
0001025; Colburn, 10/28/98pm, at 63:17 - 64:6; Colburn,
10/29/98am, at 30:14 - 31:13; Colburn, 10/29/99pm, at
35:10-20.
239.1.2. After its agreement with Microsoft, it became even more
apparent that Internet Explorer was not materially superior to Netscape Navigator, as
AOL had continuing complaints about Internet Explorer.
- The time it took Microsoft to develop a componentized cross-platform browser was a problem for AOL. Colburn,
10/28/99am, at 60:15-61:9.
- In August 1997, AOL believed that the "IE4 browser is huge
and is tangled up with OS in Win98 product." Netscape
Navigator, on the other hand, was cited for having a "Much
Smaller Disk Footprint." GX 818.
- Following a technical meeting with Netscape, AOL noted that
Netscape's browser would take only about 4-6 months to
componentize, shipped with functional parity on all platforms,
and had lower memory requirements and more ease-of-use
features than Internet Explorer. GX 1150.
- Colburn testified that Internet Explorer hardly got a "ringing
endorsement" from AOL. Colburn, 10/28/99pm, at 21:16 -
22:10.
239.2. Second, Microsoft's argument that access providers chose Internet
Explorer because it was "better" than Netscape Navigator is inconsistent with the
restrictions on distributing other browsers that Microsoft imposed.
- Microsoft required access providers to restrict their distribution of
Netscape Navigator (rather than to merely promote Internet
Explorer). Fisher, 6/1/99am, at 66:18-25.
- Cameron Myhrvold admitted that the distribution restrictions resulted
from a fear that, if provided a "side-by-side" choice of Internet
Explorer and Netscape Navigator, users would choose Netscape
Navigator. Myhrvold, 2/10/99am, at 62:7 - 64:20.
239.3. Third, Microsoft's argument is also inconsistent with the evidence
that, absent promotion through Windows, AOL would not have accepted Microsoft's
exclusionary terms.
- See supra Part V.D.4.b(4); ¶¶ 251-255.
239.4. Dean Schmalensee's contention that AOL (and other OLSs and
ISPs) agreed to Microsoft's terms because the total aggregation of value Microsoft
offered -- including Internet Explorer -- was large (Schmalensee Dir. ¶¶ 421, 483) is
merely another way of saying that Microsoft spent a lot to obtain its exclusionary terms.
4. Microsoft's agreements have caused substantial competitive
harm
240. As it anticipated, Microsoft's agreements had, and continue to have, a
substantial exclusionary impact. Microsoft's restrictions prevented access providers
from meeting consumer demand by providing another browser (which would likely have
been pre-configured for the service) and made it difficult for users to locate and install
another browser. The result of Microsoft's exclusionary restrictions was substantially to
increase Internet Explorer's browser market share, diminish rivals' market share, and
facilitate Microsoft's maintenance of its operating system monopoly.
a. Microsoft's agreements raised rivals' costs
241. Microsoft's restrictive agreements substantially raised the costs to rivals of
obtaining and retaining browser market share.
241.1. Microsoft's requirement that ISPs and OLSs distribute and promote
only Internet Explorer (or mostly Internet Explorer) hindered browser rivals because
users, in particular novice users, tend to use the browser supplied by their access
provider.
- A 1996 ISP marketing update stated that ISPs are important to
Microsoft's "Internet mission" because most new users are first
exposed to the Internet through their ISP and if users get the IE
set-up to work with their ISP, they will be "less likely to switch to
Netscape or another browser later." GX 93.
- Even Dean Schmalensee conceded that "AOL customers and the
customers of online services use the browsing software provided by
their online services." Schmalensee, 1/19/99pm, at 62:20 - 63:6.
- In a testament to an ordinary user's lack of initiative in changing the
software as received from the original source, William Harris
testified that "... it is generally understood in the computer industry,
that consumers have a high proclivity to accept default settings and
configurations on software and computer-based services." Harris
Dir. ¶ 92.
- See infra Part VII.A.2.b; ¶¶ 366.
241.2. This is particularly true of AOL users.
241.2.1. A large segment of AOL users are novices who are
especially likely not to switch browsers once presented with the AOL client built on
Internet Explorer.
- GX 814A ("the typical AOL user is an Internet novice"); GX
1062, at page 2 (AOL study entitled "AOL Web Browser
Usability Test" concluding that "the most alarming fact
discovered in the Novice group is that most do not know the
difference between being on AOL and being on the Internet.
Those Novice users thought that once they signed on to
AOL, they had already accessed the Internet."); GX 415, at
MSV 10566 (only three percent of AOL's users in 1997
considered themselves "advanced" Web surfers, compared
to thirteen percent of Navigator's users and twenty-five
percent of Internet Explorer's).
- Colburn testified that users tend to use the software provided
by AOL. He believes that AOL's users use Internet Explorer
because AOL has "virtual exclusivity" with Microsoft.
Colburn, 10/28/99am, at 56:14-22.
- Colburn also stated that a user would have to be "technically
savvy" to understand the process for using Netscape
Navigator with the AOL service. Colburn, 10/28/99am, at
46:15 - 47:8.
- Brad Chase, upon announcing the AOL/Microsoft deal,
confirmed that users will not be "faced with an either/or
choice" of browsers. Instead Internet Explorer would be "the
standard choice" for all customers. Although AOL users are
permitted to download Navigator, this option will not be
displayed "in a prominent way" and "for all intents and
purposes... AOL will be moving its 5M customers to a new
client integrated with Internet Explorer 3." GX 180.
241.2.2. Moreover, after AOL signed its agreement with Microsoft
to "exclusively promote, market, and distribute" Internet Explorer, AOL's users were
"force fed" Internet Explorer. Even if a user continued using another browser, the user
was prompted to take Internet Explorer every time the user attempted to sign off AOL.
- The share of AOL subscribers with the latest version of
Internet Explorer installed on their machines rose to over
90% after AOL signed its agreement with Microsoft.
Microsoft executives attributed this rise to the fact that AOL --
in Microsoft's own words -- "force fed" Internet Explorer to
AOL users by automatically distributing Internet Explorer to
users each time a user attempted to log off AOL. AOL also
used what Microsoft called the "deadline approach," in which
users who had an older browser version could not enter the
AOL service unless they downloaded the latest browser
version or upgraded using CD sent in the mail. GX 814A.
- In December 1996, less than nine months after AOL signed
its agreement with Microsoft, Bill Gates wrote that Microsoft
had little incentive to negotiate with AOL to include the OLS
as a default channel on the Windows desktop because "we
are getting all their users for IE through other efforts." GX
346.
241.3. Even for users inclined to try-out another browser, Microsoft's
restrictions raised rivals' costs because users are unlikely to bear the (largely non-monetary) costs of successfully obtaining a browser through other channels and then
attempting to configure it for their ISPs/OLSs service.
- David Colburn testified that getting another browser for AOL
requires some "complex steps" and "technically savvy" users.
Colburn, 10/28/99am, at 46:15 - 47:7. Colburn believes that it is
"difficult" for AOL's customers to obtain and use Netscape
Navigator. Colburn, 10/29/99pm, at 66:7 - 67:6.
- Cameron Myhrvold testified, regarding Southwestern Bell's (an ISP
that does not have a restrictive contract with Microsoft) offering of
Netscape through its web site, that, although obtaining and installing
Internet Explorer through SBC's web site might be "technically
possible," it is not "very attractive to have to go in and manually
configure it and then be followed by a note that says you're not
going to get any technical support." Myhrvold, 2/10/99pm, at 81:5 -
82:8.
- Myhrvold also testified, in referring to Internet Explorer's difficulties
before Microsoft required its distribution as the default browser, that
"in many cases, even if a user had acquired Internet Explorer on his
own, he would not receive any help from the ISP in configuring
Internet Explorer for the ISP's service." Myhrvold Dir. ¶ 26.
- See also infra Part VII.A.2.b; ¶¶ 366.2-.4.
b. Microsoft's contracts substantially excluded rival web
browsers
242. The impact of Microsoft's efforts to raise rivals' costs was to garner
substantial browser market share at rivals' expense. Microsoft's internal documents
and the testimony of its witnesses, the AdKnowledge data, and Internet Explorer's
comparative lack of success in channels where Microsoft has not secured exclusionary
agreements -- all prove the exclusionary impact of Microsoft's agreements.
(1) Microsoft's internal analyses evidence the impact
of its restrictions
243. Microsoft's internal documents show that obtaining preferential distribution
through ISPs and OLSs had a significant impact on Internet Explorer's usage and
substantially increased Internet Explorer's market share.
- Microsoft concluded that its agreement with AOL (and its CompuServe
subsidiary) alone tied up 65% of the subscribers considered to be in the
"Top 80" access providers by the end of 1997. GX 835, at MS98 0112834
(cited in Fisher Dir. ¶ 216).
- Microsoft further reported in January 1998 that "IE share" on AOL
GX 424, at MS7 000591 (sealed); This is consistent with AOL's own
estimate that, as of January 1998, the "current share" of Internet Explorer
on AOL was "90+". Microsoft executives commented that such a high
browser share would "really change the way we work with AOL; there are
few users left to upgrade, so we don't need to keep beating them up about
this." GX 814A.
- As of December 1997, Microsoft estimated that Internet Explorer had a
"run rate" of 76% of the referral server ISPs. GX 425, at MS98 0102442.
During this same period, another Microsoft document stated that 10 of the
top 12 ISPs ship IE4 today" and "63% overall ship IE default". GX 1063, at
10.
- By January 28, 1998, Microsoft believed that Internet Explorer's share
among the top ten ISP/OLSs (GX 427, at MS98 0116511) (sealed)
and that 85 of the top 100 access providers shipped Internet Explorer as
their preferred or exclusive browser. GX 420, at MS98 0113045.
- Professor Fisher summarized Microsoft's own calculation of the impact of
Microsoft's restrictions on ISPs: "According to a Microsoft document, at year
end 1997 Microsoft enjoyed a 94 percent weighted average share of browser
shipments by ISPs who agreed to make IE their default browser, compared with a
14 percent weighted average share of browser shipments by ISPs who did not
make IE their default browser. Microsoft's weighted average share of browser
usage by subscribers to ISPs who made IE their default browser was over 60
percent; Microsoft's weighted average share of browser usage by subscribers to
ISPs who did not make IE their default was less than 20 percent." Fisher Dir. ¶
224 (referring to GX 366); see also GX 11.
(2) The exclusionary impact of Microsoft's
agreements is confirmed by the AdKnowledge
data
244. Data collected by a company called AdKnoweldge confirms that Microsoft's
exclusionary agreements with ISPs and OLSs have had a substantial impact on Internet
Explorer's market share.
245. AdKnowledge collects hit data, which measures the intensity with which a
particular browser is used and is the most relevant metric of market share in this case.
245.1. AdKnowledge collects "hit" data and measures intensity of use.
- Adknowledge is a company that markets web advertising
management services and as part of that service uses a set of
servers that delivers web page advertisements when users request
particular webpages. Warren-Boulton Dir. ¶ 145; Fisher Dir. ¶ 225
- As part of its ordinary activities, Adknowledge collects information
on which browsers call up a particular web page containing banner
ads "served" by AdKnowledge. In the language of the industry,
Adknowledge tracks the number of "ads served" and its data is
commonly referred to as "hit data." Gildor Dep., 10/6/98, at 31:11 -
32:8 (DX 2569); Warren-Boulton Dir. ¶145; Fisher Dir. ¶ 225.
- As part of tracking the hit data, AdKnowledge records information on
the type of browser being used and on the user's "domain name,"
which in certain cases can be used to determine the user's ISP.
Warren-Boulton Dir. ¶ 145; Fisher Dir. ¶ 225.
- Dean Schmalensee conceded that hit data, such as the
AdKnowledge data, measures the intensity with which a particular
browser is used. Schmalensee Dir. App. ¶ 44.
245.2. Intensity of use, as will be explained, is the most appropriate
measure of market share in this case.
- See infra Part VII.A; ¶¶ 360.1-.2.
- Professor Fisher testified that it is appropriate to measure browser
share as ISVs would assess it for the purposes of determining to
which platforms to develop applications. Fisher, 6/1/99pm, at 20:13
- 22:8.
246. The AdKnowledge data, as Professor Fisher and Dr. Warren-Boulton
testified, show a broad increase in the usage of Internet Explorer, and decline in usage
of Netscape, over the time period in which Microsoft engaged in its exclusionary
practices.
- The AdKnowledge data show that Internet Explorer's overall market
share increased from approximately 20% in January 1997 to 49%
by August 1998. See infra Part VII.A.3; ¶ 369.1.1.
- The AdKnolwedge data show that Netscape's usage share declined
from 77% to 48% over the same period. See infra Part VII.A.3; ¶
369.1.1.
247. The AdKnowledge data also demonstrate the substantial impact of
Microsoft's exclusionary agreements. Plaintiffs' economists estimated this impact by
comparing the share of browsers used by subscribers of access providers that signed
restrictive agreements with Microsoft (such as AOL) with the share of browsers used by
subscribers of access providers that did not sign restrictive agreements.
247.1. Plaintiffs' economists compared several categories of ISPs that
were parties to agreements with Microsoft or Netscape which required varying degrees
of preferential treatment for Internet Explorer or Netscape, against a control group
consisting of a set of ISPs that had no contractual obligations to either Microsoft or other
browser manufacturers:
- Plaintiffs' economists assembled data for several categories of
ISPs. The categories included
(1) AOL (along with its subsidiary,
CompuServe);
(2) a category Microsoft described as "IE Preferred";
(3) a category consisting of all hits recorded by AdKnowledge
including all hits from ISPs, OLSs, and other firms that provide
Internet access;
(4) a category of ISPs whose shipments of rival
browsers was contractually limited by Microsoft to a certain
percentage of overall shipments ("shipment restrictions"); and
(5) a
category consisting of "Netscape Partners," ISPs (principally the
Regional Bell Operating Companies called the RBOCs) that have
granted Netscape certain preferences.
Warren-Boulton, 12/1/98pm,
at 17:17 - 18:7.
- These categories were compared against a control group of ISPs
that were not subject to contractual restrictions. The control group
is called in Microsoft's documents "IE parity" and is distinguished in
those documents from two other groups that Microsoft describes as
"IE Preferred" and "Netscape Preferred." GX 835, at MS98
0112826.
247.2. Through two sets of comparisons, plaintiffs' economists
demonstrated that Microsoft's ISP and OLS agreements had a substantial exclusionary
impact.
247.2.1. First, Professor Fisher and Dr. Warren-Boulton
demonstrated that Internet Explorer's overall share rose far more sharply -- by
approximately 20% -- than Internet Explorer's share among users of ISPs that were not
subject to Microsoft's contractual restrictions.
- Internet Explorer's overall share of browsers rose from 20%
in January 1997 to 49% in August 1998, while Netscape's
share fell from 77% to 48%. GX 4, GX 5, GX 1445, GX 1480;
Warren- Boulton Dir. ¶ 146; Fisher Dir. ¶ 228. By contrast,
Internet Explorer's share of usage among subscribers of the
control group ISPs rose only from 20% to 30% over the same
period. GX 3; GX 5; GX 1445; GX 1480; Warren Boulton Dir.
¶¶ 144, 149; Fisher Dir. ¶ 228.
- This difference led Dr. Warren-Boulton to conclude that, "[i]f
no ISPs had been party to exclusionary agreements,
assuming no other changes, it is reasonable to expect that
IE's market share would be its share of customers of these
unconstrained ISPs -- approximately 30 percent in August
1998. The differences between that share and IE's actual
overall share -- 49 percent -- shows the impact on the market
of the Microsoft ISP agreements." Warren-Boulton Dir. ¶
150; see also id. ¶ 151 (explaining that the dramatic
difference between AOL's share and overall share cannot be
explained by factors other than the contractual restrictions).
247.2.2. Second, Professor Fisher and Dr. Warren-Boulton showed
that Internet Explorer's share of usage among ISP users increased as the ISP's
contractual obligations to favor Internet Explorer became more severe. This is precisely,
they explained, what one would expect if Microsoft's agreements had exclusionary
effects.
- Dr. Warren-Boulton prepared a chart based on the
AdKnowledge data showing the change in Internet Explorer's
share among the various categories of ISPs (the 5 listed
above) over the same time period. GX 1318. That bar graph
shows that, the more exclusionary the terms of Microsoft's
contracts, the greater the increase in Internet Explorer's
share. GX 1318.
- Dr. Warren-Boulton testified that the striking correlation
between Internet Explorer usage share and the degree of
contractual restrictions on promoting and distributing browser
rivals is precisely what one would expect if the agreements
had a substantial exclusionary impact. Warren- Boulton,
12/1/98pm, at 17:17 - 18:7.
- Professor Fisher also demonstrated, in results he described
as "striking," that more severe contractual restrictions
correlated with higher Internet Explorer share. Fisher Dir. ¶
228. Professor Fisher undertook a comparison similar to Dr.
Warren Boulton's, but using only three categories of ISPs: (1)
AOL/CompuServe; (2) all ISPs; and (3) the "parity" control
group. GX 4. Professor Fisher's analysis showed that, while
Internet Explorer's share of the Internet Explorer Parity
"control group" rose less than 10% in the relevant period,
Internet Explorer's share of the "All ISP" group increased
nearly 30%. The share increase in the Internet Explorer
parity group, as Professor Fisher testified, includes changes
in share due to Internet Explorer's increased quality (and
other factors, such as Microsoft's other exclusionary
conduct). Accordingly, the 20% difference between the "All
ISP" group and the control group reflects the impact of
Microsoft's restrictive agreements on Internet Explorer's
overall market share. Fisher Dir. ¶¶ 227-228; GX 4; GX
1445.
247.3. The AdKnowledge data also show that the impact of Microsoft's
exclusionary contract with AOL was especially severe.
- Internet Explorer's share of AOL -- which had the most restrictive
agreement with Microsoft -- rose 60% (from 25% in the first three
months of 1997 to 85% in June, July, and August of 1998), while
Internet Explorer's share of usage on ISPs that were contractually
neutral rose less than 10%. GX 1318.
- While Internet Explorer's share of the parity group increased only
10% from January 1997 to August 1998, Internet Explorer's share of
AOL and CompuServe users rose from 22% to 87% during the
same period, and its share for all ISPs rose from 22% to 49%.
Fisher Dir. ¶ 228; GX 4; Warren Boulton Dir. ¶ 148.
247.4. The degree of anticompetitive impact Professor Fisher and Dr.
Warren-Boulton found based on their analyses of the AdKnowledge data is, if anything,
conservative because the control group itself consists of ISPs affected by Microsoft's
other predatory and anticompetitive conduct, such as its bundling of Internet Explorer
and predatory pricing.
- Dr. Warren-Boulton testified: "Moreover, differences between the IE
parity group and the other groups, if anything, understate the
exclusionary impact of Microsoft's practices because the IE parity
itself may have been affected by Microsoft's exclusionary conduct."
Warren-Boulton Dir. ¶ 151; see also Warren-Boulton, 12/1/99pm, at
38:1-6 (explaining that the control group may have been affected by
several of Microsoft's anticompetitive practices, such as the OEM
tying).
- Professor Fisher also testified that caching leads to an
"understatement of the effects of Microsoft's restrictive practices."
Fisher Dir. ¶ 226, n.6.
(3) The exclusionary impact of Microsoft's
agreements is confirmed by Internet Explorer's
comparative lack of success in other channels
248. Internet Explorer has not fared as well in channels that are not subject to
Microsoft's exclusionary agreements or control. As partly illustrated by the
AdKnowledge data, Internet Explorer has a lower share in unconstrained distribution
channels.
- Microsoft itself refers to the unconstrained channels as "demand driven."
GX 807.
- Cameron Myhrvold conceded that Internet Explorer does not do as well in
the retail channel, where Microsoft does not have restrictive agreements,
as it does when distributed through ISPs and OEMs. Myhrvold,
2/10/99am, at 32:8-21.
- Microsoft repeatedly pointed out during the trial that Netscape has a higher
market share among corporate and educational customers. Defendant's
Opening, 10/20/98am, at 29:25 - 30:12. Dean Schmalensee cited a Zona
research study of the browser market (DX 60) to argue that "Netscape
continues to have a large share of corporate users." Schmalensee Dir. ¶
538. A 1997 marketing memo from Brad Chase wrote that "We have not
done a very good job with this segment...Netscape still has this lead on
us." GX 512, at MS7 004152.
- James Barksdale agreed that Netscape has a relatively large share of
corporate and enterprise users and testified that: "That proves the point I
am making... where we have more access to the market, we are doing
much better than where we have been estopped from half of the
distribution channels... When we get to compete head to head, we do
pretty good." Barksdale, 10/26/98pm, at 45:6 - 47:19.
- Users of Internet browsing software downloaded Netscape's browser
almost 2.5 times as often as Internet Explorer in both the first and third
quarters of 1998. GX 1845; GX 1846 (estimating the download figures for
Netscape and Microsoft at 6.7 million and 2.7 million respectively in 1Q98
and 6.7 million and 2.8 million in 3Q98).
- See also infra Part VII.A.
c. Microsoft's arguments that its ISP and OLS agreements did not
have a significant exclusionary impact are belied by the
evidence
249. Microsoft witnesses advanced various arguments to the effect that its ISP
and OLS agreements did not substantially affect browser market share. The arguments
are unsound and, in many instances, incredible.
(1) Microsoft's restrictions were not ineffective
250. The testimony of Microsoft's witnesses that its restrictions were benign is at
odds with the facts and misleading.
250.1. First, contrary to the contention of its witnesses, Microsoft's
contractual restrictions had a significant impact on the ISPs and OLSs that were parties
to them.
250.1.1. Cameron Myhrvold asserted that, under their agreements
with Microsoft, ISPs and OLSs could "provide customers with whatever Web browsing
software the customer requested" and were "never required to distribute IE to any
specified percentage" of users. Myhrvold Dir. ¶ 5 (emphasis in original). But he could
not defend this assertion.
- Myhrvold admitted (as is clearly spelled out in the plain
language of the contracts themselves) that if, an ISP fell
below the shipment percentage, Microsoft had the
contractual right to remove it from the Internet referral server.
Myhrvold, 2/10/99am, at 51:11 - 52:21; 53:14-18.
- See also GX 1144, at MS6 5001130 (Microsoft and Spry, Inc.
Internet Sign-Up Wizard Referral Agreement, § 3.1) (sealed);
GX 1146, at MS6 5000924 (Microsoft and Mindspring
Internet Sign-Up Wizard Referral and Microsoft Internet
Explorer License and Distribution Agreement, § 3.1) (sealed);
GX 1213, at MS6 5000388, -389 (Microsoft and AT&T
Promotion and Distribution Agreement, §§ 3.3 and 4.1)
(sealed); GX 804,at AOL 0001735, -738, -740 (Microsoft and
AOL License and Marketing Agreement, §§ 6.1, 7.2, and
7.4).
250.1.2. Similarly, Brad Chase misrepresented the terms of
Microsoft's contract with AOL when he testified that "AOL has always been free to
provide non-Microsoft browsing software to subscribers who request it." Chase ¶ 73.
This testimony is at odds with the explicit terms of Microsoft's contract with AOL.
- Chase himself admitted that Microsoft's agreements "limit
the ability of the OLSs to promote and distribute non-Microsoft Web browsing software." Chase Dir. ¶ 98. And he
wrote at the time of the contract that the exceptions allowing
AOL to use another browser were "pretty remote." GX 180.
- AOL (like other ISPs/OLSs) had to abide by shipment
restrictions that prohibited it from distributing other browsers
more than a certain percentage. Fisher, 6/1/99am, at 65:24 -
66:25 (testifying that Microsoft does not merely require ISPs
to ship Internet Explorer to a certain percentage of its
subscribers; instead, Microsoft prevents the ISPs from
shipping more than a certain percentage of non-Microsoft
browsers to its customers).
- Colburn testified that, during the negotiations (in which Chase
was involved), Microsoft "attempted to secure exclusive
distribution and promotion for Internet Explorer, with no or
few exceptions for distribution or promotion of a competitive
browser . . . . Microsoft obtained virtual exclusivity for its
browser on AOL, preventing AOL from providing any
significant promotion or distribution of Netscape's Navigator
browser." Colburn Dir. ¶ 28. In addition, among other
restrictions, AOL's contract with Microsoft prohibited it from
volunteering information to its members about how to
download Netscape Navigator. Chase, 2/11/99pm, at 46:11 -
47:6; 56:21 - 57:9.
250.2. Second, contrary to the assertion of Microsoft's witnesses (Chase
Dir. ¶ 167), downloading is not an efficient channel of distribution for web browsers.
250.2.1. The video tape that Brad Chase sponsored to show the
supposed ease with which users may download Netscape Navigator from AOL (DX
2162) does not accurately represent the experience of an end user attempting to obtain,
install, and use Netscape Navigator with AOL.
250.2.1.1. First, Chase's video skipped the entire browser
installation process.
- Compare GX 1665 (a videotape prepared by the
plaintiffs demonstrating the download and installation
steps that Chase's videotape skipped) with Chase,
2/11/99am, at 25:24 - 26:3.
261.2.1.1.1. That process requires additional complex
steps, knowledge, and a significant amount of time, during which, as Chase admits,
many things may occur to cause the user to lose his connection.
- Chase testified that there are any number of
problems that one may encounter while
downloading a browser, with disconnection of
the phone line the most common. Chase,
2/16/99am, at 37:9 - 38:3.
261.2.1.1.2. Brad Chase's video omits the following
steps and ignores the following problems:
- After the user clicks on the button to make the
"Download Manager" screen appear on AOL,
there are no instructions telling the user what to
do next. Chase, 2/16/99am, at 31:8-14.
- And after the download process is complete
and the user clicks "Okay" to return to the AOL
service, Netscape Navigator does not appear
anywhere on the screen. Nor do appear any
instructions. The user must know that the next
step is to find the downloaded file and execute
it. Chase, 2/16/99am, at 35:25 - 36:12.
- In order to find the download file, a user must
exit AOL altogether, "minimize" the AOL service
screen, or "navigate to the explorer and explore
through the files to go to the download folder."
Chase, 2/16/99am, at 39:14 - 41:1. There are
no instructions telling the user what to do, or
even that anything additional needs to be done.
Chase, 2/16/99am, at 39:14 - 41:1.
- Once the user locates the "My Computer" file,
the user must find the file setup.ex. The user
has to remember where AOL put this file in the
earlier download manager screen,
approximately 45 minutes earlier. Again, there
are no instructions for the user to follow.
Chase, 2/16/99am, at 42:6 - 43:11; GX 1665.
- In short, the plaintiffs' videotape confirmed, as
Chase himself wrote in an internal email, the
setup process is "too hard for users to figure
out. Only a little more than half of the people
that download active set-up end up installing
the browser. I think they don't figure out what
to do once they download the set-up stub." GX
214.
250.2.1.2. Second, for the video, Microsoft used a high
speed internal corporate connection. This type of connection is used only for "new
corporate installations" and not by home users, which represent AOL's customer base.
With an Internet connection typically used by home users, the download process takes
significantly longer than depicted in Chase's video.
- Chase himself concedes that "not many" users have
this type of connection. Chase, 2/11/99am, at 26:21 -
27:11.
- The download time (not counting the installation
process) alone took between 30 and 60 minutes
during the three trials made by the plaintiffs' expert in
filming GX 1665. Chase, 2/16/99am, at 34:7-10.
- US West estimated that it would take an average of
"45 minutes" for residential customers to download
Netscape Communicator. Bozich concluded that it
"took a long time." Bozich Dep., 1/13/99, at 122:9 -
123:22.
250.2.1.3. Third, Chase acknowledged that AOL is not
permitted by Microsoft's contract to put a text message on the screen that advises users
that they may download Netscape Navigator. AOL is not even permitted to tell users
what key words to type into the AOL search function to locate the download site for
Navigator.
- Chase conceded that Microsoft's contract with AOL
did not allow AOL to advise users that they may
download Navigator. For instance the Microsoft does
not allow AOL to put a text message on the screen
telling users how to download Netscape's browser.
Microsoft had limitations on "how much AOL could
promote Netscape Navigator within their service."
Chase, 2/16/99am, at 29:14 - 30:13.
- Colburn testified that "we were greatly restricted under
the Microsoft contract as to where we could advertise
Navigator, downloads for them, whatever, and so
there was relatively little place for them to get
promotion." Colburn, 10/29/98pm, at 36:7-11.
250.2.2. Mr. Chase confirmed the complexities involved in this
entire process of downloading and installing Netscape Navigator from AOL when he told
the Court that the explanation of that process gets "complicated." That conclusion is
supported by other evidence, in addition to plaintiffs' videotape:
- Responding to a court's question relating to an AOL user
downloading Netscape Navigator, Chase testified, "This gets
a little complicated, your honor. Let my try to explain.
There's a series of processes that are involved here." Chase,
2/16/99am, at 25:15-17.
- Mr. Myhrvold's videotape makes the point that getting a
browser that is not preinstalled on the computer requires both
time to acquire the browser and begin the installation and
"the time and effort and knowledge needed to run the setup
program, which, for a large number of users, would actually
be cumbersome and not straightforward." DX 2166;
Myhrvold, 2/9/99pm, at 22:13 - 23:2.
- Chase agreed that, for some users, it would be "cumbersome
and not straightforward to try to install the browser
themselves" Chase, 2/11/99pm, at 14:7 - 16:21.
(2) Microsoft's agreements frustrated access
providers' desire to offer customers a choice of
browsers
251. Microsoft's witnesses argue that its agreements did not have a substantial
impact because many ISPs want to offer only a single browser. (Schmalensee Dir. ¶¶
407, 435, 436). But this argument, too, is inconsistent with the evidence.
251.1. ISPs and OLSs wanted to offer a choice of browsers in order to
meet consumer demand.
- Cameron Myhrvold testified that most ISPs support both Netscape
Navigator and Internet Explorer because "that is what their
customers demand." Myhrvold Dir. ¶ 17. ISPs, Myhrvold agreed,
"generally like to give their subscribers a choice of browsers."
Myhrvold, 2/9/99pm, at 72:12-14.
- According to an internal e-mail from Myhrvold to, among others,
Brad Chase and Joachim Kempin, Myhrvold wrote it was "damn
hard" for Microsoft to get ISPs to favor Internet Explorer because
"ISPs are agnostic on the browser" and "it's against their nature to
favor a browser." He elaborated: "I have had a hard time guiding
the ISPs to IE loyalty even when I made them sign explicit terms
and conditions in a legal contract." GX 440.
- Myhrvold further testified that Microsoft imposed its shipment
restrictions -- which prohibited ISPs and OLSs from offering other
browsers, even to customers that requested them, if total shipments
of other browsers exceeded a certain level -- precisely because
Microsoft was afraid that access providers would offers users a
choice. If users had a "side-by-side" choice of browsers on their
merits, Microsoft believed, users would be more likely to choose
Netscape. Myhrvold, 2/10/99am, at 62:7 - 64:20.
251.2. AOL wanted -- but was not allowed -- the flexibility to offer its
users a choice of browsers.
- David Colburn testified that AOL wanted the flexibility to be able to
integrate two different browsers into its client software, and thus
provide its users with an easily accessible choice of browsers.
Colburn Dir. ¶ 28 ("It was AOL's objective to have both Navigator
and Internet Explorer available to its members, allowing them to
choose which browser to use."); Colburn Dir. ¶ 26 ("AOL wanted
the flexibility of being able to integrate different browsers into its
client software,"); Colburn, 10/28/98pm, at 67:23 - 69:12 (AOL could
have given Netscape and Microsoft's browsers "coequal" positioning
on the AOL client and thus offered consumer choice).
- This is precisely why, as Colburn testified, AOL fought
(unsuccessfully) not to be subjected to the restrictions. Colburn,
10/29/98pm, at 46:18 -49:18.
251.3. Brad Chase's contrary testimony -- that AOL wanted to distribute
only one browser to its users (Chase Dir. ¶ 37) -- is itself contradictory and not credible:
- Chase said (at trial) that AOL wanted to integrate only one browser
technology with the AOL client. Chase, 2/17/99am, at 55:8 - 58:16.
But, when confronted with Colburn's testimony that "AOL wanted
the flexibility of being able to integrate different browsers into its
client software," Chase argued -- falsely -- that AOL's contract with
Microsoft permitted it to have another "AOL client alternative" with
another browser if it wanted. Chase, 2/17/99am, at 58:23 - 59:11.
- Although Mr. Chase testified at trial that AOL "always wanted to
have one primary technology," Chase, 2/17/99am, at 60:7-22, he
testified in deposition as follows: "I recall AOL wanted flexibility. I
don't recall whether they wanted to ship both browsers, or make
available both browsers, or have one and not the other. I don't
recall." Chase, 2/17/99am, at 61:18-21 (quoting Chase Dep.,
3/25/98, at 180:17-20).
- At trial, Chase said that his deposition testimony did not refer to
AOL actually incorporating different browser technologies into
different versions of its software. Chase, 2/17/99am, at 62:2-15. In
his deposition, however, Chase testified that he would "agree that
[AOL] wanted to have the flexibility" and that he didn't "know what
they really wanted to do," Chase, 2/17/99am, at 62:16 - 63:18
(quoting Chase Dep., 9/29/98, at 185:4-8), and that he wasn't "really
sure" whether AOL would have "entered into an agreement with
both Microsoft and Netscape to incorporate their browser
technology in different versions of AOL's client software." Chase
2/17/99am, at 63:20 - 64:13 (quoting Chase Dep., 9/29/98, at
185:23 - 186:5).
- On re-direct, Chase argued that he was testifying about whether
AOL wanted to create one client technology and put both browsers
in the same client. Chase, 2/17/99am, at 68:3 - 69:23. But, at his
deposition, he was asked whether AOL wanted to incorporate
Internet Explorer and Netscape "in different versions of AOL's client
software" and answered: "I certainly - I can't say definitively."
Chase, 2/17/99am, at 70:24 - 72:6 (quoting Chase. Dep, 9/29/98, at
185:23 - 186:16).
251.4. Other access providers also wanted to give users a choice of
browsers; many of them nonetheless agreed to strict restrictions on their ability to satisfy
their customers' demands in order to obtain desktop placement or other consideration
from Microsoft.
- CompuServe agreed to Microsoft's restrictions even though it
preferred "to have flexibility in the software" that it uses. Warren-Boulton Dir. ¶ 111 (citing Knott Dep., 2/20/98, at 24:24 - 25:5).
- MCI also suffered from not being able to provide browser choice:
"There are certainly users out there that prefer browsers and e-mail
clients that are not Microsoft. And our ability to reach them and
entice them to sign up for our service is presumably enhanced by
the ability to promote and distribute those." Warren-Boulton Dir. ¶
111 (quoting Von Rump Dep., 4/28/98, at 16:25 - 17:4). Because
Netscape Navigator was the more popular browser, MCI tried to
negotiate less stringent restrictions with Microsoft. Von Rump Dep.,
1/13/99, at 327:19 - 328:6. MCI was, however, unsuccessful. Von
Rump Dep., 1/13/99, at 326:13 - 327:10.
- Robert Beran, head of Bell Atlantic's ISP service (BAIS), testified
that BAIS chose to sign an agreement for promotion through
Netscape' referral server rather than Microsoft's because the
restrictions on which Microsoft insisted -- exclusivity for Internet
Explorer -- were too onerous. The exclusionary provisions in the
Microsoft agreement did not allow Bell Atlantic to meet its objective
of letting "the customer choose which browser they wanted to use"
which Bell Atlantic viewed as part of its "job to provide customers
with access to the leading browsers". Beran Dep., 1/13/99, 117:19 -
120:15; see also Beran Dep., 1/13/99, at 117:2 - 119:12 (BAIS did
not want an exclusive with Microsoft); Bozich Dep., 1/13/99, at
121:22 - 122:7 (US West did not want to be prohibited from offering
customers a choice); Rys Dep., 1/13/99, at 175:2 - 176:14
(Ameritech wanted to offer choice).
(3) Microsoft witnesses' testimony that its ISP and
OLS agreements did not have an exclusionary
impact is unreliable
252. The AdKnowledge data and Microsoft's own documents, among other
evidence, demonstrate that Microsoft's agreements had a substantial exclusionary
impact. By contrast, the figures cited by Cameron Myhrvold in an attempt to show a
purported lack of anticompetitive effect (Myhrvold Dir. ¶¶ 65-80) are incomplete and
unreliable.
- Myrhvold was unable to give an estimate -- other than to "hazard a guess"
-- of the percentage of shipments by referral server ISPs accounted for by
Internet Explorer in 1997. Myhrvold, 2/10/99am, at 41:4-19. While
conceding that it was difficult for Microsoft to gather distribution data from
the ISPs (Myhrvold, 2/10/99pm, at 47:4-25), Myhrvold nonetheless relied
upon the sparse data available in arguing that Microsoft's contracts did not
inhibit the distribution of other browsers. Myhrvold Dir. ¶¶ 65 -78.
- Myhrvold's figures are also flawed because they are based on ISP
reporting forms that may undercount the distributions of Internet Explorer.
As Mr. Myhrvold conceded, it is impossible to tell whether the reports
counted all copies of Internet Explorer; he does not even know why the
forms contain different reporting standards for Internet Explorer and
Netscape Navigator. Myhrvold, 2/10/99am, at 60:9 - 62:6.
- Nor do Mr. Myhrvold's figures take into account that the ISPs included as
having shipped large numbers of Netscape Navigator immediately after
entering into an arrangement with Microsoft might have been exhausting
old inventory, Myhrvold Dir. ¶ 70 (citing large numbers of Netscape
Navigator distributed by Concentric during the the first three quarters of
1997), even though Myhrvold concedes that the reported ISPs were
exhausting inventories of Navigator. Myhrvold, 2/10/99am, at 54:13 - 55:4
(Microsoft did not expect companies to instantly meet the shipment
percentages because "every company in the Internet referral server took
some time to ramp up to Internet Explorer." Therefore Concentric was not
alone in falling short of its shipment percentages while it was shifting over
to Internet Explorer.).
- Myhrvold also neglected to include, in the figures supposedly illustrating a
lack of foreclosure, some of the most recent data that Microsoft has.
Myhrvold, 2/10/99am, at 59:11 - 60:8. By contrast to the implication in
Myhrvold's testimony (Myhrvold Dir. ¶¶ 70-71), Concentric's shipments of
Internet Explorer are increasing precipitously; and, in March 1998,
Concentric reported a net return (or a negative distribution) of non-Microsoft browsers. GX 1798.
- Nor did Myhrvold mention that the most recent forecast for Earthlink in
Microsoft's possession predicted Internet Explorer shipments to be 80.6%
of overall shipments from October 1997 - December 1997. Myhrvold Dir. ¶
76 (not citing the most recent Earthlink browser shipment figures); GX
1789 (most recent Earthlink predictions of Internet Explorer shipments).
He also failed to mention that the earlier figures on non-Microsoft browser
shipments by Earthlink were a result of a special arrangement with
Microsoft; an internal Microsoft email giving background on Microsoft's
referral server program states: "The Earthlink deal was special because we
needed to get a big ISP to sign up for the IE Referral Server Program and
Earthlink was the first to sign." GX 228, at MS98 0113059.
- Another stark example of Microsoft's courthouse estimates differing from
its own internal tracking documents can be found in the Netcom numbers:
According to Microsoft's internal documents, Netcom's Internet Explorer
shipments as of the FY 1998 Mid Year Review were 40%. GX 366. A
contemporaneous Microsoft email confirms that Netcom was shipping
Internet Explorer to approximately 40% of its customers. GX 228. By
contrast, Microsoft's filings with the Court show, Netcom's Internet Explorer
shipments through January of 1997 as only 1.2%. Myhrvold Dir. ¶ 66.
253. Dean Schmalensee also opined that Microsoft's agreements did not have a
substantial exclusionary effect. His analysis is incorrect because, among other factors, it
is based on flawed data from Market Decisions Corporation ("MDC").
- See infra Part VII.A.5.6; ¶¶ 373-379.2.2.
(4) Microsoft's failure to enforce certain restrictions,
and its partial waiver of them on the eve of this
litigation, do not eliminate the agreements'
anticompetitive effects
254. Microsoft witnesses also argued that its restrictions did not have a
significant impact because Microsoft did not actively monitor compliance or enforce their
terms. Myhrvold ¶ 36 (arguing that Microsoft "never attempted to enforce" the IEAK
contractual provision requiring ISPs to make Internet Explorer their "preferred browser");
Myhrvold, 2/10/99pm, at 42:19-22 ("We don't place any restrictions on that. We don't
say this is the behavior you must do to meet that--to meet that condition."). This
argument is wrong.
254.1. First, ISPs were contractually bound to honor, and in large
measure did in fact honor, Microsoft's restrictions. Although Mr. Myhrvold on the stand
sought to portray Microsoft's contractual restrictions as "requests," he conceded under
questioning by the Court that Microsoft "conditioned" its licensing agreements on these
exclusionary terms and that ISPs are thus "arguably in violation of their license" if they
do not make Internet Explorer their preferred browser.
- Myhrvold conceded that the licenses are conditioned on Internet
Explorer being the preferred browser. "It does require it makes it
their preferred browser." Myhrvold, 2/10/99pm, at 42:16-17;
Myhrvold, 2/10/99pm, at 41:15 - 43:7.
254.2. Second, Microsoft did, in fact, enforce its restrictions. Microsoft
actively monitored compliance with the shipment restrictions limiting the distribution of
non-Microsoft browsers.
- Microsoft monitored ISPs' compliance with its shipment restrictions
by requesting that ISPs report the number of non-Microsoft
browsers distributed. GX 368 (Microsoft e-mail requesting
competitive browser shipment estimates from certain ISP accounts,
including Netcom, Concentric, Earthlink, and Mindspring).
- Microsoft took notice when ISPs failed to meet certain requirements
and took steps to ensure compliance. Myhrvold, 2/10/99am, at
55:5-14.
- Although it is not clear whether Microsoft ever removed an ISP from
the referral server for breaching its obligations (GX 228, at MS98
0113062 (suggesting that Netcom might have been temporarily
removed from the Internet Explorer 4 referral server, which was
"negatively affect[ing] our business")), it is clear that Microsoft at
least considered removing ISPs. A February 1998 Microsoft
presentation proposed terminating Earthlink and Brigadoon from the
Referral Server for "noncompliance" of their contracts. GX 429, at
MS98 0102462.
254.3. Third, Microsoft stringently enforced the most important restrictions,
those it imposed on Online Services, in particular AOL.
- David Colburn testified that Microsoft carefully monitored any
references to Netscape Navigator on AOL's service. Colburn Dir. ¶
30. For instance, Brad Chase, the Microsoft executive in charge of
its relationship with AOL, complained when it appeared that AOL
took steps to align with Netscape, thus threatening to impair what
Chase viewed as the "virtual exclusivity" terms to which AOL had
agreed. Colburn, 10/29/98pm, at 58:9 - 59:12.
- In one email, Colburn assured Chase that AOL was not selling
advertising on the AOL service to Netscape and had, in fact,
"stamped the NS issue into the ground". GX 186.
255. Microsoft witnesses also made much of the fact that, immediately before
this case was filed, Microsoft waived some of the restrictions in its IRS agreements. (GX
374; Myhrvold Dir. ¶¶ 91-92). But Microsoft's partial waiver is of very limited
significance.
255.1. First, Microsoft did not waive any restrictions in its more
competitively significant agreements with OLSs and specifically declined AOL's request
that it waive the restrictions Microsoft placed on AOL's dealings with other browsers.
Microsoft's exclusionary restrictions, therefore, remain in effect with the largest providers
of Internet access, including AOL.
- Mr. Chase testified that Microsoft did not waive the preference
clauses for certain OLS's. Chase, 2/16/99pm, at 21:4-7.
- GX 226 (AOL letter to Microsoft stating that Microsoft's refusal to
include OLSs in the waiver excludes "a substantial percentage of
the US ISP business.").
- Professor Fisher testified that "restrictions were not waived for ISPs
who were OLSs." Fisher Dir. ¶ 188
- David Colburn testified that Microsoft has "demanded that AOL
continue with" the "exclusivity provisions or lose the right of
Compuserve to be included in the Online Services Folder and the
ICW, and for AOL to continue to be listed in the referral server."
Colburn Dir. ¶ 47.
255.2. Second, even with respect to ISPs, Microsoft did not waive all of
the restrictions.
- Microsoft still prohibits ISPs that appear in the Internet Connection
Wizard from offering or promoting Netscape or other browsers as
their "default" browser. Fisher Dir. ¶ 187.
- In addition, since a number of mostly small ISPs choose to distribute
only one browser in order to reduce support costs, the "requirement
of 'parity' for Internet Explorer in order to secure access to the ICW
may amount to a de facto requirement that the ISP exclusively
support Internet Explorer." Warren-Boulton Dir. ¶ 109.
255.3. Third, and most important, Microsoft relaxed the restrictions only
after the damage was done. As explained, Microsoft's restrictions, including the
restrictions it recently relaxed, substantially contributed to Internet Explorer's increasing
share and Netscape Navigator's decline.
- See supra Part VII.A; ¶¶ 369-370.4.2.
- Professor Fisher testified: "Whatever the extent of Microsoft's
waiver, it did not undo the harm to competition that had already
occurred." Fisher Dir. ¶ 190.
- Warren-Boulton testified that Microsoft cannot erase the "significant
anticompetitive effects" of the restrictions "simply by removing"
them. Warren-Boulton Dir. ¶ 109; see also Warren-Boulton,
11/30/98pm, at 60:7-23; 15:16 - 17:17 (explaining that, by the time
Microsoft lifted its restrictions, the game was "pretty well over.)."
- Microsoft's own documents reflect its belief that around the time it
waived its restrictions it had "won" the browser war and vitiated the
threat Netscape posed to its operating system monopoly. See infra
Part VII.A.4; ¶ 363; Part II.B.3.c; ¶ 388.2.
(5) Microsoft's agreements were exclusionary and
anticompetitive notwithstanding the small number
of subscribers ISPs and OLSs garnered from the
referral server
256. Microsoft witnesses also suggested that its agreements could not have
been anticompetitive because OLSs and ISPs obtained only a small percentage of their
subscribers through the Online Services Folder and Internet Connection Wizard
(Myhrvold Dir. ¶¶ 82-85, 124; Schmalensee Dir. ¶¶ 424-425). But this assertion ignores
that Microsoft's agreement restricted ISPs' and OLS's distribution and promotion of rival
browsers in all channels and to all customers, not merely to customers who contacted
the access provider through the OLS Folder or ICW.
- See supra Part V.D.2.d; ¶ 223.
5. Microsoft's justifications for its agreements are pretextual
257. Microsoft witnesses advanced a number of justifications for its efforts to
obtain preferential distribution of its browser through ISPs and OLSs. Their arguments,
however, can explain neither Microsoft's expensive effort to gain distribution of its
browser through access providers nor the exclusionary restrictions that it paid access
providers to accept.
257.1. First, Microsoft contends that its restrictive terms are justified
because Microsoft has an interest in preventing firms that take its valuable assistance
from turning around and promoting rivals. Chase Dir. ¶ 98. As Chase explained, "given
the commitment we were making to AOL, we wanted AOL to commit to using IE."
Chase Dir. ¶ 75. This contention is misconceived.
257.1.1. Microsoft did not need the exclusionary provisions in order
to be fully compensated for the consideration it gave to the ISPs and OLSs. It could
have asked for money or other compensation, rather than exclusion, in exchange for the
browser, technical assistance, desktop placement, and other value Microsoft provided.
- Dr. Warren-Boulton testified that "there is no reason why
Microsoft had to take its compensation in the form of
exclusionary agreements rather than a simple payment. That
ISPs commonly agree to pay for customer referrals and
promotion of their services demonstrates that selling desktop
real estate does not involve prohibitive transaction costs."
Warren-Boulton Dir. ¶ 183.
- Professor Fisher testified that "rather than trading desktop
space for financial renumeration, Microsoft placed
requirements on ISPs that hindered their ability to promote or
distribute Netscape Navigator." Fisher Dir. ¶ 192.
- Bill Gates recognized that Microsoft had the option to
"monetize the box, and sell the real estate to the highest
bidder", but instead chose to use placement on the Windows
desktop "for the browser battle." GX 1372, at 5.
- OEMs charged fees for promoting access providers' services
(Colburn Dir. ¶ 18), and there is no reason why Microsoft
could not have done so as well; see also GX 621 (Microsoft's
willingness to bundle AOL's software with Windows made a
browser deal with Microsoft potentially more profitable than a
Netscape deal because AOL would save the bounties it
would otherwise pay OEMs).
257.1.2. The breadth of Microsoft's restrictions belies its contention
that it merely wanted to ensure that ISPs did not promote other browsers to customers
that it gained from Microsoft.
- Microsoft's contract with AOL (to take an example applicable
to all of Microsoft's IRS and OLS agreements) prohibits AOL
from distributing Netscape upon customer request, if doing so
exceeds the limitations contained in the shipment restrictions,
even if that customer never had any contact with the Online
Services Folder or benefitted from any of the technical
assistance Microsoft provided AOL. See supra Part V.D.2.a;
¶ 216. AOL obtains the majority of its subscribers from
channels other than Microsoft. DX 2098, at D-2.
- Dr. Warren-Boulton testified that "Microsoft's restrictions on
the ability of ISPs, OLSs, and ICPs to promote and distribute
competing Internet browsers are unrelated to any efficiency
purpose." Warren-Boulton Dir. ¶ 182.
257.2. Second, Microsoft witnesses point out that the Internet Connection
Wizard, Online Services Folder, and associated features are intended to make it easy
for users to hook-up to the Internet (Myhrvold Dir. ¶ 43; Schmalensee Dir. ¶¶ 441-442).
But this cannot justify Microsoft's exclusionary restrictions because restricting the
distribution or promotion of other browsers is unnecessary to achieve this benefit. To
the contrary, the very video demonstration Mr. Myhrvold sponsored shows that OEMs
can give users a "seamless" experience connecting to the Internet by adding the
Netscape referral server to Windows.
- Cameron Myhrvold testified that OEMs can add the Netscape
referral server to Windows, explaining: "in this example, if you're
talking about offering a seamless experience to the user, I think it
may. If they click on that icon, they get the referral server."
Myhrvold, 2/9/99pm, at 35:9 - 37:2.
257.3. Third, Microsoft argues that "Microsoft's agreements with the ten
ISPs in the Windows 95 Referral Server were the type of cross marketing agreements
that are routine in all industries, particularly in connection with the Internet" (Myhrvold
Dir. ¶ 88; see also Chase Dir. ¶ 97 (same for OLSs)). But Microsoft's agreements are
not ordinary cross marketing arrangements.
257.3.1. The terms of Microsoft's agreements are inconsistent with
Microsoft's "cross marketing" characterization. As explained, Microsoft's "shipment
restrictions" prohibit access providers from supplying non-Microsoft browsers to
customers, even when customers specifically request other browsers and even if the
ISP obtained the customer for reasons wholly unrelated to Microsoft's promotion of it.
- See supra Part V.D.2; ¶¶ 215.3-4, 216, 217.3-5.
257.3.2. Contrary to the testimony of Microsoft witnesses, the terms
Microsoft extracted are not "routine." Netscape's agreements with the RBOC's Internet
access providers, for instance, are substantially different.
257.3.2.1. Netscape's agreements are significantly less
restrictive than Microsoft's.
- The RBOCs agreed to make Netscape Navigator their
"default" browser and to certain promotional
restrictions. But, by contrast to the terms Microsoft
extracted, the RBOCs did not agree to restrictions on
their ability to distribute other browsers, such as
Internet Explorer, in response to customer requests or
otherwise. Beran Dep., 1/13/99, at 119:14 - 120:15
(testifying that, under its agreement with Netscape,
Bell Atlantic was "free to offer other browsers").
- James Barksdale testified that Netscape's RBOC
contracts "are not exclusive . . . . There is nothing in
the contracts that prohibits the RBOCs from
distributing another browser to their customers--in any
numbers." Barksdale Dir. ¶¶ 131-132.
- Indeed, RBOC representatives testified that they
preferred dealing with Netscape because the
restrictions insisted upon by Microsoft were more
onerous and "thwarted" their "objective of providing
browser choice." Beran Dep., 1/13/99, at 117:10 -
119:12.
257.3.2.2. Even the terms in the Netscape agreements that
call for Navigator to be the default browser were in direct response to Microsoft's
restrictions on the largest access providers. In other words, they were part of
Netscape's effort to open up a channel of distribution which Microsoft had closed.
- The terms securing preferences for Netscape were
negotiated only after Microsoft diminished Netscape's
opportunities by entering into its exclusionary
agreements with the most significant access providers.
Indeed, Netscape's contracts with the RBOCs required
default status for Navigator only "so long as AT&T and
MCI . . . are both restricted by agreement from
providing Navigator to their customers." GX 1151, at
AM 00076 (Amendment to the Network Service
Provider Distribution Agreement, section 15); GX
1152,at RAA 0074 (Amendment to the OEM License
Agreement between Netscape and Bell Atlantic,
section 15).
257.3.2.3. Netscape's agreements with the RBOCs account
for only a small percentage of Internet access in the United States.
- Myhrvold conceded that the RBOCs combined have a
subscriber base between 1 and 2 million, less than
10% of AOL's subscribers. Myhrvold, 2/10/99pm, at
79:11 - 80:5.
- Barksdale testified that "the RBOCs account for less
than 5% of the total ISP marketplace." Barksdale Dir.
¶ 132.
257.3.2.4. Microsoft's agreements are not typical cross
marketing agreements for a more fundamental reason. In typical cross marketing
arrangements, the product being marketed is a profitable product. But Microsoft's effort
to purchase browser market share can be explained only as a strategy designed to
weaken Netscape and protect Microsoft's operating system monopoly.
- See infra Part V.G.2; ¶¶ 297.4.1.
E. Microsoft entered into anticompetitive, exclusionary agreements with
Internet Content Providers
258. As part of its strategy to eliminate the browser threat to its operating system
monopoly, Microsoft also entered into agreements with Internet Content Providers
(ICPs). The purpose of these agreements was similar to Microsoft's exclusionary
agreements with ISPs and OLSs. Microsoft provided ICPs valuable consideration,
including placement on the Windows desktop, in exchange for ICPs' agreement to
promote and distribute Internet Explorer and to exclude and disadvantage browser rivals.
Just as with its ISP and OLS agreements, the exclusionary restrictions Microsoft
extracted can be explained only as part of a strategy to maintain its operating system
monopoly.
1. Microsoft determined that ICPs could help it win the browser
war
a. Internet Content Providers
259. ICPs develop the "content," i.e., web pages or web sites, that makes up the
Web. ICPs create web content using a number of different technologies, such as HTML
and Java.
- William Poole testified that the World Wide Web consists of web pages or
sites that are displayed in particular formats, including, among others, the
popular HTML protocols. Poole Dir. ¶¶ 15, 17, 23.
- Professor Franklin Fisher testified that "ICPs create program content for the
World Wide Web." Fisher Dir. ¶ 193.
260. ICPs generate revenue in two principal ways.
260.1. First, like television networks, ICPs may derive revenue by charging
third-parties a fee for advertising or promotion on the ICP's web site.
- Intuit's William Harris testified "the owner of a web site can sell
advertising and sponsorships on the site." Harris Dir. ¶ 15.
- Harris further testified that, before it entered into its restrictive
agreement with Microsoft, Intuit provided promotion for Netscape by
including a "Netscape Now" button on its web site. Harris Dir. ¶ 71.
- According to Dean Richard Schmalensee, Netscape's Netcenter
website is "highly trafficked;" hence, placement on Netcenter is
extremely valuable--Sportsline agreed to pay Netscape $500,000 for
ads on Netscape's site. Schmalensee Dir. ¶469.
260.2. Second, ICPs may also derive revenue by charging users who visit
their sites a fee for viewing or using particular content.
- William Harris testified that web site owners "can earn user fees
directly from the customer." Harris Dir. ¶ 15.
- For example, "PC Data" (pcdata.com) charges users a direct fee for
accessing "Retail Hardware Category Reports." DX 2495.
261. An ICP's success depends on the popularity of its web sites because the
number of users who visit a particular web site affects both the ICPs' advertising rates
and the fees it can charge users.
- William Harris testified: "Volume of customers at a web site is the single
most important factor in determining advertising rates that the web site
owner can charge for space on the web site. Similarly, customer volume
will normally bear directly on the volume of transaction and user fees that
can be earned by the web site owner." Harris Dir. ¶ 16.
- Professor Fisher testified "ICPs valued the opportunity to have a channel
on the Microsoft desktop, because it encouraged users to visit the ICPs
website, which in turn increased the ICPs ability to promote their own
products and to sell advertising space on their web pages." Fisher Dir.
¶194.
262. Microsoft argues that there are literally millions of web sites and thousands
of ICPs (Poole Dir. ¶¶ 24, 58-60, 78). However, users tend to visit sites with which they
are familiar; as a result, a handful of the leading ICPs are of special commercial
significance.
- A 1996 ActiveX "Winning @ Internet Content Marketing Plan" explains, as
part of its strategic rationale, "The top 100 commercial Internet Sites will
drive 90% of the traffic." GX 407, at MS6 5005718.
b. Microsoft determined that inducing leading ICPs to favor
Internet Explorer and disfavor rivals would facilitate
winning the browser war
263. Microsoft concluded that influencing leading ICPs to favor Internet Explorer
and disfavor rivals was important to its goal of winning the browser war for two principal
reasons.
263.1. First, Microsoft determined that influencing the technologies that
key ICPs implement in their web sites would increase Internet Explorer's market share.
263.1.1. Web sites can be designed to take advantage of
technologies that work only (or better) with a particular browser. For example, web page
content designed using Microsoft's "ActiveX" technologies -- an Internet Explorer-specific
technology -- cannot be viewed with Netscape Navigator.
- Dr. Warren-Boulton testified: "one of the Microsoft-specific
technologies is known as 'ActiveX' . . . . The crucial feature of
ActiveX for my purposes is that it is operating system
(typically Windows) specific." Warren-Boulton Dir. ¶¶ 6-7.
- In a June 22, 1996 marketing plan, entitled "Winning @
Internet Content: Marketing Plan," Microsoft described that its
"strategic objective" was to "gain leading share for ActiveX as
the primary component architecture for the top 100 web sites
. . . . Our ability to get these sites to adopt ActiveX
technology will be vital to achieving our overall goal." GX
407, at MS6 500517.
263.1.2. Microsoft believed that web content developed with
Microsoft's technologies would help drive and increase usage of Microsoft's browser and
thus sought to induce ICPs to adopt Internet Explorer-specific technologies.
- On November 1, 1995, Microsoft's Chris Jones sent Bill Gates
an e-mail "about what Microsoft should do to get 30%
browser share." GX 334, at MS98 0104679. In the attached
memorandum, Jones wrote:
"3. Get 80% of Top Web Sites to Target Our Client.
Content drives browser adoption, and we need to go to the
top five sites and ask them 'What can we do to get you to
adopt IE?' We should be prepared to write a check, buy
sites, or add features -- basically do whatever it takes to drive
adoption."
- The "top hundred commercial Internet sites will drive 90% of
traffic." GX 407, at MS6 505717.
263.2. Second, Microsoft believed that popular ICPs were an important
vehicle for promoting Internet Explorer.
- As part of a February 22, 1996, plan discussing ways Microsoft
could increase Internet Explorer's share, Paul Maritz wrote that
reaching agreements with ICPs, particularly with respect to "key
sites," was an important component of Microsoft's strategy. GX 473,
at MS6 6006248.
- In explaining the purpose of the ICP Platinum Agreements, Poole
wrote "Microsoft sought to promote consumer use of IE because at
the time these agreements were negotiated . . . Microsoft's share of
browser usage was quite modest." Poole Dir. ¶ 65.
2. To achieve its objective of gaining browser usage share,
Microsoft entered into exclusionary agreements with ICPs
264. In order to enlist ICPs in its objective of increasing its browser market share
and disadvantaging its browser competitors, Microsoft pursued a strategy similar to the
strategy it employed with ISPs and OLSs. Microsoft offered ICPs prominent placement
and promotion within the Windows desktop, without charge, on the condition that ICPs
agreed to exclude Netscape.
a. Microsoft developed the Channel Bar believing that it
would generate substantial revenue
265. Microsoft developed the "Channel Bar" as a feature of the "Active Desktop,"
which Microsoft first released with its Internet Explorer 4 product in 1997. The "Channel
Bar" appeared on the default Windows desktop and included icons that provided end-users with direct access to particular ICPs' web sites. The distinctive feature of the
Channel Bar, apart from its placement on the Windows desktop, is its use of "push"
technology, by which users can arrange with the participating ICP to receive certain
Internet content automatically.
- Mr. Poole testified that "Microsoft's push technology in
Internet Explorer 4.0 enables computer users to set up a
schedule for information from specified web sites to be
downloaded to their computers automatically;" further, ICPs
implement Microsoft's "push" technology by creating
computer files that adhere to the "Channel Definition Format,"
or "CDF." Poole Dir. ¶ 48.
- According to Harris, "The Active Desktop was intended to be
a new user interface for Windows that would, among other
things: a) permit direct access to the Internet; b) provide a
choice of Internet web sites that could be directly accessed
from the desktop; and c) permit the user to view Internet
content on the desktop itself without separately opening the
browser." Harris Dir. ¶ 57.
266. Microsoft offered ICPs several different levels of placement on the Channel
Bar:
266.1. The highest level of placement on Microsoft's Channel Bar was
known as the "platinum" level, which included six ICPs whose icons were located directly
on the Windows desktop.
- The Microsoft Network, MSNBC News, Disney, Pointcast,
Warner Bros., and America Online were the six websites
initially visible--linked via icon--on the Channel Bar. Poole
Dir. ¶ 49.
266.2. Microsoft's "platinum" partners also included those ICPs that
Microsoft listed within certain pre-set categories (e.g., "Lifestyle & Travel,"
"Entertainment," "Sports," "Business," "News," and "Technology"). A user selecting one
of these categories was presented with links to certain platinum-level ICPs.
- For example, clicking on the "Sports" category reveals CBS
Sportsline, CNNSI, ESPN SportsZone, and MSNBC Sports.
Connected users can choose to "subscribe" to the selected
web site, subsequently having related information
downloaded on a regular basis. Poole Dir. ¶ 48.
267. Microsoft required OEMs shipping Windows 95 to install Internet Explorer 4
and ship the Channel Bar "active" to end users, i.e., "turned on" as part of the default
Windows desktop, thus ensuring that the Channel Bar would be seen by millions of
Windows users, who would then have an opportunity to invoke the channels and visit the
included ICPs.
- Poole conceded that OEMs shipping to consumers were not permitted to
turn off the Channel Bar in Windows 95. Poole, 2/8/99am, 19:8-12. ICPs in
the Channel Bar enjoyed "widespread consumer exposure to their content
via links accessible through the channel bar." Poole Dir. ¶ 68.
- Dr. Frederick Warren-Boulton testified that: "Because the IE 4 Active
Desktop was anticipated to be shipped on a substantial number of PCs, the
channels -- just like Microsoft's On-Line Service Folder and Internet
Connect Wizard -- provided ICPs with an attractive way of promoting their
services." Warren-Boulton Dir. ¶ 114; see also Fisher Dir. ¶ 194.
268. Microsoft believed that the Channel Bar would be very valuable to ICPs and
that Microsoft could generate millions of dollars in revenue by charging ICPs for
placement on the Channel Bar.
- Microsoft predicted that the Channel Bar would generate total annual
revenues of "many tens of millions, maybe a hundred million" dollars.
Poole, 2/8/99am, at 25:12-19. William Poole testified that Microsoft initially
planned to charge for placement on the Channel Bar and estimated that
each top-level channel might generate as much as $10 million annually and
that other channels might generate a couple of million each. Poole,
2/8/99am, 22:16-21, 23:9 - 24:8, 24:22 - 25:5. Morever, ICPs approached
Microsoft to determine what they would need to pay to appear on the
Channel Bar. Poole, 2/8/99am, at 33:23 - 34:8.
- Microsoft entered into an agreement with Pointcast obliging Pointcast to
pay $10 million a year for placement on the Channel Bar. GX 1804, at
MS98 0100814 (section 3.3b).
- Poole, the Microsoft executive in charge of its ICP business, told ICPs that
placement on the Channel Bar would be "invaluable." Poole, 2/8/99am, at
35:8 - 36:6.
- In an October 1996 e-mail, Yusuf Mehdi noted that there seems "like there
is an oppty to get ALOT of money from sales to OEMs by having them
purchase the 'default' channel or settings and reselling them to content
providers. This is an easier way to get out of the business of brokering
one-off deals and having the OEM pay for aggregation rights. You can
charge yearly and they do the work. If they don't want to, we sell it." GX
98.
- Microsoft told CNet, a platinum-level ICP, that "we're telling the consumer
that these" default ICPs "are the very best sites on the Web. The thing you
have to keep in mind is that we are offering the default ICPs tremendous
amounts of distribution, which is worth a great deal." GX 207, at CNET
000464.
- Disney's Steve Wadsworth testified that Microsoft's Brad Chase and Bill
Spencer pitched placement on the Windows desktop as highly valuable to
Disney, and described the desktop as Microsoft's "crown jewel."
Wadsworth Dep. (played 12/15/98 am), at 31:19 - 34:22.
b. Microsoft nonetheless decided not to charge ICPs for
placement on the Channel Bar, but rather to use such
placement as "strategic barter"
269. Instead of obtaining revenues from selling placement on the Channel Bar to
ICPs, Microsoft opted to use the Channel Bar for what it called "strategic barter."
Specifically, instead of charging ICPs, Microsoft conditioned Channel Bar placement on
ICPs' agreement to terms that severely restricted their business dealings with Microsoft's
browser rivals. ICPs agreed to these restrictions because they determined, as had
Microsoft, that Channel Bar placement would be extremely valuable.
- Poole conceded that Microsoft, rather than pursuing revenue,
characterized the purpose of desktop access as "strategic barter." Poole,
2/8/99am, at 34:9-19.
(1) Microsoft's exclusionary agreements
270. In exchange for platinum-level placement twenty-four ICPs agreed, with
regard to thirty-one separate web sites, to a number of restrictions. Although the precise
terms of each "Platinum" agreement differed somewhat, the agreements generally
included a contractual commitment by ICPs:(4)
270.1. To restrict severely their ability to promote and distribute "Other
Browsers," typically defined as the "top two" competing browser companies, excluding
Microsoft;
270.2. To implement IE-specific technologies (for instance ActiveX) in
certain web pages, even if doing so would result in what Microsoft termed "acceptable
degradation" when viewed with "Other Browsers";
270.3. Not to pay "Other Browsers" any consideration to promote the ICPs'
content; and
270.4. To use Internet Explorer exclusively in any Windows or Macintosh
client.
271. Microsoft also obtained exclusionary restrictions that applied to these firms'
promotion and distribution of browsers in channels that did not relate to their ICP
businesses. In order to obtain placement on the Channel Bar, platinum partners that
distributed software such as Intuit and PointCast had to agree to promote and distribute
Internet Explorer almost exclusively.
- Pointcast's agreement with Microsoft stipulates that "Pointcast and its
subsidiaries shall, exclusive of any other browser, market, promote and, to
the extent of Pointcast's distribution efforts in the ordinary course of its
business, distribute (via download, OEM, retail and direct distribution) the
combined Pointcast Network v2.0 and the Windows 3.xx and Macintosh
Internet Explorer v3.0 to its end users on the appropriate platforms." GX
1153, at MS98 0100811 (PointCast agreement, section 2).
- Intuit was required to "promote and distribute Internet Explorer (and no
other browser) as the browser of choice for QFN, Intuit products, and Intuit
web sites." GX 1156, at INT 00005 (Intuit agreement, section 2.2).
271.1. Prior to entering into its ICP agreement with Microsoft, Intuit
distributed Navigator with its most popular software titles, Quicken, Turbo Tax, and
QuickBooks and served as an important distribution channel for Netscape.
- William Harris testified that the Active Desktop agreement
"precluded Intuit from further including Netscape Navigator in copies
of Intuit's applications;" moreover, it did not allow Intuit the ability to
"enter into a business relationship with Netscape in order to direct
potential Navigator users to Intuit's web site." Harris Dir. ¶¶ 72, 76.
- Jim Barksdale testified, "independent software vendors have also
been an important distribution channel for Netscape software . . .
Microsoft has tried to eliminate the ISVs as a meaningful distribution
channel for Netscape." Barksdale Dir. ¶ 174.
271.2. Even before Microsoft offered Intuit placement in the Channel Bar,
Microsoft tried to induce Intuit to distribute Internet Explorer instead of Navigator. In July
1996, Bill Gates offered Intuit's CEO, Scott Cook, a $1 million "favor" if he would switch
browsers, an offer Mr. Cook rebuffed.
- In reference to Cook, Bill Gates wrote: "I was quite frank with that if
he had a favor we could do for him that would cost us something like
$1M to do that in return for switching browsers in the next few
months I would be open to doing that." GX 94.
- William Poole agreed that Microsoft wanted Intuit to ship Internet
Explorer and that Gates sought to induce Intuit to switch browsers.
Poole, 2/8/99am, at 38:19 - 41:23.
271.3. In order to obtain placement on the Channel Bar, Intuit entered into
an ICP agreement with Microsoft. In doing so, Intuit, among other things, was required to
distribute Internet Explorer with all of its software products and not to distribute or
promote rival browsers such as Netscape except in very limited circumstances.
- William Harris testified that, because Intuit deemed Active Desktop
placement vital to the success of their web sites, they "felt
constrained to agree" to the conditions of the deal which required
them to "forego business relationships with Microsoft's browser
competitor, Netscape," including their existing promotion and
distribution arrangements. Harris Dir. ¶¶ 6, 61, 71, 74, 75; see also
Warren-Boulton Dir. ¶ 118.
- The Intuit/Microsoft agreement required that Internet Explorer -- and
no other browser -- be promoted and distributed with Intuit products
(such as Quicken 98, Turbotax Personal, and Quickbooks), including
the integration of Internet Explorer 4 "on successor versions."
(6/6/97 MS/Intuit agreement, section 2.2; adds obligation for Internet
Explorer to be the exclusive browser in release of subsequent Intuit
products) GX 1156, at INT 00005; see also, GX 868.
- William Poole, summarizing the terms of Microsoft's agreement with
Intuit, wrote that Intuit would be obligated to bundle Internet Explorer
with "all new 97 and 98 releases of Intuit products" and "not enter
into marketing/promotion agreements with other browser
manufacturers for distribution or promotion of Intuit content." GX
206; see also Poole, 2/8/99am, at 55:1-22 (testifying that GX 206
summarizes the "salient" terms of the Intuit/MS agreement).
272. Microsoft also entered into other agreements relating to the Active Desktop
and Channel Bar that had a similar effect of restricting an ICP's ability to work with
Microsoft's browser competitors.
272.1. Microsoft entered into approximately 30 to 50 second-tier "Gold
Agreements," which provided for promotion in the Active Channel Guide but not on the
Active Desktop. The Gold Agreements included "parity requirements" that required an
ICP that promoted Navigator to also promote Internet Explorer on "equal footing"; in
other words, the Gold ICPs could not make Netscape their preferred ICP.
- Poole explained that Microsoft had "in the vicinity of 30 to 50" Gold
Agreements, which he described as "second tier" deals. Poole,
2/8/99pm, at 36:24 - 37:6. Poole testified that Microsoft's "gold"
agreements included "parity requirements" that necessitated "equal
footing with Netscape." Poole, 2/8/99pm, at 37:7-19.
- Poole testified that the Gold agreements were "similar to the
Platinum agreements" but "did not include promotion on the
Windows desktop." Poole, 2/9/99am, at 13:24 - 14:9.
272.2. Microsoft also entered into approximately eight to twelve "IEAK
Agreements" with ICPs, which restricted the dealings that an ICP could have with
Netscape. In exchange for distribution of its icon with the Internet Explorer
Administration Kit, participating ICPs agreed to promote Internet Explorer as the
preferred browser, to refrain from promoting other browsers on their web sites, and to
create "differentiated" content (content that would work well only with Internet Explorer).
The IEAK agreements also required ICPs that distributed a browser to distribute Internet
Explorer exclusively.
- Poole testified that Microsoft had approximately a dozen IEAK
agreements, which he explained as "other agreements that had
restrictions with ICPs and Netscape." Poole, 2/8/99pm, at 39:3-10.
Poole agreed that the IEAK agreements restricted companies from
dealing with Netscape, saying, "the partner would promote IE as its
preferred browser, and it would not promote another browser on
their web site." Poole, 2/8/99pm, at 39:11-21.
- Microsoft's IEAK agreement
- redacted -
GX 1174, at MS98 0100073-74 (sealed).
(2) ICPs agreed to these restrictions in order to get
placement on the Windows desktop
273. In order to obtain access to the Channel Bar and placement within the
Windows desktop, ICPs had to agree to Microsoft's restrictions, despite the ICPs desire
to enter into promotional and distributional deals with competing browser companies, like
Netscape.
- Harris testified that, "if not for the restrictions imposed by the Active
Desktop agreement, I believe that Intuit would have shipped both Netscape
Navigator and Microsoft Internet Explorer with its products. However, Intuit
was not permitted by the terms of the Active Desktop agreement to ship
both browsers." Harris Dir. ¶ 79.
- Disney's Steve Wadsworth testified that "Netscape was specifically
discussed in the context of Disney's desire not to have an exclusive
arrangement with Microsoft." GX 776, at ¶¶ 7-8 (Wadsworth Decl.
4/23/98); see also Barksdale Dir. ¶ 182 ("Disney told us they would have liked to
do a deal with Netscape, but as a condition of their Microsoft contract for
placement on the Windows desktop, Disney was prohibited from offering Netscape
compensation of any kind! Disney would also have been prohibited from
promoting, even informing customers, or even letting us promote, the Disney.com
placement on Netscape's products." (emphasis in original).
274. ICPs agreed to Microsoft's restrictions because they believed, as did
Microsoft, that, by virtue of placement on the Channel Bar, promotion and distribution
with every new PC preloaded with Windows would be "very valuable."
- Poole conceded that he and others at Microsoft told ICPs that being on the
Windows desktop was going to be a distribution advantage to them and
that ICPs believed that placement on the Channel Bar would be "very
valuable." Poole, 2/8/99am, at 34:23 - 25.
- Intuit believed that the Windows desktop would provide "unparalleled
distributional advantages" and was thus prepared to pay a substantial fee
for placement on the Channel Bar. Harris Dir. ¶ 60. Microsoft instead
insisted on an agreement to restrict Intuit's business relationship with
Netscape. Harris Dir. ¶ 65. It was because of placement on the Channel
Bar that Intuit agreed to Microsoft's exclusionary restrictions. Harris Dir. ¶
79.
- Disney's Steve Wadsworth testified that Disney agreed to limit its marketing
and promotion of other browsers because "having the channel and getting
the deal done and having the Active Desktop channel was worth giving up
some abilities to do some other marketing and promotion on behalf of
someone else. So we determined it was worth it, and we wanted to have
the Active Desktop channel." Wadsworth Dep. (played 12/15/98am), at
43:23 - 44:6. Disney believed that desktop placement would be valuable
because Windows "is highly distributed and holds a majority market share
among operating systems for PCs." Wadsworth Dep. (played 12/15/98am),
at 32:1-9; Wadsworth Dep. (played 12/15/98am), at 34:9 - 39:21 (OEMs
can't change Windows desktop; Disney understood that Internet Explorer
would be shipped with every PC); see also GX 776, at ¶ 4 (Declaration of
Steve Wadsworth, Vice President of Business Development and Strategic
Planning for Disney, 4/23/98; Disney believed desktop placement would be
"highly valuable"; "Although Disney could have attempted to negotiate
separate deals with numerous individual Personal Computer manufacturers
in an attempt to match the level of distribution Microsoft could provide for
the Disney channel, no single company other than Microsoft could have
provided the same value in terms of ubiquitous distribution."); GX 776, at ¶
5 (Microsoft told Disney "that the desktop is Microsoft's `crown jewel'").
Wadsworth further testified that the Channel Bar gave Microsoft
tremendous "leverage." Wadsworth Dep. (played 12/15/98am), at 42:9-15
("And it -- I was -- I felt like, yeah, these guys have all of the cards, because
they have this broad distribution capability through the Windows operating
system and the desktop. And you know, in the end, yeah, they have
substantial market share with PCs. So in my mind that made them the
1000 pound gorilla of the industry.").
- David Colburn testified that AOL agreed to Microsoft's restrictions because
of AOL's judgment that placement on the Channel Bar would be very
valuable. Colburn Dir. ¶¶ 40-47; Colburn, 10/29/98pm, at 19:21 - 21:3,
21:19 - 22:1.
- ZDNet believed that "a preferred position on the active desktop -- which
means being bundled into the operating system -- is of almost incalculable
value." GX 201.
275. The value ICPs placed on the Channel Bar, and the corresponding leverage
it gave Microsoft, is also shown by ICPs' capitulation to demands by Microsoft that the
ICPs restrict their dealings with Netscape even in ways not required by the ICP
agreements.
- Disney agreed to Microsoft's insistence that Disney could have a text link
on either the Netscape browser or home page, but not both, and that
Netscape could not use either the well-known Disney characters or its
logos in connection with its browser technology. A contemporaneous
Disney e-mail explains that Disney could not afford to resist Microsoft's
insistence because: "Unfortunately, Microsoft has the upper hand from a
business value perspective even though they don't have it from a
contractual perspective" and "it seems crystal clear to me we are in the
right on the contract, but even so, it's probably not worth it to take them on.
The value of the Netcaster channel is low, and if they take us off the Active
Desktop while this is being resolved in court, we lose substantial value."
GX 213.
- Wired Digital agreed to restrictions (such as not splitting their contention
between Internet Explorer and Netscape) that went beyond the terms of the
ICP agreements. Poole, 2/8/99pm, at 54:19 - 55:20, 56:7-19. Wired Digital
was told by Microsoft's Suzan Fine that, "despite the contract," Microsoft
"would consider it counter to the 'spirit' of the agreement with MS for us to
have any Wired-branded presence on any Other Browser, even if the
agreement as we've whittled it back doesn't technically prohibit it." GX 209.
3. Microsoft's ICP agreements were exclusionary
276. Microsoft expected its agreements, in particular its "Platinum" agreements,
to further its goal of wresting substantial browser usage share from Netscape and
protecting its operating system monopoly. Although their exclusionary impact turned out
to be less than Microsoft anticipated, the ICP agreements nevertheless contributed to
Microsoft's anticompetitive campaign.
a. Microsoft specifically intended and anticipated that its
ICP agreements would deprive Netscape of revenue,
exclude Netscape and other browser rivals, and protect
Microsoft's operating system monopoly
277. Microsoft entered into its ICP agreements in order to achieve two principal
goals: (i) to gain browser share and (ii) through gaining browser share, to neutralize the
threat to its operating system monopoly non-Microsoft browsers posed.
277.1. Microsoft's ICP agreements were capable of accomplishing these
objectives in three ways.
277.1.1. First, Microsoft's ICP agreements inhibited the promotion
and dissemination of "Other Browsers," defined by Microsoft to target only those
browsers that posed a competitive threat to Microsoft, such as Netscape.
- Dr. Frederick Warren-Boulton testified that Microsoft'
agreements "directly inhibit the promotion and dissemination
of non-Microsoft browsers in ways similar to the ISP
restrictions." Warren-Boulton Dir. ¶ 116.
277.1.2. Second, Microsoft's ICP agreements contained terms
designed to deprive Netscape of revenue by prohibiting ICPs from paying "Other
Browsers" to promote ICPs' content, a typical way Netscape obtained revenue.
- Poole testified that a "typical business model" was "for a
content company to pay Netscape to promote their product,
their content." Poole, 2/8/99pm, at 5:8-11.
- Dr. Warren-Boulton testified that the prohibition on ICPs'
"paying compensation to 'Other Browsers'" inhibited "the
continued development of Netscape's browser by depriving
Netscape of important ICP partners and revenues from
promoting those ICPs." Warren-Boulton Dir. ¶ 117.
277.1.3. Third, Microsoft's requirement that ICPs adopt Microsoft-specific technologies and create content viewable by other browsers with "acceptable
degradation" was designed to bias users' choice of browser toward Internet Explorer and
away from other browsers that employed technologies Microsoft did not control.
- Dr. Warren-Boulton testified that, "by conditioning access to the
Windows desktop on ICPs agreeing to use Windows-specific
technologies in their web sites, Microsoft biases consumers' choice
of browsers toward Internet Explorer and away from browsers that
do no support their Microsoft-controlled technologies." Warren-Boulton Dir. ¶ 119.
277.2. Microsoft specifically anticipated that its agreements would have
these effects and facilitate its objective of gaining substantial browser share:
- Poole conceded that a primary purpose of the ICP agreements was
to gain browser usage share. Poole, 2/8/99am, at 67:1-4, 67:18 -
68:1, 66:15-25.
- Harris testified in regard to the Microsoft/Intuit agreement that: "At
the time I negotiated the agreement, it was generally understood by
me, and the industry generally, that Microsoft viewed increasing
browser share as an important goal and that Microsoft sought to
surpass Netscape in browser share. Microsoft executives had
spoken publicly about the importance Microsoft placed on increasing
browser share. At about that time, Bill Gates said in a public forum
at which Intuit was represented, that there was one and only one
measure of Microsoft Corporation's success in the coming year:
'browser share.' Moreover, during the course of the negotiations,
Microsoft's representatives asked about the number of copies of
Intuit's software that Intuit expected to distribute with Internet
Explorer. My understanding was that Microsoft wanted to estimate
the impact the agreement between Intuit and Microsoft would have
on Microsoft's browser market share. For all of these reasons, I
believed that the restrictions placed on Intuit in the agreement
between Intuit and Microsoft related to Microsoft's goal of increasing
its share of the browser market at Netscape's expense." Harris Dir.
¶ 69.
- An October 18, 1996, internal Microsoft document entitled "IE 4.0
Business Model," identified "Increase IE market share" as the
number one priority of the Internet Explorer 4.0 business pan. GX
852, at MS6 6005670; Poole 2/8/99pm, at 43:1-23 (testifying that
one of the purposes of the ICP agreements was to increase
Microsoft's browser usage share).
- In a February 11, 1997 "Internet Client & Collaboration: 3 Year
Business Outlook" presentation, Microsoft explained that one of its
"competitive levers" is its "integration with MS products," noting that
"Shell integration, autoupdate, distribution" are its "primary 1997
weapons," as well as its "distribution and partnerships." GX 413, at
MS6 6003212.
- Brad Chase, upon learning that Microsoft had entered into an ICP
agreement with ESPN, wrote that it was "a tremendous deal"
because ESPN's "Sportszone is one of the top few sites on the
internet." Chase specifically noted that the Sportszone web site
must use Microsoft-specific technology, including ActiveX and "key
IE 3 html extensions." GX 862.
277.3. Microsoft's specific purpose in seeking to gain browser share
through its ICP agreements was to combat the threat that other browsers -- specifically
Netscape -- posed to its operating system monopoly; indeed, the exclusionary
restrictions were primarily targeted at Netscape.
- Poole conceded that "one of the reasons that" Microsoft was "trying
to gain browser-usage share was in order to combat what" Microsoft
"viewed as the platform threat" that Netscape posed. Poole,
2/8/99am, at 18:12-15.
- Harris testified that Microsoft's William Poole told him that Bill Gates
had decreed that Platinum partners must cease working with
Netscape. Harris Dir. ¶ 68 ("Mr. Gates had mandated that all
preferred participants on the Active Desktop must agree to cease
working with Netscape as a precondition of that participation").
- Poole conceded that he told Intuit that Gates mandated that all
preferred participants on the Active Desktop were required to agree
to certain restrictions on working with Netscape as a condition of that
participation, and that these restrictions were non-negotiable. Poole,
2/8/99 am, at 53:20 - 54:1.
- Steve Wadsworth of Disney Online stated: "In negotiating the
contract with Microsoft, it was very clear that the `Other Browser'
Microsoft was concerned about was Netscape Netcaster." GX 776,
at ¶¶ 7-8 (Wadsworth Decl. 4/23/98).
- Poole conceded that, although the agreements were framed in terms
of "Other Browsers," "Netscape was the primary target" and any
other existing browser competitors were "minor." Poole, 2/8/98am,
at 69:5-11, 46:17-20.
278. Mr. Poole's testimony -- that Microsoft decided to give away placement on
the Channel Bar because the technology was "unproven" and not in order to gain
browser share to thwart Netscape (Poole, 2/8/99am, at 25:20 - 26:23) -- cannot be
reconciled with the evidence and is not credible.
- Contemporaneous Microsoft documents show that Microsoft decided to use
the Channel Bar for "strategic barter." See supra ¶ 276.
- Several ICPs offered to pay Microsoft for inclusion in the Channel Bar;
indeed, Microsoft's agreement with Pointcast specified that Microsoft was
to be paid $10 million for the first year. GX 1804, at MS98 0100814
(section 3.3b). Poole's response -- that this agreement reflected "no
actual obligation to pay cash" because PointCast could offset the $10
million if it engaged in various marketing activities (Poole, 2/8/99pm, at
31:1-8, 32:16 - 33.11) does not refute the fact that ICPs were willing to pay
for placement and Microsoft did not need to give this placement away.
- The PointCast contract, signed in November 1996, is also at odds with
Poole's testimony that Microsoft decided to give away Channel Bar
placement before that date. Poole, 2/8/99am, at 29:11 - 30:20.
- Poole's testimony is inconsistent with contemporaneous documents
evidencing Microsoft's judgment that influencing the browser technologies
ICPs employed would increase Internet Explorer's share. Compare Poole
2/8/99pm, at 48:12 - 49:8 with GX 407. Eventually, Poole conceded that
Microsoft's agreements would increase browser usage share even though
ICPs were not themselves significant browser distributors. Poole,
2/8/99pm, at 43:1-23.
279. Microsoft's ICP agreements were not only designed to exclude browser
rivals, but also had the effect of disadvantaging those rivals. Intuit, for example, agreed
to restrict its distribution and promotion of Netscape and other browsers not only to users
who learned about Intuit through Microsoft's Channel Bar, but also in other channels
through which Intuit sells software. As a result, the number of copies of Netscape
distributed by Intuit decreased dramatically.
- Harris testified that "a total of over five million copies of Netscape
Navigator were distributed with the 1997 versions of Quicken, Turbo
Tax, and QuickBooks, but over five million copies of Internet
Explorer were distributed with the 1998 versions of those products."
Harris Dir. ¶ 80.
- Absent Microsoft's exclusionary restrictions, Intuit likely would have
offered consumers Netscape Navigator in some "lightly integrated"
way. Harris, 1/4/99pm, at 67:11 - 68:7; Harris Dir. ¶¶ 76-77; Harris,
1/5/99am, at 45:3-9 ("Netscape, first of all, had been a good partner
of ours for some time and we wanted to be able to continue to deal
with them. And, in a more self-interested manner, Netscape had
access to a great deal of web-based traffic, which is, of course, the
life blood to any Internet business. And we wanted and continue to
want to work with Netscape to secure some of that traffic to our
sites").
- Dr. Warren-Boulton testified that the anticompetitive impact of
Microsoft's ICP agreements are "illustrated by Netscape's
experience with Intuit." Warren-Boulton Dir. ¶¶ 117-18.
b. Microsoft's contention that its ICP agreements were not
capable of causing significant anticompetitive effects is
unfounded
280. Microsoft witnesses asserted that its agreements with ICPs were not
capable of causing significant anticompetitive effects. Microsoft's contentions are
unpersuasive.
280.1. First, Microsoft witnesses argue the Channel Bar did not turn out to
be as popular as Microsoft and ICPs anticipated (Poole Dir. ¶¶ 54, 68). But the fact that
the Channel Bar turned out not to be a popular feature of Windows, and thus that
placement in the Channel Bar turned out not to be valuable to the ICPs, does not detract
from the fact that the promise of placement within Windows induced ICPs to agree to
restrictions that harmed Netscape.
- Professor Franklin Fisher testified that "damage to the competitive
process has already occured" even though the ICP agreements may
not have been commercially successful. Fisher Dir. ¶ 197.
- Dr. Warren-Boulton testified "that the ICPs at the time that Microsoft
offered the active desktop thought that being on the active desktop
was just absolutely invaluable, or almost of incalculable value ."
Even though being placed on the active desktop "wasn't such a
great deal for them after all," "a lot of people who signed contracts or
agreements entered contracts with Microsoft to -- agreed to
exclusivity in return for being on the active desktop." Warren-Boulton, 11/30/98pm, at 32:16 - 33.6.
280.2. Second, Dean Schmalensee's arguments that the ICP agreements
did not adversely affect Microsoft's browser rivals are not sound and miss the point.
280.2.1. Because of scrutiny that preceded this litigation, Microsoft
waived some of its restrictive agreements in April 1998. Dean Schmalensee thus argues
that, even "if the contract terms with the ICPs had been anticompetitive, they were in
force too short a time to have a significant impact" Schmalensee Dir. ¶ 470. But this
contention ignores both Microsoft's expectations and the dramatic increase in Internet
Explorer's usage share during the very period in which the restrictions were in place.
- Poole conceded that the ICP agreements would have continued
absent the pre-trial scrutiny on them. Poole, 2/8/99am, at 74:21-24.
- See infra Part VII.A.3; ¶ 369 (discussing the increasing in Internet
Explorer's usage share during the period of the restrictions).
280.2.2. Dean Schmalensee argues that the ICP agreements did
not affect the browser with which users viewed web sites covered by the agreements.
Schmalensee Dir. ¶¶471-76. He also argues that the ICP agreements were not
reasonably capable of causing anticompetitive effects because they affected only a small
fraction of web sites and because Microsoft had only a tiny share of the browser market
when those agreements were instituted. Schmalensee Dir. ¶ 469. Both these arguments
ignore the contemporaneous evidence that Microsoft entered into these agreements in
the context of its other anticompetitive actions and with the expectation that they would
impede its rivals and for the purpose of protecting its dominant position in operating
systems.
- Professor Fisher, on the subject of Microsoft's ICP
agreements, testified: "Regardless of whether such
provisions would be anticompetitive in themselves if put in
place by a company with a small share of operating systems,
they are certainly anti-competitive when Microsoft uses them
to protect its dominant position in operating systems." Fisher
Dir. ¶ 199.
280.3. Third, Microsoft asserted that the restrictions in the ICP agreements
had no bearing on Intuit's decision to distribute Internet Explorer exclusively because
Intuit intended to use and distribute Internet Explorer irrespective of the ICP agreement
(Poole Dir. ¶ 130). Mr. Poole's argument misses the point because Intuit could have
licensed Internet Explorer wholly apart from the ICP agreement. Intuit agreed to
Microsoft's restrictions on its dealings with Netscape in order to obtain placement on the
Windows desktop.
- Harris testified that "the most likely outcome" in the absence of the
restrictions "would probably have been to distribute both" browsers.
Harris, 1/5/99am, at 49:23 - 50:11. Although Harris also testified
that Intuit likely would have bundled Internet Explorer absent the
restrictions, as he explained, that decision "was independent" of
Intuit's decision to agree to Microsoft's demand that it not distribute
Navigator. Harris, 1/5/99am, at 51:9-14.
- Indeed, Harris testified that both browsers "had certain advantages
and disadvantages." Harris Dir. ¶ 78. In contrast, "only Microsoft
could offer placement on the Windows Desktop" and thus Harris felt
Intuit "had no practical alternative but to agree to Microsoft's
restrictions in order to gain access to the desktop. Harris Dir. ¶ 79.
"Had we decided to bundle the IE browser," Harris testified, "there's
no reason that we couldn't have continued to work with Netscape in
any number of ways, bundling their browser as well." Harris, 1/5/99,
at 51:9-14. Harris further explained that Intuit did not immediately
ship Netscape with Quicken 99 because Quicken 99 was finalized
during the period in which Microsoft's restrictions were in place.
Harris, 1/4/99pm, at 71:18 - 72:13.
- Poole conceded that Intuit could have obtained Internet Explorer
without agreeing to Microsoft's ICP restrictions. Poole, 2/8/99am, at
43:19 - 44:16.
4. Microsoft's ICP agreements lacked justification
281. Microsoft witnesses advanced a number of justifications for the ICP
agreements. The nature of the agreements and the contemporaneous evidence
demonstrate that the these justifications are pretextual. Microsoft's agreements can be
explained only as part of an anticompetitive strategy designed to disable the threat
Netscape posed to Microsoft's operating system monopoly.
281.1. First, Microsoft witnesses, in particular Mr. Poole and Dean
Schmalensee, assert that the ICP agreements were simply "commonplace" "cross
marketing agreements." Poole, 2/8/99am, at 55:23 - 56:7; Schmalensee Dir. ¶ 469. This
assertion -- and Poole's testimony -- is not credible. Microsoft's agreements impose
restrictions that, Poole ultimately conceded on cross-examination, have no known
analogue in the industry and, in fact, have no conceivable purpose except to blunt
Netscape.
281.1.1. Microsoft's agreements generally prohibit ICPs from paying
"Other Browsers" to promote their content. By virtue of Microsoft's narrow definition of
"Other Browser," this is a unique, naked restriction intended only to harm Netscape.
- Poole conceded that he was "not aware of any analogous
business relationship beyond these ICP deals within
Microsoft." Poole, 2/8/99pm, at 14:8-16.
- Dr. Warren-Boulton testified that "one aspect of the ICP
agreements -- the prohibition on ICPs" payment from
Netscape for promotional services -- plainly serves no
legitimate purpose. The only reason for that restriction is to
impede Netscape." Warren-Boulton Dir. ¶ 184; Warren-Boulton, 11/30/99pm, at 39:19 - 40:13 ("it is just very difficult
to see any reason why Microsoft would want to impose that
restriction in terms of any sort of efficiency story").
- Professor Fisher testified that "Microsoft sought to further
deprive Netscape of revenue by inducing ICPs to agree not to
pay Netscape for carrying or promoting the ICPs' content or
logos." Fisher further testified that the provision that
mandates this "can have no purpose other than that of
damaging those browser suppliers." Fisher Dir. ¶¶ 134, 199.
281.1.2. Microsoft's agreement with Intuit also lacks any industry
analogue.
- Poole conceded he was unaware of any other agreements in
which "Microsoft agrees with one independent software
vendor that Microsoft will give that independent software
vendor things of substantial value, like placement on the
desktop, in return for that ISV's agreement not to promote or
market the products of another ISV." Poole, 2/8/99am, at
58:23 - 59:7.
281.1.3. The Netscape agreements identified by Microsoft (Poole
Dir. ¶ 61; Schmalensee Dir. ¶ 469) are very different from Microsoft's ICP agreements.
- Netscape's agreements did not prohibit ICPs from receiving
value from Microsoft for promoting Microsoft's content. Poole,
2/8/99pm, at 41:8-25 (agreeing that he did not mean to imply
that "Netscape's agreements had the same kind of
restrictions as Microsoft's agreements").
- In contrast to Microsoft's ICP agreements, the AOL/Netscape
agreement contains no prohibition on, for example, AOL
dealing with Microsoft in areas other than instant messenger.
GX 1256.
281.2. Second, Microsoft argues that its ICP agreements are justified as a
means of "associating" the Internet Explorer brand and technology with well-established
other brands (Poole Dir. ¶¶ 62, 64-68, 84; Poole 2/9/99am, at 24:4 - 25:13). But
Microsoft's "brand association" argument is also pretextual. It is belied by both
Microsoft's conduct and the ICP agreements themselves. When these agreements are
read as a whole it is apparent that they permit ICPs to associate with other brands,
including other browser companies, and are designed, instead, simply to harm Netscape
by inhibiting its ability to receive compensation, promotion, or distribution from leading
ICPs.
- The agreements permit ICPs to deal with Netscape (Poole Dir. ¶ 66),
which is inconsistent with Microsoft's brand association rationale, but
they prohibit ICPs from paying Netscape; this prohibition harms
Netscape and is unrelated to brand association. Poole Dir. ¶ 84.
- The ICP restrictions apply only to "Other Browsers," which generally
is defined to include only Netscape and one other browser and to
exclude client software developed by PointCast, Marimba, AOL, and
Chat or Pager clients, e.g., GX 1156, at INT 00003 (section 1.23).
281.3. Third, similarly pretextual is Microsoft's contention that its
agreements were intended merely to "showcase new technologies," such as ActiveX and
Dynamic HTML (Poole Dir. ¶¶ 62, 90-100; Poole, 2/9/99am, at 24:4 - 25:13). That
rationale has nothing to do with Microsoft's insistence that ICPs such as Intuit not
distribute Netscape apart from the Channel Bar or its prohibition on ICPs' paying
Netscape for promotion. It is inconsistent with Microsoft's conduct and other provisions
of the ICP agreements as well.
- Poole conceded that Microsoft "did almost nothing to enforce the extent to
which its Platinum ICP partners were taking advantage of these
technologies" and therefore showcasing them. Poole Dir. ¶ 93. Although
Poole testified that the "differentiated content" provisions, which required
ICPs to implement Internet Explorer-specific technologies and to create
content that could not be viewed well with Netscape, were central to
Microsoft's brand associate justification (Poole Dir. ¶ 84), Microsoft did not
enforce those provisions. Poole, 2/9/99am, at 54:15 - 55:8.
- Microsoft did not merely require ICPs to "showcase" Microsoft
technologies; it extracted agreements that such content could not work well
with other browsers -- in the words of the agreement, with "acceptable
degradation." GX 1156, at INT00018. Poole never explained why
"showcasing" Microsoft technologies required "degrad[ing]" the experience
of users employing other browsers. Harris Dir. ¶ 73 (Microsoft/Intuit ICP
agreement required Intuit to deploy Microsoft-specific technologies such as
Microsoft's Channel Definition Format, and Dynamic HTML, which
conflicted with Intuit's general policy of making its Internet content broadly
accessible).
281.4. Fourth, Microsoft argued that its agreements were designed to
increase user interest in the Internet (Poole Dir. ¶¶ 64-68; Poole, 2/9/99am, at 24:4 -
25:13), but that justification finds no support in contemporaneous documents and is
completely at odds with the evidence. If Microsoft were interested in increasing user
interest in the Internet, it would permit, rather than restrict, ICPs ability to promote,
market, and distribute its content and other browsers as widely as possible.
- Professor Fisher testified that, if "Microsoft were really interested in
selling windows, it wouldn't have any interest" in restricting Netscape
"since people who wanted to use the Netscape browser with
Windows would be happier" with Windows if Netscape was included.
Fisher, 6/1/99am, at 66:8-25.
- Dr. Warren-Boulton testified: "Microsoft has a legitimate interest in
ensuring that Windows users are able to acquire high quality
browsers at low prices, because that would increase the demand for
Microsoft's operating system. But even if achieving this objective
were furthered by Microsoft's decision to offer a quality browser
product, its further efforts to increase IE's share by excluding
Netscape and making it more difficult for users to obtain Netscape's
browser could only reduce the value of its operating system to
consumers." Warren-Boulton Dir. ¶¶ 187, 189.
281.5. Fifth, Microsoft argues that ICPs entered into the exclusionary
agreements only because "Microsoft offered them more value than Netscape was willing
or able to offer" (Schmalensee Dir. ¶ 483). That argument, however, is simply another
way of saying that Microsoft offered a very large bribe (that is, not charging for Channel
Bar placement or the pertinent Microsoft technology) to induce these firms to limit their
dealings with Netscape. It cannot justify the agreements' exclusionary restrictions
282. Microsoft's justifications for its restrictions are pretextual not only on their
own terms, but also because Microsoft's expensive effort to gain browser usage share,
of which its ICP agreements were a component, can be explained only as part of an
effort to preserve its operating system monopoly.
- Professor Fisher testified that Internet Explorer "was a product which
Microsoft not only gave away for free, but basically bribed people to take.
They gave them preferred places on the Desktop for which they could have
charged. But beyond that, they also spent hundreds of millions of dollars
on the development of this no-revenue product, and then they gave away
the technology. That is not a profitable act, except for the protection of the
operating system's monopoly." Fisher, 6/1/99am, at 39:22 - 40:4.
- Dr. Warren-Boulton testified that: "The available evidence indicates that
Microsoft pursued " its browser-related practices "for the purpose of
preserving its Windows operating system monopoly and gaining monopoly
power in the browser market, and pursued them without regard to whether
would have been were profitable in their own right." Warren-Boulton Dir. ¶
185.
F. Microsoft entered into exclusionary agreements with other firms that
restricted their ability to promote, support, and distribute non-Microsoft browsers
283. Microsoft also induced other firms to enter into exclusionary agreements
designed to increase its share of the browser market and protect its operating system
monopoly.
1. Microsoft used its leverage over office productivity suites to coerce
Apple to enter into an exclusionary agreement that favored Internet
Explorer and severely disadvantaged browser rivals
284. Microsoft, by threatening to cease development of its Office for Macintosh
productivity suite, coerced Apple into making Internet Explorer the default browser on all
Macintosh operating systems and to disadvantage competing browsers.
a. To facilitate winning the browser war, Microsoft sought to
obtain default status for Internet Explorer on the Macintosh
285. Microsoft made inducing Apple to favor Internet Explorer and disfavor
competitive browsers a key goal.
285.1. Apple develops and markets computer systems. Apple's operating
system for its desktop computers is known as the Macintosh Operating System or the
"Mac OS."
- Tevanian Dir. ¶¶ 7, 10.
285.2. Although the Macintosh has a tiny share of PC sales compared to
Intel-compatible personal computers, Macintosh users are particulary active users of the
Internet.
- Avadis Tevanian, the Senior Vice President of Software
Engineering at Apple Computer, testified that "given the
demographics of our users, Macintosh users were more likely to go
on the Internet, and so the percentages don't actually translate."
Tevanian, 11/9/98am, at 58:18-59:2.
- A January 27, 1998 Internet Explorer 5 for Macintosh draft
transition plan stated that "it's becoming increasingly apparent that
the cross-platform browsers directly affect overall IE market share
exponentially." GX 370.
285.3. Microsoft accordingly sought to obtain preferences for Internet
Explorer on the Macintosh.
285.3.1. Although Apple had bundled Internet Explorer with its
operating system, Mac OS 8.0, since April 1997, Microsoft remained dissatisfied;
indeed, Microsoft asserted, failure to make Internet Explorer the default browser
breached an oral agreement between the two firms.
- Tevanian testified that "Gates was unhappy that, although
we were including it with Mac OS 8, we were not including it
as the default." Tevanian, 11/9/98am, at 25:16 - 26:5; see
also Tevanian Dir. ¶¶ 30-32. Then-Apple CEO Gil Amelio
wrote to Gates attempting to appease him on this issue. GX
1053 (7/3/97 letter from Amelio to Gates: "I'd like to
comment on the inclusion of Internet Explorer with our
release of Mac OS 8. I know that this is a source of great
irritation to you.").
285.3.2. Gates made increasing Internet Explorer share on
Apple's platform a "key" goal, and identified making Internet Explorer the "standard" or
default browser as the means of accomplishing that.
- As early as June 23, 1996, Gates stated his goals for the
Apple relationship: "I have 2 key goals in investing in the
Apple relationship -- 1) Maintain our applications share on
the platform and 2) See if we can get them to embrace
Internet Explorer in some way," and proposed to Apple that
Apple "ship IE on all their systems as the standard browser."
GX 260 (6/23/96 Gates e-mail to Microsoft executives).
b. To accomplish its objective, Microsoft used its leverage over
Apple, specifically Apple's dependence on Microsoft's Office
productivity suite, to coerce Apple to enter into an
exclusionary agreement that favored Internet Explorer and
disfavored rivals
286. To accomplish its objective of increasing its browser share with Macintosh
users, Microsoft coerced Apple, by threatening to cancel continued support for
Microsoft's Office productivity suite for the Macintosh, into agreeing to favor Internet
Explorer and restrict its promotion, distribution, and support of Microsoft's browser
rivals.
(1) Microsoft's Office productivity suite ("Mac Office") was
and remains vital to Apple's business
287. Microsoft's continued support for its Office productivity suite for the
Macintosh is vital to Apple's computer business.
287.1. Microsoft Office is an "office productivity suite" that includes word
processing, spreadsheet, and presentation programs. Microsoft's version of Office that
runs on the Macintosh is known as Microsoft's Office for Macintosh or "Mac Office."
- Tevanian Dir. ¶¶ 34-35.
287.2. Because Office, including Mac Office, is the dominant product in
its critical category, Microsoft's continued support for Office on the Macintosh is vital to
Apple's business.
287.2.1. Office, including its Macintosh version, is the dominant
office productivity suite.
- An internal Apple strategy presentation states: "Excel owns
100% of the Mac productivity spreadsheet market" and
"Word has 92 % of the word processing productivity market."
GX 1036, at MAC 10347; Tevanian, 11/9/98am, at 30:17-21.
- Tevanian Dir. ¶¶ 34-35.
287.2.2. Both Apple and Microsoft recognize that Microsoft's
continued support for Mac Office is vital to the Macintosh's survival.
- Tevanian testified that for most Apple customers, "Office was
a critical application that they depended on. And if Office
were no longer to be available and if new versions were not
to be produced, then we would basically lose our
customers." Tevanian, 11/9/98am, at 28:8-15.; Tevanian
Dir. ¶ 35.
- Microsoft documents recognize the significant leverage Mac
Office gave Microsoft over Apple. E.g., GX 267 (1/21/98
Bradford e-mail to Gates and others, explaining that
"MacOffice is the biggest Apple carrot. From a browser, mail
client and java perspective, Apple has other options."); GX
263 (6/27/97 Waldman e-mail to Gates et al.: "The threat to
cancel Mac Office 97 is certainly the strongest bargaining
point we have, as doing so will do a great deal of harm to
Apple immediately.").
287.2.3. In the summer of 1997, Microsoft had substantially
completed development work for a new version of Mac Office (Mac Office 97, later Mac
Office 1998). Failure to release or support this product would have had a devastating
effect on Apple's business.
- Tevanian testified that without good versions of Office
available for the Macintosh: "There's the first effect, which is
customers can't get Office, which is, of course, devastating
in and of itself. The second effect, which makes the problem
even worse, is if other software developers, who are
developing other applications, which are also important to
customers, see that Microsoft is pulling away and Office is
pulling away, then they would pull away themselves because
they would see the whole platform as just dying out."
Tevanian, 11/9/98am, at 29:13 - 29:21; Tevanian,
11/9/98am, at 27:17 - 28:25; Tevanian Dir. ¶ 35 ("Because
Microsoft Office completely dominates the market for office
productivity suites, it is critical to the commercial viability of
the Apple Mac OS to have a version of Microsoft Office that
can run on the Mac OS operating system."); GX 1036, at
MAC10347 ("As the largest Macintosh developer,
[Microsoft's] abandonment of the platform would have
serious ramifications."); Tevanian Dir. ¶ 15 (explaining how
the commercial viability of an operating system is dependent
on the availability of well accepted, broadly used application
programs).
- Tevanian further testified: "Because the prior Office for
Macintosh version was poor in terms of performance and
stability, Apple computer users were especially anxious to
obtain a new and improved version of Microsoft Office."
Tevanian Dir. ¶ 36; Tevanian, 11/9/98am, at 31:24-32:5
(explaining that Apple customers "did not at the time have
access to a good version of Office; the previous version of
Office was just really bad"); Tevanian, 11/4/98pm, 23:4-19,
25:7 - 26:4 (same); GX 263 (6/27/97 Gates e-mail
acknowledging Microsoft's neglect of the Mac business).
- Tevanian testified that Microsoft had substantially completed
development work on the MacOffice update by June 1997
and had shown Apple personnel a preliminary beta.
Tevanian Direct ¶ 37; see also GX 263 (6/23/97 Waldman e-mail to Gates describing work accomplished).
(2) Microsoft used the threat of stopping
development of and support for Mac Office to
extract Apple's agreement to favor Internet
Explorer and to restrict its distribution of rivals
288. Recognizing the importance of updating and supporting Mac Office,
Microsoft used the threat of canceling such support to coerce Apple into entering an
exclusionary agreement.
288.1. Through the spring and summer of 1997, Apple and Microsoft
were engaged in a wide-ranging discussion concerning the settlement of, among other
things, a patent dispute, and the issue of how Apple bundled Internet Explorer.
- Tevanian testified that in the spring of 1997, Microsoft insisted on
merging negotiations to resolve the disputes over patents,
browsers, and other issues to seek a comprehensive solution.
Tevanian Direct ¶¶ 33-34; see also GX 1046 (4/27/97 e-mail from
Microsoft's Maffei to Apple's Scalise containing 12 point proposal).
288.2. Even though it had substantially completed development for its
new version of Mac Office, Microsoft threatened to cease supporting Mac Office and
other Microsoft applications ported to the Macintosh unless Apple agreed, among other
things, to make Internet Explorer the default browser for the Mac OS and to restrict
Apple's distribution and promotion of non-Microsoft browsers.
288.2.1. Microsoft knew that Apple needed the new release of Mac
Office and Microsoft's commitment to support it.
- In a June 27, 1997, e-mail to Bill Gates, Ben Waldman, the
Microsoft manager in charge of Mac Office 97, stated, "The
threat to cancel Mac Office 97 is certainly the strongest
bargaining point we have, as doing so will do a great deal of
harm to Apple immediately." GX 263 (6/27/98 Waldman e-mail to Gates).
- Tevanian testified that "Microsoft was aware that Apple
desperately needed to maintain support for Microsoft Office
for Macintosh. In addition, there was a strong demand in the
market for Microsoft Office 98 for Macintosh based on the
reports of its development." Tevanian Dir. ¶ 36; Fisher Dir. ¶
153 ("As Microsoft knew, withdrawal of support for this crucial
application would have a devastating effect on the viability of
the Macintosh operating system.").
- Tevanian further testified that Microsoft's refusal to release
the new version of Office, even if the outdated version
remained available, would have had the net effect of
canceling the Mac Office program because applications that
are not updated fall behind. Tevanian, 11/4/98pm, at 23:4 -
24:8; Tevanian, 11/9/98am, at 32:11-16 ("Generally
speaking, applications need to be updated every once in a
while to take advantage of new features in the operating
system or, in some cases, to deal with just other issues that
may come up. So any -- any application that is in use by
customers will often need to be upgraded on a frequent
basis.") .
288.2.2. Microsoft used the threat of withdrawing support for Office
to coerce Apple's acceptance of its exclusionary terms.
- Tevanian testified: "In mid-May 1997, Microsoft's negotiators
told Apple's negotiating team that Microsoft would remove its
support for Microsoft application programs for the Mac OS
operating system if Apple refused to resolve the disputes
concerning the patents, the browsers and other aspects of
the companies' relationship on terms acceptable to
Microsoft." Tevanian Dir. ¶ 34.
- Ben Waldman, who was in charge of Mac Office wrote to
Gates: "The pace of our discussions with Apple as well as
their recent unsatisfactory response have certainly frustrated
a lot of people at Microsoft. The threat to cancel Mac Office
97 is certainly the strongest bargaining point we have, as
doing so will do a great deal of harm to Apple immediately. I
also believe that Apple is taking this threat pretty seriously . .
. ." GX 263 (2/27/97 Waldman e-mail to Gates) (emphasis
added). Gates' response was to ask Waldman when the
product would ship and "Can we avoid Apple knowing how
far along we are for the next 30 days?" GX 263 (6/27/98
Gates to Waldman e-mail).
- Waldman later assured Microsoft CFO Greg Maffei that he
has done his part to convey the threat: "Did you convey your
concerns to Apple (i.e. that you were recommending
cancellation to Bill), and did they have a response? When I
spoke with [Apple's] Gable yesterday, I told him that we had
almost given up on any hope of progress, and that you
would be recommending to Bill that we cancel Mac Office
97. . . . In any case, he took notes, and I think he got the
message that we would, in fact, cancel. (Each time I talked
about cooperating, I qualified it with `if we continue.'" GX
264 (6/27/97 Waldman to Maffei, Gates e-mail).
- Paul Maritz admitted that Microsoft threatened to cancel
MacOffice in connection with the multi-part negotiations
between the two firms. Maritz, 1/28/99pm, at 27:13 - 28:4.
Gates' testimony that Microsoft never threatened Apple that
it would cancel development of Mac Office (Gates Dep.,
played 11/2/98pm, at 25:9 - 26:3, 29:8 - 38:15) is not
credible in light of this admission.
288.3. Because of Apple's dependence on Mac Office, Apple agreed, in
an August 5, 1997, "Technology Agreement," to Microsoft's restrictive terms.
- GX 1167 (8/5/97 Technology Agreement Between Apple and
Microsoft).
- Tevanian testified that, but for Microsoft's threat to withdraw
support for Mac Office, Apple would not have agreed to these
restrictive terms. Tevanian Dir. ¶¶ 38, 42.
- Apple's Fred Anderson explained to Jim Barksdale that Apple felt it
had no choice but to give in to Microsoft's threat: "Apple needed to
insure that Microsoft would continue to provide Microsoft Office for
Mac or we were dead. [MS was] threatening to abandon Mac.
Trading card was making Internet Explorer default browser." GX
595; Barksdale, 10/27/98am, at 18:11 - 19:15.
289. Apple thus agreed, in exchange for Microsoft's commitment to provide an
up-to-date version of Mac Office and other support for 5 years, to a number of
restrictions on its ability to distribute and promote non-Microsoft browsers.
289.1. First, Apple agreed to bundle Internet Explorer as the "default"
browser on all Macintosh computers and the Mac OS operating system for five years.
Default status means that Internet Explorer will automatically be launched when the
user invokes the "Browse-the-Internet" icon preinstalled on the Mac OS desktop.
- GX 1167 § 3.1 (requiring Apple to bundle Internet Explorer as the
"default selection" for Macintosh with all system software releases
of the Mac OS as long as Microsoft supports Office for the Mac);
Tevanian Dir. ¶ 38; Tevanian, 11/9/98am, at 38:25 - 40:11, 41:11-25.
289.2. Second, if Apple bundles another browser, it cannot be placed on
the desktop and must be placed in a folder, where it is more difficult for users to find.
- GX 1167 §3.1.
- Tevanian explained that Section 3.1 of the Agreement requires that
"any other browsers that we would bundle would be somewhere in
the operating system, not on the desktop, so the user would have
to go find them to actually -- to find out where they were and see
that they were there." Tevanian, 11/9/98am, at 40:12 - 41:5, 42:1-4.
289.3. Third, Microsoft insisted that Apple not include Netscape Navigator
in the "default" installation for the Mac OS 8.5 upgrade. Unless a user is aware of, and
makes the effort to select, a "custom" installation option, Navigator will not be installed
on users' computer hard drive when the upgrade product is loaded.
- Tevanian explained that, because of Microsoft's insistence that
Netscape not be shipped as the default browser, the Mac OS 8.5
operating system upgrade product does not include Navigator in
default installation of the operating system upgrade. This means
that Navigator is not installed on the computer hard drive during the
default installation, which is the type of installation most users
choose. Therefore, users who use the default installation will be
unable to access Navigator on the computer without retrieving the
CD-ROM and doing a custom install. Tevanian, 11/4/98pm, at 68:7
- 69:8; Tevanian, 11/9/98pm, at 44:18 - 45:13.
- On December 5, 1997, Microsoft's Don Bradford wrote to Gates:
"Navigator will be available on companion media but users must
use 'Custom Install' to get it." GX 266.
289.4. Fourth, the agreement requires that Apple "will not be proactive or
initiate actions to encourage users to swap out of Internet Explorer for Macintosh."
Apple and Microsoft have interpreted this to prohibit Apple from promoting competing
browsers.
- GX 1167 § 3.1.
- GX 266 (12/5/97 Bradford e-mail to Gates et al.: "all of Apple's . . .
marketing collateral will be updated by" the Summer of 1998 "to use
only IE").
- Tevanian testified that, "our interpretation of this sentence has been
that we can't promote browsers other than Internet Explorer. And,
in fact, that has been represented to us as being the case, by
Microsoft, who has actually been policing how we do our marketing
and what we say in our speeches and actually been critical of us in
certain cases where we have mentioned other browsers. Q: And
you are referring to the time that Mr. Jobs said at the MacWorld
Expo that people could switch out or switch Netscape Navigator as
the default browser in place of Internet Explorer? A: That would
have been one of them." Tevanian, 11/4/98pm, at 61:6 - 62:3; see
also Tevanian, 11/9/98am, at 42:10-22; Tevanian Dir. ¶ 41.
289.5. Fifth, Apple agreed to "encourage its employees to use Microsoft
Internet Explorer for Macintosh for all Apple-sponsored events and will not promote
another browser to its employees." This agreement prevented Apple from using
competitive browsers in demonstrations at trade shows and other public events.
- GX 1167 § 3.2.
- After the agreement was signed, Apple's Guerrino De Luca
instructed Apple employees that there were "several cases in which
we currently, and understandably, use other browsers in our
demos, events, etc. We must limit or avoid these cases." GX
1105.
- Tevanian, 11/9/98am, at 43:1 - 44:17 (explaining how Apple has
instructed its employees not to use other browsers in
demonstrations and how Microsoft has policed this provision).
289.6. Sixth, the Technology Agreement contains other provisions that
favor Internet Explorer and severely restrict competitive opportunities for competing
browsers, including (i) requiring that Microsoft's Internet Explorer logo be displayed on
"all Apple-controlled web pages where any browser logo is displayed" and (ii) granting
Microsoft the right of first refusal to supply the default browser for any new operating
system that Apple develops during the term of the agreement. The agreement also
commits the two parties to work together on Java development.
- GX 1167 §§ 3.2, 3.7 & 3.8.
290. Microsoft policed Apple's marketing activities and promotions in order to
enforce its agreement.
- Tevanian testified that, while at times Apple employees have had to
demonstrate Netscape's browser at trade shows for technical reasons,
"often a Microsoft person will see that happening and they will contact us
and let us know that they don't think it's appropriate." As a result,
according to Tevanian, "even though, in some cases, we need to mention
other browsers, we always try to make sure that we have the right
perspective with the promotion of IE versus the other browsers."
Tevanian, 11/9/98am, at 43:15 - 44:17; Tevanian, 11/4/98pm, at 61:6 -
62:3.
- Microsoft's Ben Waldman wrote to, among others, Gates on February 13,
1998, that "the spirit" of its agreement with Apple "is that apple should be
using" Internet Explorer "everywhere, and if they don't do it, then we can
use Office as a club." GX 268.
c. The restrictions Microsoft coerced Apple into accepting had
significant exclusionary effects
291. Microsoft's coercion of Apple to make Internet Explorer the default browser
and to restrict Apple's promotion and support of competing browsers has had, and will
continue to have, significant exclusionary effects.
291.1. First, Microsoft's restrictions -- ensuring Internet Explorer is the
default browser; relegating other browsers to less favorable placement; requiring
Netscape's omission from the default installation for the Mac OS 8.5 upgrade; and
preventing Apple from promoting the usage of other browsers -- raise rivals' costs,
reduce their usage on the Macintosh, and harm consumers.
291.1.1. As Microsoft recognized, most computer users (including
Macintosh users) tend to use browsers that come preinstalled on their machines in an
easy-to-locate place on the desktop. Accordingly, Microsoft has ensured that most
purchasers of the Mac OS will use Internet Explorer and will not use competing
browsers, thus advancing its intended goal of increasing Internet Explorer share on the
Mac OS at the expense of competitive browsers.
- Tevanian
testified that
Apple's
experience
showed that
"customers
seldom
reconfigure
their systems
to change the
default
browser,"
Tevanian Dir.
¶ 41. "Most
of our
customers
just use the
system as we
send it to
them, which
is, they just
double-click
on the . . .
Internet icon
and don't
bother . . . to
change the
default. In
fact, most of
them don't
even know
that they
can."
Tevanian,
11/9/98am,
at 45:19 -
46:23.
- An internal Microsoft planning document recognizes the
same self-evident proposition: "On MacOS we are finally in
the enviable position of being the default browser on the OS.
This is a huge advantage especially in the case of the large
chunk of customers who use whatever browser comes
with/is integrated with their operating system. GX 370, at 2
(emphasis added); Tevanian, 11/9/98am, at 48:4 - 12.
- Tevanian testified: "The Technology Agreement gives
Microsoft significant advantages in its efforts to defeat
Netscape Navigator and gain total control of the browser
market. . . . Making Microsoft's Internet Explorer the default
browser on the Mac OS did not confer any substantive
technical benefit on users, but it would help Internet Explorer
to become the most commonly used browser among Mac
customers." Tevanian Dir. ¶ 41 (emphasis added).
- Tevanian explained that "most users simply won't switch" or
"won't know how to switch, so the natural evolution would be
for more people to use IE. . . . As more and more customers
use IE versus Navigator, it helps Microsoft to consolidate
their position and gain more control of the market."
Tevanian, 11/9/98am, at 49:5-24.
- Professor Fisher testified: "You have a browser on the
desktop, typically IE. You could find another browser, if you
looked for it . . . but the question of whether you can do it
and the question of whether people will do it are two quite
different things. . . . If they've got a satisfactory thing already
there in front of them, they typically won't bother to go and
find something which gives essentially the same service but
requires some difficulty." Fisher, 1/12/99am, at 18:6-17.
- See also supra Part V.D; ¶ 220; infra Part VII.A; ¶ 370.
291.1.2. Although it is possible for users to obtain browsers in
other ways and to change the default browser, the process for doing so is not known to
most users, and is difficult and confusing enough for the typical user that it serves as a
significant impediment to typical users switching default browsers.
- Microsoft introduced a series of screen shots showing the
process for changing the default browser. DX 1783; DX
1784. A review of the screen shots and Tevanian's
testimony shows that the process is sufficiently obscure,
especially in light of the restrictions Microsoft has placed on
Apple, that it serves as a significant impediment to typical
users switching defaults. See generally Tevanian,
11/4/98pm, at 65:12 - 75:23; see also Tevanian, 11/9/98am,
at 45:19 - 46:23 (most users are not aware they can switch
default browsers); see supra ¶ 290.4 (Technology
agreement prohibits action by Apple to encourage users to
switch default browsers).
- Tevanian explained that, "because of Microsoft's insistence,
we do not install Netscape by default, so the user will never
see the Netscape pop up at all. They have to take an extra
action to install Navigator." Tevanian, 11/4/98pm, at 68:17-21. Tevanian elaborated: The "operating system includes
Netscape Navigator on the CD-ROM. If you use the default
installation process, which is what almost every user does, it
is not installed on your hard drive. You need to go back to
your CD-ROM at a later date to specifically install it."
Tevanian, 11/4/98pm, at 70:8-14.
- Tevanian testified that switching browsers on Mac OS 8.5 is
"actually pretty complicated. You have to know what
package it's in and search for it. It's hidden underneath the
package. . . . [W]hen you get the operating system on the
CD, it comes with lots of different components, and if you go
into what we call a custom installation, which, again, most
people don't do because it's an advanced feature, you get a
list of all the things that are installed by default as well as a
list of other things that you can optionally install. So, you
would first have to know to go there, and then you would
have to know which package the browser is in." Tevanian,
11/4/98pm, at 76:2-19.
291.2. Second, absent Microsoft's restrictions, Apple would not have
disfavored competitive browsers and favored Internet Explorer.
- Tevanian testified: "If Microsoft had not exercised its monopoly
power in the office application market by threatening to stop
supporting Office for Macintosh, Apple would not have resolved the
disputes on the terms outlined above." Tevanian Dir. ¶ 42.
- Apple has a fundamental interest in a competitive browser market
so that it is not dependent on one supplier. GX 593; GX 268
(internal Microsoft e-mails discussing Apple's interest in Netscape
remaining a viable competitive alternative).
- Macintosh users met the announcement that Apple had agreed to
make Internet Explorer the default browser with derision, as the
very videoclip introduced by Microsoft showed. DX 1850; played
11/9/98am, at 53:15 - 55:24.
d. Microsoft's coercion of Apple to agree to exclusionary
restrictions lacks justification
292. Microsoft's coercion of Apple to favor Internet Explorer and disfavor rivals
lacks any legitimate business purpose.
292.1. First, Microsoft's threat to abandon a profitable product can be
explained only as a predatory act designed to help facilitate maintenance of Microsoft's
operating system monopoly.
292.1.1. Because Microsoft had already incurred most of the sunk
costs of product development of the new Mac Office by the end of June 1997, canceling
the release of this historically profitable product at that point would have meant the loss
of both those costs and the anticipated profit.
- Tevanian testified: "By June 1997, Microsoft had
substantially completed the development work on Microsoft
Office 98 and, in fact, had shown a preliminary `Beta' version
of the product to some Apple personnel. Although Microsoft
had made a substantial investment getting Office 98 for
Macintosh ready for market, it was willing to risk an outright
loss of that entire investment to force Apple to terms."
Tevanian Dir. ¶ 37.
- Maritz admitted that Microsoft "had a version of Mac Office . .
. ready to ship." Maritz, 1/25/99pm, at 60:9-16; see also GX
263 (6/27/97 Ben Waldman e-mail to Gates detailing the
work already accomplished on the Mac Office 97, including
"thousands of bug fixes," performance improvements, and
many new features, and noting "we are close to shipping").
- Maritz's testimony that he asserted in March 1997 that there
was little reason to continue Mac Office development (Maritz
Dir. ¶ 377) is misleading because it is clear that Mac Office
development was not canceled in March and in fact
continued until the Technology Agreement was signed. GX
263 (6/27/97 Waldman e-mail to Gates).
292.1.2. Canceling Mac Office in 1997 would also have resulted in
substantial losses of goodwill for Microsoft as a result of breaking promises to
customers, poor public relations, and reduced employee morale.
- In asking Gates not to cancel the project, Waldman offered,
among others, the following reasons: "Keeping our word --
customers, press, etc. As you noted in April, we've given
our word to our customers that we will do this product.
We've also been clear on this point with press and analysts.
Pulling out at this point, no matter how high our frustrations
with Apple, would be a blight on our integrity," and "canceling
this now would be devastating" to the developer team. GX
263 (6/27/97 Waldman e-mail to Gates); Maritz Dir. ¶ 377
(arguments in favor of continuing Mac Office development
related to maintaining good will of existing customers and
public relations).
- See also GX 263 (6/27/97 Waldman e-mail to Gates: "our
indecision so far has caused quite a bit of harm").
292.2. Second, there is no technical reason for Microsoft's insistence that
Apple favor its browser over rivals.
- "Making Microsoft's Internet Explorer the default browser on the
Mac OS did not confer any substantive technical benefit on users,
but it would help Internet Explorer to become the most commonly
used browser among Mac customers." Tevanian Dir. ¶ 41.
292.3. Third, Microsoft's contention that the restrictions on which it
insisted were a very minor part of the overall agreement (Maritz Dir. ¶¶ 382-383) and
added only after the other principal elements of the August 5, 1997, deal (including a
patent dispute) were worked out (Maritz Dir. ¶ 385), is neither credible nor important.
292.3.1. The agreement specifically links Microsoft's support for
Office to the exclusionary browser-related terms.
- The Technology Agreement expressly conditions Apple's
obligation to bundle Internet Explorer as the default browser,
as well as the other restrictions on competitive browsers, on
Microsoft's continued development of Office for Macintosh.
GX 1167, §§ 2.1, 3.1 ("For so long as Microsoft is in
compliance with Section 2.1 above and has not elected to
cease future development and releases of Microsoft Office
for Macintosh . . . Apple shall bundle . . . Microsoft's Internet
Explorer . . ."). Both obligations last the five-year term of the
contract. GX 1167, §§ 2.1, 3.1.
- By contrast, neither the patent cross license nor the stock
purchase agreement mention Internet Explorer or Office for
Macintosh. GX 583; Maritz, 1/25/99pm, at 71:25 - 72:17.
292.3.2. Paul Maritz's second-hand testimony -- that issues
regarding Office had been resolved before issues concerning the browser (Maritz Dir. ¶
385) -- is also belied by the contemporaneous documents, which demonstrate that
Microsoft had been insisting on obtaining default status for Internet Explorer since 1996.
- On June 25, 1996, Gates proposed to Apple that Apple
"immediately ship IE on all their systems as the standard
browser" GX 260.
- Microsoft's Maffei proposed to Apple on April 27, 1997, that
Apple offer Internet Explorer as its default browser. GX
1046.
- Maritz conceded that he did not know whether or not the
issue of default browser was raised before July 20, 1997 --
as clearly reflected by GX 1046 -- because his knowledge
was limited to what Maffei had told him about a conversation
that Maffei had with Jobs under a tree in Palo Alto on July
20, 1997. Maritz , 1/25/99pm, at 36:2 - 41:20.
- Maritz later admitted that he knew that one of Gates' and
Maffei's intentions was to make Internet Explorer the default
browser for Apple. Maritz, 1/25/99pm, at 54:1-7; 50:16 -
53:2 (discussing GX 579) (7/1/97 Gates e-mail to Maritz: the
original proposal to Apple had three elements, the first of
which was Microsoft browser getting a privileged position.);
Maritz, 1/25/99pm, at 51:4-16; Maritz, 1/28/99pm, at 29:7-21.
- Maritz conceded that none of the contemporaneous
documents corroborate his testimony -- based on what
Maffei told him -- that the issue of Internet Explorer's status
had been raised only after other issues had been resolved.
Maritz, 1/25/99pm, at 63:5 - 64:24.
- Tevanian testified that Microsoft indicated to Apple in
negotiations that making Internet Explorer the default was an
important part of the agreement and a "deal breaker."
Tevanian, 11/9/98am, at 49:25 - 50:9.
292.3.3. Microsoft's contention is also undermined by the fact that,
even after Apple agreed to make Internet Explorer the default browser, Microsoft
continued to use the threat of withholding Mac Office support as a "club" to ensure
Apple continued to advantage Internet Explorer.
- On January 21, 1998, Microsoft's Don Bradford wrote: "Mac
Office is the biggest Apple carrot. . . . Negotiating from a
single point, centered around MacOffice, will give us better
leverage. . . . Think we should use this opportunity to push
Apple into stronger support of IE." GX 267 (1/21/98 Bradford
e-mail to, among others, Gates).
- On February 13, 1998, Ben Waldman wrote to Bradford,
Gates and other Microsoft executives regarding Apple's
usage of Internet Explorer: "Though the language of the
agreement uses the word 'encourage' I think that the spirit is
that Apple should be using it everywhere and that if they
don't do it, then we can use Office as club." GX 268 (2/13/98
Waldman e-mail). Bradford replied: "Do agree that Apple
should be meeting the spirit of our cross license agreement
and that MacOffice is the perfect club to use on them." GX
268 (2/13/98 Bradford e-mail) (emphasis added).
- In July 1998, Waldman wrote to Gates, regarding Apple: "I
suspect that they are starting to get a bit uncomfortable with
me running all the Mac stuff, because now they can't play off
different parts of MS against each other, and I am holding
them to their commitments (they care a lot about Office, and
I'm using that to get them to be more supportive of IE)." GX
908
292.3.4. That the exclusionary terms were included as part of a
larger set of agreements does not either eliminate the anticompetitive purpose and
effect of those terms or provide any justification for their inclusion in the agreements.
292.4. Fourth, Microsoft's insistence on imposing exclusionary restrictions
on Apple demonstrates that developing a version of Internet Explorer for the Macintosh
and giving it away for free was intended, not simply to enable its customers to
standardize on Internet Explorer across multiple operating systems, but rather to
exclude Netscape.
- See infra Part V.G; ¶ 298.
2. Microsoft also induced RealNetworks not to support Netscape
293. Microsoft also induced RealNetworks not to support Netscape.
293.1. In order to receive distribution of its software through Windows,
RealNetworks entered into an agreement that provided for distribution of RealNetworks'
software with Internet Explorer in exchange for RealNetworks's agreement to notify and
consult with Microsoft before supporting any programming interfaces published by
Netscape or Sun, and to not support such interfaces if Microsoft offers an equivalent
solution.
- The July 18, 1997, agreement between Microsoft and
RealNetworks requires, among other things, RealNetworks,
in return for distribution of its media player with Internet
Explorer (and thus with Windows) to
- redacted - GX
1369 § 4.1, §6.6 (sealed).
293.1.1. This agreement placed a substantial impediment on
RealNetworks's ability to support Netscape's or Sun's efforts to establish a rival platform
that might erode the applications barrier to entry.
- RealNetworks' Jacobsen explained: "In essence, Microsoft
has rights of first discussion if we're going to go support
something that would compete with DirectDraw or
DirectShow. . . . The impact on us is to put a speed bump in
discussions with Sun or Netscape because before we could
conclude an agreement with Sun or Netscape, we would
have to go to Microsoft to have discussions with them." GX
1455 at 147:11 - 148:21.
- Microsoft, he further explained, would be given the
opportunity to propose an alternate to the Sun or Netscape
interface, and only if that solution did not work would
RealNetworks be permitted to support Sun or Netscape
interfaces. GX 1455 at 149:21 - 151:4. If this clause were
invoked, Mr. Jacobsen testified, he expected that Microsoft
"would engage is some serious discussions in between
carrots and sticks" to convince RealNetworks not to support
the competitive interfaces. GX 1455 at 150:23 - 151:4.
293.1.2. The restrictions Microsoft placed on RealNetworks'
dealings with Netscape and Sun lack justification.
- Microsoft has not advanced any technical reason that
RealNetworks' software could not support both Microsoft's and the
alternate interfaces. Microsoft's own media software supports
both Internet Explorer APIs and Netscape APIs. Engstrom,
2/24/99am, at 15:12-15 (Microsoft has designed the Windows
Media Player to work with the Netscape browser plug-in APIs, in
addition to the Internet Explorer APIs).
- Robert Muglia's suggestion that the provision is justified
because Microsoft simply wanted prior notice of
RealNetworks partnering with Netscape or Sun (Muglia
Supp. Dir. ¶ 28) cannot explain the requirement of prior
"discussions" with Microsoft, nor the prohibition against
supporting the interface when Microsoft provides an
equivalent solution.
3. Microsoft conditioned access to early beta releases of Windows and
other technical assistance on ISVs' agreeing to make Internet
Explorer the default browser and to adopt Microsoft-controlled
Internet standards
294. Microsoft similarly used its leverage over ISVs, through so-called "First
Wave" agreements, to extract preferences for Internet Explorer and adoption of
Microsoft-controlled Internet standards.
294.1. Because of the importance of "time to market" in the software
industry, it is crucially important to ISVs that they obtain early beta releases and other
technical information relating to Windows in order to develop applications.
- Gordon Eubanks acknowledged that "early access to select
specifications and beta releases of Windows NT 5 and SDKs on an
ongoing basis" is both "important" and "necessary" to Symantec.
Eubanks, 6/16/99pm, at 16:4 - 17:19.
- Microsoft's other ISV witness, Michael Devlin, also acknowledged
that "The irony of being called an independent software vendor is
that while we're not controlled by" major platform vendors, "our
products rely on services provided by those platforms," many
important aspects of an ISV's business rely on the willingness of
platform vendors to cooperate with the ISV, and Rational is
becoming increasingly dependent its ability to support the Microsoft
platform. Devlin, 2/4/99am, at 14:2 - 15:16.
- William Harris of Intuit testified that Intuit depends on "Microsoft for
the information, specifications, training, development assistance
and development tools necessary to develop our products in an
effective and timely manner." Harris Dir. ¶¶ 27-28.
- An internal Oblix e-mail reports, regarding finalization of its First
Wave agreement with Microsoft that "we need to be totally in bed
with them as over the long run people like Enteva (who are focused
only on MSFT) may be our only real competition." GX 2072.
294.2. In its so-called "First Wave" agreements, Microsoft conditioned
access to early beta releases of Windows, other technical information, and the right to
use certain Microsoft logos, on ISVs' agreeing to set Internet Explorer 4.0 as the default
browser for applications with an HTML-based user interface, and to adopt Microsoft's
"HTML Help" to implement the application's help system.
- GXs 2071; 2400-2463.
294.3. Because of the importance of the technical information and early
beta releases, numerous important ISVs entered into these agreements and thereby
facilitated Microsoft's control over Internet standards and its success in its browser
campaign.
- Among the important ISVs subject to these restrictive terms are
Symantec (GX 2071);
- As early as May 1995, Bill Gates instructed Microsoft executives to
develop a new version of Help tied to Microsoft's browser as a
means of forcing Microsoft browser use. GX 20, at 5 (5/26/95
Internet Tidal Wave memo: "eliminate today's Help and replace it
with the format our browser accepts including exploiting our unique
extensions so there is another reason to use our browser").
- See infra Part VII.D; ¶¶ 398-400.
294.4. Microsoft has suggested no justification for conditioning an ISV's
access to Windows beta releases on its adopting Internet Explorer as the default
browser or using Microsoft's HTML-based help system.
G. Microsoft set a predatory price for Internet Explorer
295. Microsoft set a "better than free" price for Internet Explorer for the specific
purpose, and with the effect, of weakening browser rivals and thereby maintaining its
operating system monopoly.
1. Microsoft set a zero price for its browser for the purpose of
depriving Netscape of revenue and protecting its operating
system monopoly
296. Microsoft recognized before the release of Windows 95 (and continued to
recognize before the release of Windows 98) that charging for Internet Explorer would
generate additional revenue and contemplated charging OEMs and others for it as part
of a software add-on product called, before Windows 95's release, "Frosting."
- In January 1995, Microsoft concluded that "frosting without Ohare
represents a $63MM opportunity, and with Ohare a $120MM opportunity.
We're talking about $57MM difference. It appears that as many as 1.5MM
frosting customers will buy it for the internet access." GX 142.
- A February 10, 1995 Microsoft memorandum entitled "The Case for
Shipping O'Hare with Frosting"evaluates four possible packaging
alternatives (none of which are to bundle O'Hare with Windows 95) for the
first version of Internet Explorer, and recommends that O'Hare should be
put in the Windows 95 add-on product called Frosting to "scoop
incremental revenue from the frosting product." The memo cites a study
conducted in January 1995 by Microsoft that "shows that we can nearly
double frosting sales by including O'Hare in the Frosting pack. . . . This
increases Frosting year 1 sales from an estimated $81 MM to $151 MM,
with marginal incremental COGS, and incremental profits of $61 MM." GX
606.
- See GX 140, at MS98 0107151 (1/13/95 Slivka report noting that shipping
O'Hare with the release of Windows 95 "Helps sell more units of Win95
Frosting (assuming O'Hare is not in Win95)"); GX 143 (Microsoft
document commenting: "We shouldn't give our stuff away."); GX 63 (7/95
Jones e-mail recommending regarding Internet Explorer that Microsoft
should "Figure out pricing and promote agressively [sic]. We need
someone who will go and sell this thing . . . ."); GX 114 (Chase notes that
a "pro" of a proposal to charge for aspects of Internet Explorer is that it
"starts people getting to think about everything won't be free"; a "con" is
that it would "hurt IE share efforts"); GX 118, at MS7 005732 (Windows
marketing group reported research that, upon "hearing IE 4 is free" the
price users were willing to pay for the Windows 98 upgrade product
dropped from $100 to "$10 to $30").
297. Microsoft nonetheless chose not to charge for the browser, and to continue
not to charge for the browser.
297.1. Microsoft included Internet Explorer in Windows at no separate or
extra charge.
- On July 3, 1995, Microsoft released its "Microsoft Internet
Jumpstart Kit" to OEMs "at no additional charge." GX 36.
- Dean Schmalensee conceded that he is unaware of any
circumstances in which Microsoft has charged for Internet Explorer,
either as part of the operating system or as a separate application.
Schmalensee, 1/20/99am, at 21:10-18.
- Microsoft's Brad Silverberg is unaware that Microsoft ever identified
or allocated a portion of the price of Windows to Internet Explorer.
Silverberg Dep., 1/13/99, at 685:25 - 686:5.
297.2. Microsoft also licensed Internet Explorer at a zero price and
committed to doing so "forever" when it distributes Internet Explorer separately from
Windows.
- Bill Gates announced on December 7, 1995, that Microsoft would
make the "Internet add-on" "available at no cost." GX 502, at MS98
0116232; Silverberg Dep., 1/13/99, at 686:6-14 (same).
- Dr. Warren-Boulton testified: "Microsoft's decision to give IE away free
to the installed base of Windows users meant sacrificing substantial
revenue from two sources. First, Microsoft lost revenue from not licensing
IE at a positive price as a stand-alone application -- whether through
downloads directly to end users or through positive licence fees to ISPs,
OLSs and ICPs. Second, Microsoft lost revenue from retail sales of
Windows 98 upgrades because providing IE free to the installed base
reduced the demand for the Windows 98 upgrade and the revenue
Microsoft earns from that source." Warren-Boulton Dir. ¶ 190.
298. Microsoft chose to give Internet Explorer away for free for the purpose of
blunting the threat Netscape posed to Microsoft's operating system monopoly.
- Professor Fisher concluded that "Microsoft studied Netscape's business
model and studied its source of revenue. Microsoft priced its browser for
free and bundled its browser and put a lot of effort into . . . promoting,
bribing, and forcing people to take its browser. A good deal of that
appears to me and appears, I think, from Microsoft documents, to have
been directed at thwarting the threat that Netscape represented to the
operating system monopoly." Fisher, 1/7/99am, 75:19 - 76:1.
- In July 1997, Microsoft's Moshie Dunie noted that selling the Windows 98
"shell" and browser separately "would certainly increase significantly
Win98 upgrade sales." GX 113. Dunie continued that, although "there is
the browser share counter argument" increasing upgrade revenue was "an
intriguing thought." GX 731. Paul Maritz responded: "It is tempting, but
we have to remember that getting browser share up to 50% (or more) is
still the major goal." GX 731; GX 514 (same document). Maritz conceded
at trial that he thought the proposal "to take some features out of Internet
Explorer 4.0 and charge for them separately" was "a proposal with merit,
but that it was outweighed by the desire to increase browser share."
Maritz, 1/26/99pm, at 25:17 - 27:16, 27:18 - 29:8.
298.1. Microsoft determined to give Internet Explorer away for free to "cut
off Netscape's air supply."
- Steven McGeady of Intel testified that Paul Maritz told Intel
representatives in a meeting in the fall of 1995 that Microsoft
planned to "cut off Netscape's air supply" by giving the browser
away for free so that Netscape could not invest in it. McGeady,
11/9/98pm, at 53:8 - 55:16; McGeady, 11/12/98am at 73:21 - 76:4
(same). A January 1998 New York Times article also quoted
McGeady as reporting that Microsoft representatives told him: "we
are going to cut off their air supply. Everything they're selling, we're
going to give away for free." GX 1640, at 4.
- Russell Barck of Intel testified that Maritz "said the term 'embrace
and smother' with respect to a strategy with respect to Netscape."
Maritz, 1/26/99am, at 55:19 - 57:1 (quoting Barck's deposition);
Rob Sullivan (a person for whose competence and integrity Maritz
has a high regard) also testified that Maritz said the phrase
"embrace and smother," and that he "understood that concept to
mean that Microsoft intended to deprive Netscape of revenue and
viability." He understood that Microsoft would achieve this "by
giving away their products, by embracing the Internet standards
and extending them in a way that favored the Windows platform."
Maritz 1/26/99am, at 57:2 - 59:8 (quoting Sullivan's deposition).
298.2. Beginning in 1995, as part of its plan to "cut off Netscape's air
supply" and for the purpose of determining how to deprive Netscape of revenue
necessary to compete effectively, Microsoft studied Netscape's sources of revenue.
- Microsoft was aware that Netscape made most of its revenues from
browser sales. Paul Maritz's February 1996 memorandum "Internet
Browsers" includes a graph showing more than half of Netscape's
"Revenue Mix" comes from its "clients." The same document lists
"More $'s even than Netscape" as one of Microsoft's strengths. GX
473, at MS6 6006240, MS6 6006256.
- In the same April 1996 memorandum in which he argued the
importance of maximizing browser share even though Internet
Explorer is a "no revenue product," Brad Chase described "own
corporate browser licensing" as "one of the biggest potential
revenue opportunities for Netscape." GX 39, at MS6 5005720.
Professor Fisher testified that the juxtaposition of these statements
suggests that "Microsoft was interested in, quote, winning the
browser battle, unquote, not because of the revenues it would bring
in directly, but because of the effect that would have in protecting
Microsoft's operating-system monopoly." Fisher, 1/12/99am, 40:16
- 41:16.
- Steve Ballmer asked his staff in August 1996 to do a "drill-down on
Netscape's browser revenues to understand where they make
money." Bengt Akerlind responded that "Netscape can no longer
make any money on the browser in the OEM market." He also
noted that "Customers/ISPs don't want to talk" about their
payments to Netscape because "they all know we are out to get
them." GX 343. Cameron Myhrvold testified that he understood
"them" in that sentence to refer to Netscape. Myhrvold, 2/10/99pm,
26:3-15.
- In an August 1996 email entitled "Netscape Browser Breakdown,"
Amar Nehru wrote to Cameron Myhrvold, Joachim Kempin, and
others: "Steveb asked us to coordinate a drill-down on Netscape's
browser revenues to understand where they make money and get
back to him in 2 weeks. He suggested I contact all of you. I'd be
grateful if your organizations could help us get data to answer this
question. In the latest quarter ended June 30, 1996 (Q2-96),
Netscape browser revenues were $45mm worldwide. For FY 1997
(Jan 1 - Dec 31), Netscape's browser revenue is projected at
$270mm worldwide. Net, we are trying to categorize the $45mm
and $270mm figures by channel and sub-channel to see how this
can pencil out." GX 969 (emphases in original).
- In November 1996, Amar Nehru sent an e-mail to the Executive
Staff with a long report on "Netscape Revenues." The e-mail
provides in-depth analyses of Netscape's product revenues,
including revenues from browsers. "Browser revenue for the
quarter [Q2-1996] amounted to $45 million (a 32% increase over
the last quarter) representing 60% of total Netscape revenue. Of
the $45 million, ISP's commanded the largest share at 40% of
browser revenue, with direct sales to LORGs via site licenses
coming in second at 28% share." GX 100, at MS98 0122161.
- See GX 39, at MS6 5005720 (Chase emphasized that Netscape's
survival depends on their ability to upgrade a significant chunk of
their installed base to Communicator. They also count on it as a
significant source of revenue and wrote: "Own Corporate browser
licensing. This is one of the biggest potential revenue opportunities
for Netscape. As soon as we have Win 3.1 and Mac clients, we
should have absolute dominant browser share in the corporate
space."); GX 424 (sealed).
- Professor Fisher concluded that "Microsoft, at Bill Gates's personal
direction, undertook detailed studies of Netscape's sources of
revenue and what Netscape required to survive as an effective
competitor. At the time Microsoft made the decision to supply IE
without charge, Microsoft estimated that at the time Microsoft made
its decision to supply IE without charge, from 20 percent to 50
percent of Netscape's revenues came from licensing its browser.
(Bill Gates 8/27/98 Dep. Tr. 236.) Microsoft's decision to price its
browser below cost (indeed, at a zero or even negative price) was
thus made when it knew that Netscape was charging for its browser
and that Netscape depended on those revenues to continue to
compete effectively." Fisher Dir. ¶ 125.
298.3. Netscape originally charged for its browser and, absent Microsoft's
decision to give its browser away "forever free," would have continued to do so.
298.3.1. Netscape charged for its browser before Microsoft
launched its predatory campaign.
- James Barksdale testified: "The commercial release of
Netscape Navigator 1.0 occurred on December 15, 1994.
By the end of the second quarter of 1995, Netscape had
collected over $10 million in revenue generated by the
browser alone. By the end of 1995, Netscape had collected
approximately $45 million in revenue from browsers."
Barksdale Dir. ¶ 18, 57.
- Barksdale confirmed that: "Although Netscape distributed the
beta (i.e. pre-release) version of Netscape Navigator 1.0 free
on the Internet, Netscape's business model in the early days
reflected our intention to charge customers to use the
browser. Consistent with this intent, soon after Netscape
rolled out its retail release on December 15, 1994, Netscape
made it clear to the world that Netscape would charge for
Navigator. The initial price for a Navigator license was $39."
Barksdale Dir. ¶ 20.
- Cameron Myhrvold testified that, in the summer of 1995,
Navigator Personal Edition was priced at $39. Myhrvold Dir.
¶ 22.
- Dean Schmalensee conceded that Netscape sold Navigator
at a positive price. Schmalensee, 6/21/99am, at 13:9 - 15:9.
298.3.2. Netscape dropped the price of its browser to zero only in
response to Microsoft's predatory strategy.
- Dr. Warren-Boulton testified that, because the "incentives" to
earn sales on complementary products "were present when
Netscape first decided to charge a positive price for its
browser." It "was not the potential for the generation of
ancillary revenue that brought the market price of the
browser down the zero, but rather Microsoft's actions."
Warren-Boulton Dir. ¶ 192.
- Barksdale testified that Microsoft's pricing policy forced
Netscape to also give away its browser. "On the revenue
side, all our browser revenue disappeared because it
became increasingly difficult to charge for a product that our
principle competitor was offering for free or 'better than
free.'" Barksdale Dir. ¶ 225. He later testified that Netscape
made the browser for free because it was "forced to" by
Microsoft's pricing policies: "There was no alternative."
Barksdale, 10/27/98pm, at 13:14-20.
298.3.3. Microsoft's contention that its zero price simply mirrored
Netscape's "free but not free" strategy is wrong.
- During cross examination, Barksdale rejected the suggestion that
Netscape's pricing strategy was "free but not free," Barksdale,
10/21/98pm, at 43:8-19. He testified: "to have no revenue and just
market share, I didn't consider that then, now or ever, to be a viable
business strategy." Barksdale, 10/21/98pm, at 47:19 - 48:1.
- James Clark, founder of Netscape, testified: "Netscape has never
given away the browser. It allowed people to download it for free
for one brief period during the beta; but after that, it never gave it
away." Clark Dep. (played 10/27/98pm), at 18:2-5. Marc
Andreessen affirmed that Netscape believed early that it would
give away browsers to education and nonprofit users, but planned
to charge all others: "In fact, that was fundamental to the
company's business plan." Those plans did not change until
January 1998, when MS's pricing rendered the browser market
"noneconomic." Andreessen Dep. (played 10/27/98pm), at 14:11 -16:25.
- Dan Rosen, of Microsoft, was aware that Netscape was selling "site
licenses" to enterprises for browsers in June 1995. GX 25 (Rosen's
notes reflecting June 2, 1995 meeting with Netscape).
- Cameron Myhrvold testified that Netscape was losing sales
to ISPs in competition with Microsoft precisely because
Microsoft gave away Internet Explorer while Netscape
charged for the browser. Myhrvold Dir. ¶¶ 109-111.
- Dean Schmalensee acknowledged that Netscape always
charged corporate users for its Navigator browser.
Schmalensee Dir. ¶ 273. "Netscape had to lower its prices
(or charge fewer customers) in response to IE 3."
Schmalensee Dir. ¶ 279. He further recognized: "The
release of IE 4 put even more pressure on Netscape's ability
to charge for Web browsing software." Schmalensee Dir. ¶
280.
298.4. Microsoft published statements about Netscape's precarious
financial position in order to increase the damage to Netscape from the revenue loss
resulting from Internet Explorer's zero price.
- In an October 1995 Q&A for the New York Times, Bill Gates
described Netscape's high stock market valuation as a "challenge"
for the company. "Netscape has little income, but investors have
valued its stock at more than $2 billion. When a company's shares
have a high value, expectations from investors, including
employee-owners, are correspondingly high. Failure to meet those
expectations can be damaging." GX 333.
- A July 1996 Business Week article, entitled "Netscape: Sitting
Pretty -- Or Sitting Duck?," concluded "Netscape, publicly, is
unfazed . . . but Microsoft, a global software empire with expected
fiscal 1996 sales of $8.6 billion and $2 billion in aftertax profits, has
one enduring edge. 'One thing to remember about Microsoft,' says
Chairman William H. Gates III. 'We don't need to make any
revenue from Internet software.' Who could forget?" GX 84.
- A July 1996 Infoworld report stated that Bill Gates, in comments to
reporters, made a point to "position Netscape as a 'middleware'
company. He then reminded the assembled press critters that,
historically, middleware companies do not last long. Any lead
Netscape has, Microsoft hopes to erase in the home stretch, or -- to
quote [Gates] -- 'What part of the fact that Microsoft owns Windows
don't you understand?'" GX 1248; Barksdale Dir. ¶ 120.
- A January 1997 Forbes article quotes Steve Ballmer as saying,
"We're giving away a pretty good browser as part of the operating
system. How long can they survive selling it?" GX 103, at p.2.
- Bill Gates anticipated the impact of such statements in a memo to
his senior executive staff in May 1996: "At some point financial
minded analysts will begin to consider how much of a revenue
stream Netscape will be able to generate." GX 41, at MS6
6012956.
- Mr. Barksdale confirmed: "Given the power that Microsoft, and in
particular, Mr. Gates, has in influencing the computer industry and
analysts, Microsoft's negative comments, as intended, directly
affected Netscape's ability to compete effectively. It was not a
totally uncommon event for a customer to question whether it made
sense to do business with Netscape because of Microsoft's public
position that it was going to crush Netscape's business." Barksdale
Dir. ¶ 115.
298.5. Microsoft witnesses' assertions that its pricing of Internet Explorer
was not intended to harm Netscape to the contrary are not credible.
298.5.1. Mr. Gates' testimony that Microsoft's inclusion of Internet
Explorer in Windows at no separate charge had nothing to do with Netscape (Gates
Dep., 1/13/99, at 478:7 - 480:4) is not credible.
- In June 1996, Gates reportedly said to the Financial Times:
"Our business model works even if all Internet software is
free. We are still selling operating systems. What does
Netscape's business model look like (if that happens)? Not
very good." GX 71. McGeady testified this statement is
consistent with what Maritz said in his presence. McGeady,
11/12/98pm, 42:4-16; Barksdale Dir. ¶ 119.
- Microsoft held a briefing for press and analysts on the first
anniversary of Bill Gates' 1995 "Pearl Harbor Day" speech,
and announced that Microsoft would give away Internet
Explorer and then include it in its operating systems. Mr.
Barksdale testified that the Seattle Times "reported that
during the briefing Microsoft executives Greg Maffei and
Paul Maritz gloated over the $30 drop in Netscape's stock
price that resulted from the Gates announcement, and
reported that another of their colleagues said of the
precipitous drop in the stock price, 'That's not enough.'"
Barksdale Dir. ¶ 117; GX 1570.
298.5.2. Mr. Gates' testimony that he did not know whether
Microsoft employees collected information concerning Netscape's revenues (Gates
Dep., 1/13/99, at 455:14 - 456:16, 481:13 - 486:25) is not credible.
- Gates in July 1996 responded to an outline entitled "8/19
Netscape Exec Meeting Agenda" by noting that "what is
really important" includes Netscape's "Future growth plan.
Any data analyst's have about how they will grow revenues.
People are expecting Netscape to make a lot of money.
How does that pencil out." GX 980; GX 981.
- Gates specifically requested such data on "how much
software companies pay Netscape" from his staff in
December 1996. He wrote: "In particular I am curious about
their deals with Corel, Lotus and Intuit. All of these ship a lot
of units of Netscape. In our discussions we must have some
kind of sense of the revenue which Netscape gets from this."
GX 345.
298.5.3. Cameron Myhrvold's testimony that Microsoft studied
Netscape's revenues simply to determine how better to compete (Myhrvold, 2/10/99pm,
at 30:7 - 31:19), is not credible.
- Cameron Myhrvold wrote, in October 1997, "NetScape's
client revenue -- is it rising (hope not), falling (I think so), and
do we think they are getting any money from ISPs/netops for
their browser? -- If so, which netops are still paying them
and construct a hunting list for us to go after." GX 117
(emphasis in original).
- See GX 701 (12/97 Myhrvold e-mail describing "our progress
agianst [sic] the netscape hunting list," referring to
Microsoft's successes in cutting deals with Netscape's ISP
accounts); Myhrvold, 2/10/99pm, 33:22 - 34:14 ("I certainly
wondered whether Netscape could keep up the pace of
innovation if they weren't making money from that" ISP
"channel . . . So, really, it would be a question as to whether
Netscape would choose to continue to invest in the browser
if they weren't making money there.").
2. Microsoft incurred hundreds of millions of dollars in costs in
its effort to gain browser usage share
299. Although Microsoft gave Internet Explorer away for free, it spent hundreds
of millions of dollars on developing and marketing Internet Explorer in order to gain
browser usage share and blunt the platform threat.
299.1. Microsoft spent tens of millions of dollars each year from 1995 to
the present developing Internet Explorer to run with Windows 95 and Windows 98.
- Microsoft represented in an interrogatory answer that its
development expenses for Internet Explorer were on the order of
$100 million each year. Fisher, 1/12/99am, at 35:13-19;
Schmalensee, 1/20/99am, at 47:8-18. Schmalensee Direct ¶ 211
("Microsoft reportedly has spent more than $100 million annually in
developing Internet-related technologies for Windows . . . .").
- Dean Schmalensee testified that the figure of half a billion dollars is
broadly consistent with his understanding of how much Microsoft
has spent overall to develop Internet Explorer. Schmalensee,
1/20/99am, 48:9-15.
- According to Brad Chase: "The cost of rebuilding complex software
from the ground up is high." He further testified that Microsoft
"dedicated a team of more than 100 developers to the Internet
Explorer 3.0 effort. To put that number in perspective, the original
Internet Explorer 1.0 team consisted of five or six developers."
Chase Dir. ¶ 20.
- Microsoft's February 1998 - redacted -
GX 428, at MS7 00389 (sealed).
299.2. Microsoft also spent millions developing Internet Explorer for other
operating systems.
- See supra Part V.B.1.c.(1); ¶ 113.
299.3. Microsoft spent millions marketing and promoting Internet
Explorer.
- A memo to Bill Gates entitled "think week" about "How to Get To
30% Share In 12 Months," stated: "Content drives browser
adoption, and we need to go to the top five sites and ask them
'What can we do to get you to adopt IE?' We should be prepared
to write a check, buy sites, or add features -- basically do whatever
it takes to drive adoption." GX 334, at MS98 0104679; GX 684
(same document).
- In an "Internet Browsers" presentation, Maritz concluded that to
gain share, Microsoft would engage in "Massive seeding of IE.
Magazines, tradeshows, via partners, high profile events etc. IE
being free is a key advantage to push . . . . Pay for premium
merchandising positions." The presentation contemplated for
Internet Explorer "Broad advertising on the web & with traditional
media?" GX 473.
- In May 1996, Bill Gates sent a memo to senior executives
describing "lots of ways to spend money" to promote "the browser,"
including advertising ("Clearly we need to do a lot of this . . . .") and
expenditures on distribution including "massive airdrops"). GX 41.
- In July 1997, Paul Maritz noted the high cost of marketing Internet
Explorer and the reasons for it: "There is talk about how we get
more $'s from the 1000+ people we have working on browser
related stuff" than from increasing Windows 98 branding, "but I
have not lost sight of the fact that Browser Share is still an
overwhelming objective. You may notice that I have kept IE
marketing spend [sic] at very high level through FY '98. [sic] and
resisted pressure to reduce this or switch it to other products. I also
said 'no' on the proposal to charge separately for the Shell." GX
112 (emphasis in original).
- See Mehdi Dep., 1/13/99, 655:4-20 (IE FY97 marketing budget
roughly $30 million); GX 511 (Chase wrote to Microsoft senior
executives in April 1997: "Browser share needs to be a top priority
around the world. Marketing budgets, including mine, should be
budgeted about equal to this year (we are doing a bottom's [sic] up
IE budget now, last year including some drg [Developer Relations
Group] efforts I was around $69M)."); GX 795; GX 696 (sealed).
299.4. Microsoft also spent millions, both in direct payments and through
giving away valuable Windows "real estate" and other property, to induce third parties to
favor Internet Explorer and disfavor rivals.
299.4.1. Microsoft paid OLSs and ISPs to gain browser usage
share.
299.4.1.1. Microsoft, through its ICW and OLS Folder
contracts, paid ISPs and OLSs to favor Internet Explorer and severely restrict their
distribution and promotion of browser rivals.
- Microsoft "made a considerable investment in order to
establish the Windows Referral Server," (Myhrvold
Dir. ¶ 49) including spending between four to five
million dollars a year to lease the network, in addition
to the development of software and 24-hour per day
maintenance of servers. Myhrvold, 2/10/99am, at
27:17 - 28:6.
- Microsoft traded placement in the referral server and
the online services folder for the promotion of Internet
Explorer and the exclusion of rivals. See supra Part
V.D.3.b; ¶¶ 227-236.
299.4.1.2. Microsoft paid ISPs bounties, and (in some
cases) direct cash payments, to convert their installed base to Internet Explorer.
- Microsoft gave Netcom a nine dollar discount for
every Internet Explorer distributed to the installed
base. Myhrvold, 2/10/99pm, at 11:18 - 13:2; GX 81.
- Microsoft planned to convert "existing Nscp users"
through the "ISP bounty program," according to a
January 1997 presesntation entitled "NC & Java
Challenge." GX 51, at MS7 005539.
- Myhrvold explained the mechanics behind Microsoft's
payment to ISPs to convert their installed base to
Internet Explorer. Microsoft offered discounts off of
the referral fees owed to Microsoft for subscribers
gained through the Internet Referral Server for
"distributing Internet Explorer to existing users of
Netscape Navigator." Myhrvold Dir. ¶ 62. GX 1141,
at MS6 5000017 (Exhibit D4;
- redacted -
(sealed); GX 1144 (Spry agreement) (sealed) ; GX
1146 (Mindspring agreement) (sealed).
- Microsoft entered into a contract with AOL to promote
Internet Explorer through distributing it to AOL's
installed base. The contract provided that AOL would
be paid $0.25, up to a limit of one million, per
conversion to Internet Explorer. GX 1019 (AOL
Access Software Advertising and Promotion
Agreement); GX 978 (summary of Internet Explorer
promotional agreement). Brad Chase summarized
the agreement in an internal Microsoft email: "It works
as follows: if they convert 4M of their users to a client
using IE 3 by 2/1/97, i will give $1.5M. If they get an
additional 2M by 4/2/97, I will give them $500K more."
GX 976.
- In addition, the contract between Microsoft and
- redacted -
in the Microsoft Referral Server provided that
would receive - redacted - for each
new subscriber who already uses Internet Explorer or
upgrades to Internet Explorer. GX 1069, at MS98
0101395 (Exhibit 3) (sealed).
- See also Part V.D.3.a; ¶ 223.
299.4.1.3. Microsoft paid ISPs to use Internet Explorer-specific technologies, which Microsoft expected to influence web site standards and
increase Internet Explorer usage.
-
- redacted -
GX 1132
(sealed).
-
- redacted -
GX 1141
(sealed);
- redacted -
GX 1144 (sealed);
- redacted -
GX 1146
(sealed).
- Dr. Warren Boulton concluded that, because Active X
is "operating system (typically Windows) specific," the
effect of the provision giving discounts for the use of
Active X "is to reward ISPs that configure their
services in a way that reduces the cross-platform
threat to Microsoft's operating system monopoly. The
reason is that ISP use of Microsoft-specific
technologies reinforces the dominance of the
Windows platform." Use of "such technologies by
ISPs serves to blunt the cross-platform threat" that
"rival browsers might pose." Warren-Boulton Dir. ¶
108.
299.4.1.4. Microsoft paid ISPs in other ways to favor
Internet Explorer over rivals.
- In a presentation on "Internet Browsers," Maritz listed,
among other inducements, the following to get
Internet Explorer share: "ISPs. Allow ISPs to be in
Windows and the Internet Starter Kit . . . . Provide
customization opportunities so ISPs can brand their
offering and add specific features. Co-marketing
funds to encourage ISP partners to promote our
browser and get new customers for them." GX 473,
at MS6 6006248.
- Microsoft paid ISPs and others to buy out contracts
that the ISPs had with Netscape to distribute
Netscape's browser. Schmalensee, 1/20/99am, at
56:19 - 57:23.
- Microsoft offered, in attempting to induce AT&T to
enter into a contract favoring Internet Explorer, to pay
off up to $17 million in minimum commitments owed
to Netscape. GX 179. In addition, Microsoft created
a co-marketing fund for the distribution of Internet
Explorer that consisted of a $5 credit for every copy of
Internet Explorer distributed, not to exceed $5 million.
Myhrvold Dir. ¶ 29.
- See also supra Part V.D.3; ¶¶ 223-239 (detailing the
large value Microsoft bartered with AOL and other
ISPs for exclusionary terms); Barksdale ¶ 142 (citing
GX 75).
299.4.2. Microsoft paid ICPs in order to gain browser usage share.
d
299.4.2.1. Microsoft exchanged placement on the channel
bar for ICPs' agreement to restrict their dealings with browser rivals and adopt Internet
Explorer-specific technology.
- See supra Part V.E.2; ¶¶ 263-275.
299.4.2.2. Microsoft paid ICPs in other ways.
- In his February 1996 memorandum "Internet
Browsers; 1. Netscape's actions; 2. How to Win," Paul
Maritz outlined plans for ICPs: "Build 'first wave' like
programs to provide value to ICPs that build on our
platform and use our logo" and to set aside a "special
marketing pool of $30M to build co-marketing
opportunities with the key site." GX 473, at MS6
6006231, MS6 6006248.
- In June 1996, Brad Chase reported to Paul Maritz and
Brad Silverberg on a "tremendous" deal just struck
with Starwave, which operates of the ESPN
Sportszone website -- which Chase characterized as
"one of the top few sites on the internet . . ." -- and the
Family Planet website. Under the terms of the deal,
Microsoft agreed to pay Starwave a flat fee of
$500,000 plus a per-head bounty, up to a total of $1.2
million. In return, Starwave committed throughout
1996 and 1997 both to undertake activities to promote
Internet Explorer and not to undertake activities with
Netscape. GX 862.
- Barksdale testified about several similar episodes.
Barksdale Dir. ¶¶ 186, 188, 189, 190, 200; GX 72;
GX 79; GX 85; GX 90; GX 94; GX 1250.
299.4.3. Microsoft used costly Market Development Agreements
("MDAs") with OEMs to gain browser usage share.
299.4.3.1.
-
- redacted -
GX 1498 (sealed); see also GX 1506 (HP)
(sealed); GX 1503 (IBM) (sealed); GX 1493 (AST)
(sealed); GX 1509 (Hitachi) (sealed); GX 1511
(Packard Bell) (sealed); GX 192 (NCR).
-
- redacted -
GX 1169 (HP) (sealed); GX 1171 (Dell) (sealed); GX
1171A (Dell) (sealed); GX 1173 (Gateway) (sealed);
GX 680 (Toshiba) (sealed).
-
- redacted -
GX 1169 (HP) (sealed); GX 1171 (Dell)
(sealed); GX 1171A (Dell) (sealed); GX 1173
(Gateway) (sealed); GX 680 (Toshiba) (sealed).
- redacted -
Compaq ($1.00/CD) and HP.
GX 163 (Compaq) (sealed); GX 1169 (HP) (sealed).
-
- redacted -
GX 979 (sealed).
299.4.3.2. Microsoft offered OEMs MDA discounts and
other consideration to make or promote Internet Explorer as the default, preferred, or
exclusive browser.
- See supra Part V.C.2.a.(2); ¶ 203.
- In a May 27, 1998, draft "IE5 OEM Marketing
Review," Microsoft listed as "PHASE IV - Encourage,
PR, Launch," that it would "Explore joint marketing
opportunities to solicit Netscape users to convert to
IE5." GX 233, at MS98 0125666.
- In March 1996, Nick Zaharias reported to other
Netscape executives that Dell's Director of Software
Procurement told him that "Microsoft is willing to make
Internet Explorer 'better than free' in exchange for a
positioning statement that would make MSIE the
'browser of choice' or 'preferred product.'" He said
Zenith Data Systems had had a similar offer. GX
182; GX 236.
- In November 1996,
- redacted -
GX 758 (sealed). A January 1997 internal Compaq
e-mail reports that Joe Williams of Microsoft had
confirmed in principle Microsoft's offer to share ISP
revenues with Compaq. GX 1041.
- Barksdale testified that his salespeople had reported
to him several instances in which "OEMs were offered
a discount on Microsoft products, including Windows,
if they would make Internet Explorer their 'preferred'
browser." Barksdale Dir. ¶ 165 (citing GX 87, GX
188, GX 199). The Netscape salespeople reported
that the threat to raise Windows royalties "has always
been done verbally and they never left any evidence."
GX 188.
- - redacted -
- - redacted -
299.4.3.3. Microsoft offered OEMs MDA discounts and
other consideration to adopt IE-specific technologies.
- See supra Part V.C.2.(a); ¶ 203.1.
299.4.4. Microsoft paid ISVs to gain browser usage.
299.4.4.1. Microsoft entered into First Wave agreements
that gave ISVs preferential access to Microsoft "Beta" releases of Windows in exchange
for preferential terms for Internet Explorer.
- In exchange for, among other things, access to so-called "Beta" releases of Microsoft operating system
products, participating ISVs agreed that: "If the user
interface is HTML based, Internet Explorer 4.0 must
be set as the default browser." GX 2071 (Symantec)
(sealed); Microsoft entered into dozens of such or
similar agreements with leading ISVs. GXs 2400 -
2497 (sealed).
299.4.4.2. Microsoft sought to bribe ISVs in other ways to
gain browser usage share.
- In July 1996, Bill Gates told the CEO of Intuit, Scott
Cook, that if Cook "had a favor we could do for him
that would cost us something like $1M to do that in
return for switching browsers in the next few months I
would be open to doing that." GX 94.
299.4.5. Microsoft paid end users and other firms to gain browser
usage.
- Barksdale testified concerning numerous instances in which
Microsoft offered end users significant consideration for
exclusively using Internet Explorer. Barksdale Dir. ¶ 187
(testifying that Intelligent Electronics was offered $100,000
as part of a deal requiring exclusive use of Internet
Explorer); GX 104 (in January 1997, Microsoft offered
Intelligent Electronics $100,000 as part of a deal requiring
exclusive use of Internet Explorer.) Barksdale Dir. ¶ 199
("Microsoft offered to upgrade telecom New Zealand's 9000+
Win 3.1 terminals to Windows 95 for free if Telecom would
use InternetExplorer as its internal browser."); Barksdale Dir.
¶ 202 (International Paper); see also GX 74; GX 77.
- Netscape understood that Microsoft was "going into all major
accounts" that Netscape contacted in Brazil, and offered,
among other things, to pay "$1.00 to take each navigator out
of the account," "Support and provide all of their products
and give mktg dollars to support vendors in trade shows,
conferences etc.," and "Give MS Explorer for free for 2
years." GX 1251.
3. Microsoft also sacrificed revenue from other products to gain
browser usage share
300. Microsoft also deliberately sacrificed revenue from other products in order
to implement its campaign to gain browser usage share.
300.1. Microsoft's coercive conduct, including its screen restrictions and
its requirement that all OEMs, regardless of preference, distribute Internet Explorer,
diminished the value of Windows to OEMs and thereby reduced the price OEMs were
willing to pay for Windows.
- Dr. Warren-Boulton testified: "Microsoft's tying of IE to its operating
system made distribution of rivals browsers infeasible or more costly for
OEMs and thus reduced the OEMs' demand for Windows." Warren-Boulton Dir. ¶ 189.
- Microsoft's screen restrictions, as explained, imposed significant
costs on OEMs, inhibited their ability to differentiate their products
and best serve users, and thus reduced the value of Windows to
both OEMs and end users. See supra Part V.C.1.b.(1); ¶ 171.
- Dean Schmalensee confirmed that economists look to more than
just the price terms of a product in determining the real cost of the
product to customers. He testified in a previous case (Data
General), where he believed that the defendant had engaged in a
tie-in of two products, that: "You must realize, parenthetically, by
price, an economist means not only the dollars paid, but everything
else that affects price." Schmalensee, 1/19/99am, at 41:14 - 45:4.
300.2. Microsoft's use of desktop placement to gain browser usage share
reduced its revenues from MSN and other products (such as from selling distribution
with Windows itself).
- See supra Part V.D.3.b.(2); ¶ 232 (detailing Gates' concern that
putting AOL in the "Windows box" would put a "bullet through
MSN's head").
- Professor Fisher testified that the opportunity cost to gain share for
Internet Explorer Microsoft incurred included "offering places on its
desktop real estate that was valuable to the recipient and for which
Microsoft could otherwise have charged." Fisher, 6/1/99am, at
69:19 - 70:9; see also GX 868D.
- Dr. Warren-Boulton testified: "Microsoft's agreements with ISPs and
OLSs provided those firms with preferential access on highly
desirable terms to valuable Desktop real estate. This is an unique
asset; its value was enhanced by the OEM screen restrictions; and
it could have generated substantial direct revenue for Microsoft if it
had been sold rather than bartered or exchanged for exclusivity
agreements." Warren-Boulton Dir. ¶ 189.
4. At the time it incurred its immense browser-related costs,
Microsoft did not anticipate recoupment except through
weakening browser rivals and thereby protecting its operating
system monopoly
301. Microsoft anticipated recouping its browser-related costs by weakening
browser rivals and thereby protecting its operating system monopoly.
- Brad Chase wrote, in an April 4, 1996, memorandum entitled "FY 97
Planning Memo 'Winning the Internet platform battle,' under the heading
"Why should you care?,': "This is a no revenue product, but you should
worry about your browser share as much as BillG because" Microsoft "will
loose the Internet platform battle if we do not have a significant user
installed base. The industry would simply ignore our standards. Few
would write Windows apps without the Windows user base." GX 39, at
MS6 5005720 (emphasis in original).
- Paul Maritz wrote on June 20, 1996: "Without browser share, everything is
very hard. So job #1 is browser share. We also have to persuade approx
5 million persons to start using IE over the next 6 months." GX 42.
- An internal Microsoft presentation for a meeting hosted by executive Brad
Silverberg states, under the heading "Internet Battle": "This is not about
browsers. Our competitors are trying to create an alternative platform to
Windows. They are smart, aggressive, and have a big lead." GX 40
(emphasis in original).
- Paul Maritz wrote on July 11, 1997: "There is talk about how we get more
$'s from the 1000+ people we have working on browser-related stuff, but I
have not lost sight of the fact that Browser Share is still an overwhelming
objective. You may notice that I have kept IE marketing spend [sic] at
very high level through FY'98, and resisted pressure to reduce this or
switch it to other products. I also said 'no' on the proposal to charge
separately for the Shell." GX 112 (emphasis in original).
- Brad Chase wrote on April 8, 1997: "Last year, and before that, we went
on a jihad as we saw the threat to our platform from Netscape Navigator."
Chase further explained that "it's critical that we maintain our focus on
gaining browser share." GX 511.
- In April 1997, Chase wrote, "IE share is critical. Without it, we lose the
desktop, which translates to Windows and Office revenue over time." GX
59. He had also expressed this idea earlier that month in memorandum,
"FY98 Planning Memo 'Preserving the Desktop Paradise'." GX 510, at
MS7 004127.
- Paul Maritz wrote on July 14, 1997, in response to a suggestion that
Internet Explorer 4 be shipped separately from Windows 98 and sold for a
positive price as part of the Windows 98 upgrade product, that charging
for Internet Explorer "is tempting, but we have to remember that getting
browser share up to 50% (or more) is still the major goal." GX 113.
- Dr. Warren-Boulton testified: "The available evidence indicates that Microsoft
pursued the practices I have examined for the purpose of preserving its Windows
operating system monopoly and gaining monopoly power in the browser market,
and pursued them without regard to whether they would have been profitable in
their own right. Accordingly, it is my opinion that Microsoft's intent in engaging
in this course of conduct, when considered as a whole, was predatory." Warren-Boulton Dir. ¶ 185.
- Professor Fisher testified: "Microsoft internal documents make clear that
Microsoft undertook its browser development not to make money from browsers,
not because doing so would 'make sense from a business standpoint' on its own,
but to prevent Netscape's browser from facilitating competition with Microsoft's
monopoly operating system." Fisher Dir. ¶ 124. Professor Fisher further
testified that "It is important to note that this is not merely colorful language
that could be interpreted either as aggressive competition or as evidencing
a predatory intent (for example, language like: 'Our goal is to get 100% of
the business' or even like 'Let's kill the competition'). This is language that
accurately describes the purpose and effect of Microsoft's conduct--distribute its browser at a zero or negative price in order to eliminate
competition." Fisher Dir. ¶ 126.
302. There is no contemporaneous evidence that Microsoft anticipated any other
way of recovering its massive Internet Explorer related costs.
- Professor Fisher explained the purpose of looking at contemporaneous evidence:
"what matters is what is expected (or can reasonably be expected) at the time the
action in question is taken." Fisher Dir. ¶ 49.
- Professor Fisher testified: "This was a serious expenditure of money.
What was happening here with the browser was a big effort for Microsoft.
They spent a lot of money -- hundreds of millions -- to develop the
browser. They gave away valuable real estate. They, in effect, paid
people to take it. And this was a no-revenue product -- explicitly a no-revenue product. Serious businesses -- and I certainly take Microsoft to
be a serious business -- don't typically engage in activities like that, unless
there is some relatively formal or even -- relatively formal showing that it's
going to bring in revenues, and, therefore, be a profitable thing to do. I
know of no document that suggests that at all, and I know of no document
-- and I certainly know of no document that can be called anything like a
formal business plan that shows those revenues and shows that this is
going to be a profit-maximizing choice." Fisher, 1/12/99am, at 34:11 -
35:12.
- Professor Fisher further testified: "Microsoft's documents do not say,
'we're doing this with Internet Explorer because Internet Explorer is going
to bring in a lot of money.' In fact, contemporaneous documents do not
suggest that Microsoft cared at all about -- and some of its actions also
confirm this -- that Microsoft cared at all about the ancillary revenues that
might" be "derived from giving away Internet Explorer." Rather,
"Microsoft's documents are full of statements" that "'This is a no-revenue
product, but you should care about it just as much as does Bill Gates.
Without winning the browser war, we lose.'" Fisher, 6/1/99am, at 40:7-25;
68:18 - 69:10 (explaining that Microsoft's documents confirm that
Microsoft engaged in its browser-related conduct "to protect" its "monopoly
power," particularly the document states that "without browser share, we
lose -- and then it makes mention of both, I think, Windows and office").
- Dean Schmalensee conceded that he did not "see any analysis of the
revenues that Microsoft expected to receive, or any written indication of
the revenues that Microsoft expected to see from the browser in 1994 or
1995 or 1996 or 1997." Schmalensee, 6/24/99pm, at 16:12-21.
303. Because Microsoft expected to protect its operating system monopoly by
weakening browser rivals, it priced the browser without regard to cost.
- During the negotiations for the Internet Explorer promotional agreement
with AOL, Colburn was told that "Microsoft had no limitations on what it
could spend to gain market share for Internet Explorer." Colburn Dir. ¶¶
38-39; GX 689 (AOL email reporting that Microsoft could spend any
amount to gain market share for Internet Explorer).
- Dr. Warren-Boulton testified that he had seen no documents dating from
the time that key decisions about Internet Explorer were made to indicate
that Microsoft ever performed a calculation comparing the costs it
"incurred through its pricing policies and exclusionary agreements"
regarding Internet Explorer to the "revenues Microsoft could have
expected to gain absent any effect on the competitiveness of the browser
and operating system markets." Warren-Boulton Dir. ¶193.
- Paul Maritz's trial testimony about whether or not Microsoft tracked the
development expenses of the browser when it was actually being
developed and priced is internally contradictory, confusing, and incredible.
- First, Maritz said that he could quantify "how much money it cost"
Microsoft "to develop Internet Explorer" by looking at, among other
development expenses, salaries of employees, capital equipment,
and corporate overhead. Maritz, 1/26/99pm, at 6:11-23. He further
testified, in contradiction to Dean Schmalensee's testimony that
records at Microsoft of this sort did not exist, that these figures
"would have been prepared in the normal course of business," and
that he personally became aware that Microsoft was tracking
Internet Explorer development expenses in the "middle of fiscal
year 96." Maritz, 1/26/99pm, at 8:5 - 10:10.
- At his deposition, however, Martiz testified that he did not know
whether Microsoft ever kept track of how much money it had spent
on browsers, or whether Microsoft ever made an estimate of how
much money it has spent on browsers. After he was shown this
testimony, Maritz changed his trial testimony to admit that Microsoft
never made an estimate of how much it spent to develop its
browser and assert only that he could give an estimate of those
costs if asked to figure it out today. Maritz, 1/26/99pm, at 10:14 -
13:22.
- Indeed, Maritz's testimony led the Court to observe: "I have here in
my notes that prior to looking at his deposition, he said that the
development of browser costs were tracked by Microsoft from '94
on, and then he became aware of it sometime in fiscal '96 . . . from
that I inferred that he knew that there was some specific accounting
of the investment in the browser. And then after he looked at his
deposition, he seemed to think that the figures that he had were
only bits of information which related generally to the development
of Windows." Maritz, 1/26/99pm, at 17:10-21.
304. Microsoft's incurring of its massive browser-related costs and pricing the
browser at zero did not otherwise "make sense from a business standpoint" and,
therefore, cannot be explained except as part of a predatory strategy to weaken its
browser rivals and thereby protect its operating system monopoly.
- Professor Fisher testified that, absent maintenance of its operating system
monopoly, Microsoft's conduct does not make sense and is not profitable:
- "Without the gain to Microsoft that will result from preserving its
highly profitable operating system monopoly and from monopolizing
the browser market, Microsoft's conduct does not 'make sense from
a business standpoint.' It is giving away, indeed paying people to
take and distribute, something that it has spent a lot of money to
develop and distribute and something for which the leading
competitor was charging." Fisher Dir. ¶ 127.
- "It is only when Microsoft's gains from preserving and extending its
monopoly are included that Microsoft's conduct is profitable."
Fisher Dir. ¶ 128.
- "Microsoft's price for its browser, together with its other actions, is
not profit-maximizing except for its effect of preserving Microsoft's
operating system monopoly (and possibly gaining further monopoly
profits by monopolizing the browser market and its ancillary
revenues)." Fisher Dir. ¶ 241.
- "Microsoft was interested in 'winning the browser battle' not because of the
revenues it would bring in directly, but because of the effect that would
have in protecting Microsoft's operating-system monopoly. In order to do
that, they were not merely interested in how well they would do. They
were also particularly interested in being sure that Netscape did not do
well in browsers." Fisher, 1/12/99am, at 40:25 - 41:9.
- Dr. Warren-Boulton reached the same conclusions.
- "Microsoft's conduct, in the aggregate, was not expected to be
profitable except for the market power Microsoft expected to gain
from the exclusion of browser rivals and therefore was predatory."
Warren-Boulton Dir. ¶ 195.
- "The available evidence indicates that Microsoft pursued the
practices I have examined for the purpose of preserving its
Windows operating system monopoly and gaining monopoly power
in the browser market, and pursued them without regard to whether
they would have been profitable in their own right. Accordingly, it
is my opinion that Microsoft's intent in engaging in this course of
conduct, when considered as a whole, was predatory." Warren-Boulton Dir. ¶ 185.
305. Microsoft's pricing of Internet Explorer was not profitable (absent monopoly
recoupment).
305.1. Microsoft's pricing of Internet Explorer was predatory because the
development and distribution costs exceeded the revenues that Microsoft could
reasonably have anticipated from the zero price.
- Professor Fisher testified that Microsoft's actions "were simply not
profitable at all on any standard." Fisher, 6/1/99am, at 37:21 - 38:4.
- Dr. Warren-Boulton testified: "The evidence I have seen supports
the inference that Microsoft took exclusionary actions and incurred
costs without regard to whether its actions were profit-maximizing
or even profitable absent the future revenue gains from
weakening rival browsers and thereby preserving its Windows
operating system monopoly and from gaining monopoly power in
the browser market. Instead, Microsoft viewed winning browser
share at almost any cost as being of overwhelming strategic
importance." Warren-Boulton Dir. ¶ 194.
305.2. Microsoft's pricing was predatory, even considering only the costs
associated with Microsoft's provision of Internet Explorer separately from Windows.
- Professor Fisher testified that Microsoft's conduct was not profitable
because "Microsoft gave away Internet Explorer. It was to be
forever free. Microsoft's documents describe it correctly as 'This is
a no-revenue product.' Now, this was a product which Microsoft
not only gave away for free, but basically bribed people to take.
They gave them preferred places on the desktop for which"
Microsoft "could have charged. But beyond that, they also spent
hundreds of millions of dollars on the development of this no-revenue product, and then they gave away the technology. That is
not a profitable act, except for the protection of the operating
system's monopoly." Fisher, 6/1/99am, at 39:14 - 40:6.
5. The effect of Microsoft's predatory pricing of Internet Explorer
has been to impede rivals, harm consumers, and facilitate
Microsoft's objective of blunting the browser threat
a. Microsoft's predatory pricing injured competition
306. Microsoft's predatory pricing of Internet Explorer increased its share at
rivals' expense.
- As Professor Fisher explained, it is the combination of offering a browser
that was roughly equivalent to Netscape Navigator, and offering it at a
zero price, that increased Internet Explorer's share. Fisher, 6/2/99am, at
8:5-17 (testifying that he doesn't "deny that an improved IE was required
to make Microsoft's strategy succeed. Predation pricing, to succeed, has
got to be the offering of an unprofitably low price for a product that, at the
lower price, consumers will want. That means you've got to have an
adequate product that consumers will really want at the low price. So long
as IE was quite inferior . . . offering it at a zero price would not be sufficient
to persuade customers to take it.").
- Dean Schmalensee acknowledged that Earthlink "represented one of the
many ISPs that struggled to justify paying to distribute Netscape's client
products when they could distribute the improved internet Explorer for
free." Schmalensee Dir. ¶ 277.
307. Microsoft's predatory pricing injured its principal rival, Netscape, in other
ways as well.
307.1. Microsoft's predatory pricing deprived Netscape of browser
revenue, thereby impeding its ability to innovate.
- Graphs of Netscape's quarterly browser licensing revenues show
that those revenues had been reduced to zero by Q1 1998. GX 9;
GX 10. Likewise, Netscape's 1997 Annual Report identified its
client stand-alone revenues for 1997, 1996 and 1995 as $105.5
million, $181.2 million, and $77.5 million, respective. The Report
concludes: "The decreases in all periods as a percentage of total
revenues as well as the absolute dollar decrease in 1997 were due
to increased price pressure from Microsoft Corporation, a
competitor that offers its browser with no licensing fees. In January
1998, Netscape announced that it would offer its client software for
free. As a result, Netscape does not expect to generate any further
significant client stand-alone revenue." GX 367.
- James Barksdale testified: "We have already cut back on some of
the things we wanted to do and extensions and expansions . . . .
We depended on revenue to fund all of these new ideas. . . . So
the money we were making was what was allowing us to do these
things. If we don't bring in the revenue, by definition, you were
trapped, and you were less innovative and less responsive to
market opportunities." Barksdale, 10/27/98pm, at 24:5- 25:3;
see also Barksdale, 10/21/98pm, at 55:3 - 56:25 (Microsoft's pricing
strategy has led to less investment, and therefore innovation, in
browsers at Netscape); Barksdale, 10/27/98pm, at 20:4-12
(browser revenue was "absolutely" vital to Netscape's ability to
continue to improve the product: "We had a payroll to make").
- Marc Andreessen testified that "it became clear to us in the '96-97
time frame that it was not an economically feasible proposition to
continue that development path. We would never generate a
return." He testified that this problem arose from pricing pressure
on browsers "ultimately down to zero," lack of access to OEM, ISP,
and other channels, and a broad range of sales and marketing
tactics by Microsoft. Andreessen Dep., 7/15/98, at 130:4 - 131:9
(DX 2555). He further testified that Netscape's change in focus
from the client to the server was motivated by "every reporter and
analyst in the world believing that Netscape was going to go
bankrupt because we were dependent on that revenue . . ." from
the browser "and also every customer in the world believing that
Netscape was going to go bankrupt. Not every customer, but
many." Andreessen Dep., 7/15/98, at 137:16 - 138:7 (DX 2555);
see also Andreessen Dep., 7/15/98, at 138:8-21 (DX 2555)
- Scott Bosworth of IBM testified that Netscape's concern that the
browser was "no longer a viable area for it to invest in" was the
main reason that the Java browser work was dropped. Bosworth
Dep., 10/16/98, at 80:10 - 81:5 (DX 2609).
- Dr. Warren-Boulton testified that "given the zero pricing for
browsers, given the absence of revenue from that source,"
Netscape "had decided to reduce its investment in updating the
browser." Warren-Boulton, 11/24/98am, at 74:8-13.
307.2. Microsoft's predatory pricing deprived Netscape of distribution
through OEMs and ISPs, further injuring Netscape's ability to maintain share.
- In a memo to FY'97 WWSMM Attendees in April 1996 regarding
"FY97 Planning Memo 'Winning the Internet platform battle,'" Brad
Chase wrote: "you should go out to all the significant ISPs and on-line services in your country in May and close licensing
agreements. You should also be able to break most of Netscape
licensing deals and return them to our advantage because our
browsers are free." GX 465, at MS6 6002322.
- Cameron Myhrvold noted the impact of Internet Explorer's
"preferred" licenses to ISPs (for which Microsoft did not charge):
"Technically they can also distribute other browsers but in fact very
few do simply because of our better economics." GX 193.
- Internal correspondence in January of 1996 between Netscape's Peter
Thorp and Ram Shriram, reveals that in negotiations with PSI, an ISP that
was interested in licensing Navigator, PSI indicated that "Microsoft is
offering to give them the world for free. They really want to do this deal
and go with Netscape, but free tough to argue with." GX 65; Barksdale
Dir. ¶ 140 (discussing GX 65).
- The President of Global Telecom wrote that "Microsoft gave me a deal
I couldnt [sic] refuse. Free dialer, browser, developer kit, freely
distributable, etc. . . . I know Netscape is better, but $0 vs $18K is
impossible to beat." GX 73; Barksdale Dir. ¶ 149 (discussing GX 73).
- As a result of Microsoft's "constriction of Netscape's distribution
channels," by 1997 Netscape's browser revenues were significantly
reduced. Barksdale Dir. ¶ 219. A chart prepared by Barksdale
demonstrates Netscape's revenue growth flattening then declining
through the first quarter of 1998. Barksdale Dir. ¶ 219.
- See also GX 108 (MidOhioNet canceled Netscape account
because Internet Explorer was free); GX 109 (same for Bliss
Computer Services); GX 111 (same for Web Services Group); GX
115 (same for Mercury Internet Services); GX 116 (same for
Seescape).
307.3. Netscape had to offer inducements similar to Microsoft's to retain
market share, and that further deprived it of revenue needed to compete.
- See infra Part VII.A.2.b; ¶ 363.3.
308. Microsoft's predatory pricing also retarded the development of other
browsers.
- James Gosling testified: "The HotJava browser is a software application that
was released by Sun in 1995. At the time the HotJava browser was developed,
Sun contemplated undertaking the revisions and improvements necessary to
maintain it as a competitive product for desktop computers such as Windows PCs.
However, after Microsoft announced that its Internet Explorer browser would
always be given away for free, Sun concluded that it made little business sense at
that time to compete vigorously to sell a consumer browser application to compete
against a product that was being given away for free." Gosling Dir. ¶ 37; see
also Gosling, 12/3/98pm, at 80:17 - 81:3 (testifying that Sun never sold
HotJava "as a commercial browser" because, "given that the market price
for browsers, those days, seemed to be zero, it hardly seemed like a
sensible thing to do.").
- Scott Bosworth testified that IBM did not, in early 1998, seriously consider
sourcing a browser for use with JavaOS from a supplier other than
Netscape or Sun or seriously consider building such a browser itself,
because IBM believed no such investment was likely or profitable. He
testified, "we all recognized the fact that anyone investing heavily into the
browsers [sic] at this point in time was a pure and risky expense level with
little return on that investment from a browser standpoint. No one
believed that we should go get in the browser business." Bosworth Dep.,
10/16/98, at 118:17 - 120:18 (DX 2609).
b. Microsoft's predatory pricing facilitated monopoly
recoupment and injured consumers
309. Microsoft's predatory pricing of its browser harmed consumers because it
contributed to the diminution of the platform threat Netscape posed and thereby
facilitated maintenance of Microsoft's operating system monopoly.
- Professor Fisher concluded that Microsoft is already recouping in the form of
preserving its monopoly power and that "its financial recoupment will occur from
preserving the returns to the monopoly power in operating system, returns that
might have been dissipated had it not acted in the way in which it did." Fisher,
1/12/99am, at 31:25 - 32:15.
- See generally infra Part VII.A.
6. The after-the-fact justifications Microsoft offered for its better-than-free pricing of Internet Explorer are pretextual and
inconsistent with the evidence
310. Microsoft's contemporary documents indicate only a concern with
eliminating Netscape as a platform threat. Microsoft's very different, after-the-fact
explanations for its browser pricing are pretextual.
a. Microsoft's assertion that it reasonably expected its
browser-related expenditures to be profitable because of
expanded demand for Windows is pretextual
311. Microsoft's argument that it expected the free pricing of Internet Explorer to
be profitable because it would increase demand for Windows is pretextual.
311.1. First, there is no contemporaneous evidence that Microsoft
believed increasing demand for Windows would cover its immense browser
expenditures. To the contrary, Microsoft was concerned only with increasing browser
share.
- See supra ¶¶ 301-302.
311.2. Second, there is no evidence that the additional demand for
Windows created by making the browser free (rather than that demand created by
offering the browser at a positive price) could compensate for Microsoft's immense
browser-related costs.
- Professor Fisher testified: "There is no reason to believe and . . .
considerable reason not to believe -- that" the ancillary revenues
Microsoft obtains from its negative pricing of the browser can
"possibly lead to a recoupment of the amount of money that was
spent on the development of Internet Explorer." Fisher, 6/1/99am,
at 65:10-14; see also Fisher, 6/3/99am, at 25:18 - 26:1 ("The real
question is . . . was there value to Microsoft . . . beyond the value
that would have occurred had they charged separately for" the
browser "and . . . then allowed Netscape to be distributed more
widely." Although "there may be some" value, "there is" not "nearly
enough to account for what happened.")
- Professor Fisher also testified: "Among the other revenues that
Microsoft has claimed that it would get are revenues from
increasing the sales of Windows. But the sales of Windows would
have increased with any browser. And in any event, Microsoft gets
to claim, in the analysis, not all the ancillary revenues that it gets
from the sale of Windows because the browser way given away
free, and not all the ancillary revenues that it gets from the browser
anway. It gets to claim, at the most, the amount of ancillary
revenues of either type that it got because of what it did, that it
would not have gotten had it priced the browser separately."
Fisher, 6/1/99am, at 64:24 - 65:9.
311.3. Third, Microsoft's real-world conduct shows that it was not trying to
increase demand for Windows.
311.3.1. Demand for Windows is maximized by ensuring the
availability of all good complements (including browsers) and satisfying end-user
demand for a choice among complements.
- Professor Fisher testified that "the more things that will run
well on Windows . . . the more attractive Windows will be to
users." Professor Fisher further explained: "Ordinarily" the
producer of a product "would want to encourage other
people to produce better complements because that would
make" the product "better" and that giving "consumers a
choice" between complements is "going to increase demand"
for the product." Fisher, 1/7/99pm, at 52:19 - 54:2.
- Professor Fisher further testified: "As an analytical matter, if
browsers are a complement to operating systems such that
the sale of browsers that can be used with Windows will
increase demand for Windows, it should not matter who
makes the complement. But Microsoft cared greatly who
makes the complement . . . Microsoft even tried to
discourage Netscape from offering Netscape's browser for
use with Windows--an action inconsistent with browsers
being a complement to Windows, whose distribution
Microsoft wanted to maximize." Fisher Dir. ¶ 129.
- Dr. Warren-Boulton testified: "Microsoft has a legitimate interest
in ensuring that Windows users are able to acquire high quality
browsers at low prices, because that would increase the demand for
Microsoft's operating system. But even if achieving this objective
were furthered by Microsoft's decision to offer a quality browser
product, its further efforts to increase IE's share by excluding
Netscape and making it more difficult for users to obtain
Netscape's browser could only reduce the value of its operating
system to consumers." Warren-Boulton Dir. ¶ 187.
- See also supra Part V.B.3.(c).(1); ¶ 156.1.
311.3.2. There is thus no reason for Microsoft to favor Internet
Explorer over other browsers in order to increase demand for Windows.
- Professor Fisher testified that "it may be true that having
browsers widely distributed increases the demand for Windows.
That doesn't mean necessarily either that that browser has to be IE,
nor even to provide the integrated in the seamless experience way
that I mentioned before. Nor does it imply that it is profitable for
Microsoft to have done that and give it away." Fisher, 1/7/99am,
at 46:11-17.
- Indeed, "As an analytical matter, if browsers are a complement to
operating systems such that the sale of browsers that can be used
with Windows will increase demand for Windows, it should not
matter who makes the complement. But Microsoft cared greatly
who made the browsers used with Windows." Fisher Dir. ¶ 129.
311.3.3. But Microsoft took acts, to impede users' choice among
browsers and to impede the distribution and development of other browsers, which it
would not have taken were its objective increasing demand for Windows.
- Cameron Myhrvold conceded that Microsoft imposed its
exclusionary restrictions on ISPs becuase it was afraid that,
if users were provided a side-by-side choice of Internet
Explorer or Netscape Navigator, users would chose
Navigator. Myhrvold, 2/10/99am, at 62:7 - 64:20.
- Paul Martiz admitted that Microsoft sought to get companies
to agree not to promote Netscape's browser. Maritz,
1/26/99am, at 53:16 - 54:16.
- See supra Part V.A-F (detailing exclusionary practices).
- Professor Fisher testified that, if "Microsoft was interested in
increasing the sales of Windows, "it would surely have no
interest in restricting" the distribution of other browsers,
"since people who wanted to use the Netscape browser with
Windows would be happier" with Netscape. Fisher,
6/1/99am, at 65:24 - 66:8.
- Professor Fisher also testified that if "Microsoft were really
interested in selling Windows, it wouldn't have any interest
in" imposing its shipment restrictions in ISPs, which "require
that the ISP not ship more than, in this example, 15 percent
of other browsers." And Microsoft "can't have any interest in
doing that to protect its, quote, sales of IE, end quote,
because it doesn't have any, quote, sales of IE, end quote.
It's a no-revenue product." Fisher, 6/1/99am, at 66:12-25.
- Professor Fisher further testified: "Microsoft was
preoccupied not with increasing total sales of browsers but
with Microsoft's share of browser sales. Indeed, Microsoft
studied, and tried to implement, ways to disable Netscape
and reduce total browser sales. This conduct doesn't 'make
sense from a business standpoint' if browsers are viewed as
a means of increasing sales of Windows. But this conduct
makes good sense if browsers are viewed as a competitive
threat to Microsoft's Windows monopoly." Fisher Dir. ¶ 129
(emphasis in original).
- Dr. Warren-Boulton testified: "Microsoft has a legitimate
interest in ensuring that Windows users are able to acquire high
quality browsers at low prices, because that would increase the
demand for Microsoft's operating system. But even if achieving
this objective were furthered by Microsoft's decision to offer a
quality browser product, its further efforts to increase IE's share by
excluding Netscape and making it more difficult for users to obtain
Netscape's browser could only reduce the value of its operating
system to consumers." Warren-Boulton Dir. ¶¶ 187, 189.
312. Microsoft's related contention -- that its negative pricing was part of a
profitable plan to distribute widely the platform-aspects of its browser, including APIs, in
order to increase demand for Windows -- is also pretextual.
312.1. First, because Internet Explorer lacked APIs when Microsoft
committed to giving it away forever free, this contention cannot explain Microsoft's
actions.
- Brad Chase testified, "In August 1995, Microsoft embarked on a
redesign and rewrite of Internet Explorer from the ground up. Our
objective was to rebuild Internet Explorer as a set of separate
components, a process known as componentization." Microsoft
dedicated a team of more than 100 developers to the development
of this product which eventually was released as Internet Explorer
3.0. Chase Dir. ¶¶ 18-20.
- William Poole testified that "Microsoft began offering a
'componentized' version of its Internet Explorer technologies in
August 1996 with the release of Internet Explorer 3.0." Poole Dir. ¶
127.
312.2. Second, Microsoft spent millions developing Internet Explorer for
other operating systems, where it neither exposes APIs nor increases demand for
Windows 95/98, and gave those versions of Internet Explorer away for free, made it
more difficult for users to employ other browsers on other operating systems, and paid
users to use Internet Explorer rather than other browser on those operating systems.
- See supra Part V.B.1.c.(1); ¶ 113.2.2 -.3.
- Microsoft executive John Messerly wrote to Ben Slivka in June
1995 that "the importance of your browser achieving dominance will
in places override other (in this case systems) interests. Systems
want to show that Windows is as good if not better Multimedia
platform than Mac . . . In some respects, having an Ohare broswer
[sic] that screams as fast on the Mac as it does on Windows works
against that goal. But let's not loose [sic] our sense of proportion
about what this downside cost is though. It's not like netscape
won't be making their mac client as fast as possible, or like other
groups in MS aren't making their products as fast as they can
possibly be on Mac. The benefits of winning the browser war
outweigh the minor costs of making the Mac version as good as
and in lock step with the rollout schedule of the windows version."
GX 332.
- As Professor Fisher testified: "Whatever the relevance of
Microsoft's arguments about why it wanted to make Internet
Explorer available to sell more copies of Windows, those
arguments cannot apply to Microsoft's efforts to force Apple to
distribute Internet Explorer." Fisher Dir. ¶137. "Microsoft devoted
substantial time, effort, and money to developing and distributing a
version of IE for Apple computers. Microsoft gets no money from
increasing sales of Apple's operating system; indeed, since Apple
offers the main alternative to a PC using Windows, promoting
complements to Apple that increase Apple's attractiveness to users
reduces sales of Windows." Fisher Dir. ¶ 129.
312.3. Third, Microsoft could have included the APIs in Window itself and
sold the browser at a positive price.
- Professor Fisher testified, based on evidence that the "consumer
gets the same benefits if it acquires" the browser and operating
system separately and combines them, that "there is no reason why
Microsoft shouldn't offer them typically separately throughout and
let consumers decide, if those are really good benefits." Fisher,
6/1/99am, at 43:15 - 44:12; see also Fisher, 6/1/99am, at 42:1-6
("There appears to be no particular reason why Microsoft could not
have offered its browser, both together with the operating system
and separately, and offered the operating system separately, all of
these things at different charges. And because consumers wanted
it, that would have been a profitable thing to do.").
312.4. Fourth, Microsoft's free provision of Internet Explorer cannot be
explained as an effort to "stabilize" APIs because Microsoft continues to destabilize the
APIs through its frequent Internet Explorer and Windows updates.
- Professor Fisher testified that "it's not obvious that" the "API's have
to be Microsoft's API's for there to be a stable set of API's offered
to Developers" and "Microsoft's API's are not, in fact stable. They
change. And ISV's have to keep embedding pieces of the
appropriate APIs into their own software in shipping it out." Fisher,
6/3/99am, at 22:3-14.
- See supra Part V.B.3.d.(2); ¶ 164.4 (Microsoft fragments its
platform by updating Internet Explorer APIs).
b. Microsoft's argument that ancillary revenues explain its
better-than-free pricing of Internet Explorer is pretextual
313. Microsoft also argued that it expected to recoup its browser-related
expenditures from "ancillary revenues," such as revenue from search-engines (Maritz
Dir. ¶ 306; see also Schmalensee Dir. ¶ 556). This argument is implausible.
313.1. There is no contemporaneous evidence to support this argument.
- Professor Fisher testified: "Microsoft's document do not say, 'we're
doing this with Internet Explorer because Internet Explorer is going
to bring in a lot of money.' In fact, contemporaneous documents do
not suggest that Microsoft cared at all about -- and some of its
actions also confirm this -- that Microsoft cared at all about the
ancillary revenues that might" be "derived from giving way Internet
Explorer." Rather, "Microsoft's documents are full of statements"
that "'This is a no revenue product, but you should care about it just
as much as does Bill Gates. Without winning the browser war, we
lose.'" Fisher, 6/1/99am, at 40:7-25; see also Fisher Dir. ¶ 130, at (f)
("Microsoft's contemporaneous documents make clear that the company's
zero (or negative) price for its browser was not considered a way to earn
competitive ancillary revenues but a way to prevent potential competition
from alternative platforms."); Fisher, 1/7/99am, at 17:6-9 ("I do know
there is not a hint in the contemporaneous Microsoft documents that that's
what they were thinking about in terms of these revenues. That appears to
have been invented in the middle of this trial."); Fisher, 1/7/99am, at
17:13-18 ("So far as I know, there was, up to the beginning of this trial--I
may be wrong about this--but so far as I know, there are no Microsoft
documents that say we're doing this in order to get the alternative--the
ancillary revenues. The documents are full of statements about we're
doing it to protect the desktop.").
- Professor Fisher further testified that, "in terms of what Microsoft
thought it was doing -- if Microsoft was doing this stuff with the
browser because of the ancillary revenues, you would expect there
to be contemporaneous documents or business plans that show
that's why they're doing it. They [sic] wasn't anything like that, so
far as I know." Fisher, 6/1/99am, at 64:2-8; see also Fisher,
6/1/99am, at 68:10-13 ("If Microsoft was undertaking" its "campaign
to have ancillary revenues, you would expect them to be able to
produce some records that show that the ancillary revenues were
going to be sufficient to justify the costs.").
- Dr. Warren-Boulton testified that calculations made today, rather
than when Microsoft made its business decisions, are not relevant:
"if Microsoft performed such a calculation today and determined
that it earned substantial ancillary revenues from increasing its
browser usage share, that result would not be meaningful unless it
could be shown to provide a reliable guide to what reasonably
could have been anticipated by Microsoft at the time the decision
was made." Warren-Boulton Dir., ¶ 193.
313.2. Microsoft's actual conduct is inconsistent with that its ancillary
revenue explanation.
313.2.1. First, Microsoft declined to take advantage of significant
browser-related ancillary revenue opportunities.
- Professor Fisher testified: "Microsoft, in doing what it does
with its browser, from time to time took actions which, in fact,
gave up part of the ancillary revenues." Fisher, 6/1/99am, at
64:9-11.
313.2.2. Microsoft allows other firms to collect ancillary revenues
derived from the Internet Explorer start page.
- If Microsoft were really in the business for ancillary revenues,
Professor Fisher testified, it would not permit people to
change the default start page. Fisher, 1/12/99am, at 35:21 -
36:21.
- Joachim Kempin testified that, although Microsoft's OPK
prohibits OEMs from changing the Internet Explorer "start
page," he "understands" that this requirement "is not
enforced." Kempin, 2/25/99pm, at 92:7 - 94:23; GX 1201. In
fact, he testified, "I know that one OEM, in particular, has
asked me if they could" change the start page, "and I said
yes . . . . I think that was Dell." Kempin, 2/25/99pm, at 94:9-23.
313.2.2.1. Microsoft allows producers of browser shells and
(ignoring its own licensing terms) OEMs to collect ancillary revenues that Microsoft
would otherwise derive from the Internet Explorer start-page.
- Microsoft permits OEMs to install "shell browsers,"
such as Encompass, which are not actually browsers
but rather shells that sit on top of the Internet Explorer
browser and present a different user interface. These
shells rely on the underlying Internet Explorer
technology. See supra Part V.C.1.b; ¶ 185.2.1. The
shell displays the OEM's own brand, not Microsoft's,
and can be configured to point to any start page of the
OEM's choosing. The OEM and its shell-browser
partner keep any revenue it earns from selling
advertising on the start page. Warren-Boulton,
11/30/98am, at 72:3 - 74:21.
- Kempin's videotape demonstrated the Encompass
browser shell, which is built "on top" of Internet
Explorer but is customizable by third parties in ways
that allow them, rather than Microsoft, to capture
significant ancillary revenues. DX 2163.
- Microsoft represented, during the cross-examination
of Mr. Harris, that Dell is allowed to set Excite as the
default browser for Dell customers who connect to the
Internet through Dell's new Connect Direct [sic; actual
transcript reads Correct Direct], and that Compaq has
"exactly the same deal with Yahoo." Harris,
1/5/98am, at 25:18-25; see also DX 1842 (HP ships
both Internet Explorer and Encompass shell).
- Dr. Warren-Boulton testified that the Encompass
browser shell shows that: "What Microsoft is trying to
do here is to increase the percentage of IE
technologies based" on "the IE browsers, not, as is
clear," to "make a lot of money off the Internet in the
sense of advertising." Warren-Boulton, 11/30/98am,
at 79:22 - 80:19. He further explained that "the
question is at the time that they were making these
decisions, why is it that they want to increase browser
share. And I think, that, you know, it speaks exactly
to the point that you're making. Microsoft cares a
great deal about having people use browsers that use
IE technologies, even though, as you're pointing out,
there is no direct revenue to Microsoft from
advertising or other sources." Warren-Boulton,
11/30/98am, at 80:20 - 81:4.
313.2.2.2. Microsoft permits browser licensees, such as
ISPs, OLSs, and ICPs, to change the Internet Explorer start-page.
- According to the testimony of Cameron Myrhvold,
Microsoft's Internet Explorer Administration Kit
enabled ISPs to preset the default homepage so that
customers would be taken to the ISP's web site
whenever they logged onto the Internet. Myrhvold
Dir. ¶ 33.
- Professor Fisher testified that the ancillary revenues
Microsoft sacrificed included permitting "OLS's to take
their subscribers directly to the OLS's home page and
not to Microsoft's home page. That gives up some of
the" ancillary "revenues." Fisher, 6/1/99am, at 64:9-19; see also Fisher, 6/1/99am, at 69:11-18 (giving the
example of AOL).
- On cross examination of William Harris of Intuit,
Microsoft represented through its questions that
Microsoft is taking steps to make it easier for
consumers to change their browser home page.
Microsoft's lawyer asked whether Harris is aware that
Internet Explorer 5.0 "will enable web sites to display
a button that says `make this page your browser
default start page,' and all you have to do is click on
that button to change the home page automatically?".
Harris, 1/5/99am, at 27:22 - 28:2.
- Microsoft's William Poole, who was "attempting to be
helpful," told Mr. Harris at the break during his cross
examination that in Internet Explorer 5.0, Microsoft
"was making it easier to change the default browser
page and that, in fact, that was their strategy and
intent across many different venues with ISPs, with
OEMs, etc., making it easier for them to set defaults
rather than Microsoft." Harris, 1/5/99am, at 42:5 -
43:7.
313.2.3. Microsoft's indifference to collecting these ancillary
revenues stands in stark contrast to the practice of other firms which, unlike Microsoft,
have no incentive to sacrifice such revenues in order to preserve monopoly power.
- Dr. Warren-Boulton testified that the contrast between
Microsoft's incentives to sacrifice ancillary revenues by
allowing shells to be built on its browser and Netscape's
incentives instead to try to collect any available revenue from
its browser "is a nice example of the distinction between
what Netscape is trying to do in the browser market, which is
to make money, and what Microsoft is trying to do in the
browser market, which is to control the technologies."
Warren-Boulton, 12/1/98am, at 13:12 - 14:11.
313.3. Second, the ancillary revenues arguably associated with the zero
price for Internet Explorer are insubstantial.
313.3.1. Revenues from search engine contracts and the like are
not large enough to cover Microsoft's browser-related expenditures, nor are they
appropriately attributed wholly to Microsoft's browser; they certainly are not attributable
to the free pricing of Microsoft's browser.
- Microsoft represented, during Professor Fisher's cross-examination, that Microsoft receives $15 million a year from
each of two search engines, Altavista and Lycos, just for
placing the search engines on the "MSN web search" menu
of MSN.com, MSN's start page. Fisher, 1/7/99am, at 50:18 -
52:18. But, as Professor Fisher testified, such revenues
should not be attributed to Internet Explorer because
Microsoft earns them for placement on the MSN start page.
Fisher, 1/7/99am, at 53:8-19.
- Even if one looked at the ancillary revenues that Microsoft
receives today from the browser, one must only look at those
revenues that Microsoft makes because of the "better than
free" pricing of the browser. Professor Fisher noted that "the
real question is . . . was there value to Microsoft . . . beyond
the value that would have occurred had they charged
separately" for the browser "and then allowed Netscape to
be distributed more widely." Although "there may be some"
value, "I don't think there is nearly enough to account for
what happened." Fisher, 6/3/99am, at 25:18 - 26:1.
- Barksdale testified that Netscape makes money from portal
revenues, but that doing so does not require giving away the client:
"But we were doing that before. I mean, we would have that
revenue anyway." Barksdale, 10/27/98pm, at 23:23-24.
- Furthermore, Barksdale testified, the ancillary revenues he
hopes "offset some of that cost" of the browser, but relying
on those to cover the entire costs is not "economically viable.
And by the way, you would never start a business with that
business plan, I don't think." Barksdale, 10/27/98pm, at
23:12 - 24:4.
313.3.2. There is no evidence that other ancillary revenues, such
as selling more servers or advertising for other products, could cover Microsoft's
immense browser-related expenditures.
- Professor Fisher testified: "There is no reason to believe and
some reason not to believe-- and considerable reason not to
believe -- that" the ancillary revenues Microsoft obtains from
its negative pricing of the browser can "possibly lead to a
recoupment of the amount of money that as spent on the
development of Internet Explorer." Fisher, 6/1/99am, at
65:10-14.
313.4. Third, the examples of firms giving away products for free, to which
Microsoft points, cannot explain Microsoft's very expensive effort to build a dominant
browser share.
- Professor Fisher testified: "It is sometimes the case
that, for various reasons -- sometimes it's introductory
offers; sometimes it's for reasons of expanding the
market; sometimes because of the selling of ancillary
products -- that companies will give away or sell very
cheaply things which lead to those ends. And if that's
all that Microsoft had done, we wouldn't be here
today. But that's not what happened." Fisher,
1/7/99am, at 44:23 - 45:5.
313.4.1. Most of the examples Microsoft cites (Maritz Dir. ¶¶ 278-306; Maritz ¶ 313) are very different from its commitment to give away Internet Explorer
"forever free" because, among other things, they are associated with specific,
anticipated other revenues.
313.4.1.1. Apple. Apple charged for advanced versions of
QuickTime, something Microsoft does not do with IE.
- Apple's Avadis Tevanian testified that, although Apple
gives its basic version of QuickTime away for free, it
charges $29.99 for its advanced version, QuickTime
3.0 Pro. Tevanian, 11/5/98am, at 6:16 - 7:7.
313.4.1.2. Intel. Intel in some instances licenses its
software for a fee, but in other instances gives it away. There is no evidence that Intel's
modest expenditures on free software could be recouped only by preserving monopoly
power.
- Intel, although it gave away much of its software in
order to raise "the capability of the overall personal
computer platform," nonetheless "in some cases . . .
did try to license" its "software for fees." McGeady,
11/12/98am, at 34:18 - 35:13.
313.4.1.3. Adobe. Adobe gives away its "viewer" for free in
order to charge for content-creation tools.
- Professor Fisher testified that, even though it may
make sense for Adobe to give away its viewer so it
can make money selling the authoring software, that
is not true of Microsoft's effort to gain browser usage
share. Fisher, 1/7/99am, at 42:25 - 43:14.
313.4.1.4. AOL. Although AOL gives away its access
software for free, it makes money on subscriptions (the only purpose of the access
software is to access AOL's service).
- Professor Fisher testified that AOL gave away its
Booklink software to subscribers as part of the
software that enabled them to take advantage of the
AOL service. AOL then earned subscription revenue
from the AOL service. Fisher, 1/6/99pm, at 10:24 -
11:23.
313.4.2. The fact that Netscape decided to reduce the price of its
browser to zero in response to Microsoft's zero price provides no basis to infer that
Microsoft's zero price is profitable.
313.4.2.1. Microsoft's zero price was established after
Netscape had already incurred the costs of developing its browser; and the issue
Netscape faced at that point was simply how it could cover its avoidable, future costs.
Microsoft, by contrast, sunk massive costs in developing, promoting, and distributing its
browser after it decided to make it "free forever."
- See supra ¶ 299 (describing MS's huge expenditure).
313.4.2.2. Netscape (and now AOL) also obtains substantial
portal revenues and, thus, unlike Microsoft -- which surrenders such portal revenues by
permitting the start page to be changed -- can give its browser away and still recover its
future costs.
- Although Dean Schmalensee testified that the portal
revenues described in documents concerning the
AOL/Netscape merger show a profitable plan to
distribute "browsing software at a substantial negative
price" because of portal revenues (Schmalensee,
6/21/99pm, at 51:8-21, 53:4 - 54:19, 56:11-24, 75:12-23; DX 2518), he ignored that
c. Dean Schmalensee's argument that predation is
implausible is flawed
(1) Dean Schmalensee greatly underestimates the
costs, and overstates the legitimate benefits, of
Microsoft's predatory strategy
314. Dean Schmalensee argued that Microsoft could not possibly have engaged
in predation because only a modest increase in either the price for or sales of Windows
would make its actions profitable. Schmalensee, 6/21/99pm, at 56:25 - 62:15; DX 2763;
DX 2764; Maritz Dir. ¶ 36. This analysis is flawed.
314.1. First, Dean Schmalensee's analysis looks to benefits that occurred
after the fact; but predation analysis is not properly based on hindsight.
- Dean Schmalensee conceded that a predation analysis properly
examines expected revenues and costs yet admitted that he did not
"make an analysis of what revenues, if any, Microsoft expected to
receive from or as a result of the browser at the time that Microsoft
was developing its Internet Explorer browsers." Schmalensee,
6/24/99pm, at 15:5-19.
- Dr. Warren-Boulton testified: "I have seen no documents indicating
that Microsoft ever performed such a calculation at the time these
decisions were made. (Indeed, if Microsoft performed such a
calculation today and determined that it earned substantial ancillary
revenues from increasing its browser usage share, that result would
not be meaningful unless it could be shown to provide a reliable
guide to what reasonably could have been anticipated by Microsoft
at the time of the decision was made.)." Warren-Boulton Dir. ¶ 193.
- Professor Fisher testified that "what matters is what is expected (or can
reasonably be expected) at the time the action in question is taken." Fisher
Dir. ¶ 49.
314.2. Second, Dean Schmalensee drastically understates the costs of
Microsoft's predatory campaign.
- Dean Schmalensee took as the cost to Microsoft of Internet
Explorer the development costs of $100 million a year.
Schmalensee, 6/21/99pm, at 59:18 - 60:2; Schmalensee,
6/24/99pm, at 16:24 - 17:8.
- But as explained, the actual costs of the predatory campaign --
including the amounts Microsoft paid third parties to distribute its
browser and not to distribute other browsers -- were substantially
larger. See supra ¶ 299.4.
- Dean Schmalensee conceded that he did not seek to account for
the costs of marketing Internet Explorer or the opportunity costs
Microsoft incurred to increase its browser share. Schmalensee,
6/24/99pm, at 17:13 - 18:16. Nor did Dean Schmalensee take into
account assets Microsoft bartered for exclusion, such as desktop
placement (Schmalensee, 6/24/99pm, at 26:9 - 32:14) even though
he conceded that marketing costs should be taken into account.
Schmalensee, 6/24/99pm, at 17:13 - 18:24.
314.3. Third, Dean Schmalensee overstates the benefits because he
includes benefits Microsoft would have obtained even if it did not set a zero price for
Internet Explorer.
- Dean Schmalensee points to all the ways Microsoft has assertedly
"improved" Windows, including adding Internet Explorer
(Schmalensee, 6/21/99pm, at 62:6 - 69:22; DX 2764;
Schmalensee, 6/22/99pm, at 7:20 - 8:15). But he did not analyze
how much Microsoft would still have "grow[n] the Windows
business" (Schmalensee, 6/21/99pm, at 69:12-17) had it
nonetheless charged for Internet Explorer at a positive price.
- Professor Fisher testified that the only revenues that are properly
taken into account as benefits to Microsoft from its zero price are
those Microsoft could not have obtained by charging a positive
price: "In figuring out whether or not Microsoft's actions were
predatory, one should certainly take account of the ancillary
revenues which it reasonably expected to earn as a result of those
actions. But you don't get to count all those revenues as though
they wouldn't be there had Microsoft taken some other action,
because if Microsoft had sold its browser at a separately stated
price, there would still have been some amount of those ancillary
revenues which it would then have achieved. And those have to be
offset against the ones that are achieved by giving it away. You
also, of course, have to balance that against what it would then
have received for the browser had it sold." Fisher, 1/12/99am, at
37:22 - 38:8.
314.4. Fourth, Dean Schmalensee's refusal to examine why Microsoft
actually undertook its better than free pricing (Schmalensee Dir. ¶¶ 337-338) renders
his analysis unreliable.
- Dean Schmalensee previously endorsed "only one economically
defensible general policy choice: Scherer's proposal that courts
follow a rule-of-reason approach and perform a thorough
examination of the factual circumstances accompanying the
monopolist's alleged predatory behavior, how the monopolist's
officials perceive the probable effects of its behavior (i.e., Intent),
and the structural consequence actually flowing from the behavior."
GX 2334, at 1028.
- Dean Schmalensee sought to distance himself from this article by
asserting that he is "less comfortable" inferring intent from behavior
except "when . . . one has a smoking gun -- or a warm smoking
gun." Schmalensee, 6/24/99pm, at 43:14-24. But, he conceded,
"the better the intent evidence, the stronger the weight it ought to
have." Schmalensee, 6/24/99pm, at 44:23-25.
- Despite this concession, Dean Schmalensee wholly ignored in his
analysis the contemporaneous statements of Microsoft executives
that they were giving away the browser, not to expand demand for
Windows or to garner ancillary revenues, but rather for the specific
purpose of blunting the browser threat to its operating system
monopoly. See supra ¶ 301.
(2) Dean Schmalensee is wrong that successful
predation required eliminating Netscape
315. Dean Schmalensee argued that no predation has taken place because
Netscape has not been eliminated as an important browser producer (Schmalensee,
6/21/99pm, at 32:4 - 33:21); but this testimony is misconceived because eliminating the
threat Netscape posed to Microsoft's operating system monopoly required only
preventing Netscape from obtaining a dominant browser share, not driving it from the
browser market altogether.
- See infra Part VII.A.1; ¶ 359.
(3) Dean Schmalensee is wrong that predation is
implausible on the ground that AOL "holds the
key" to the browser market
316. Dean Schmalensee argued that it is implausible that Microsoft engaged in
predation because AOL could, anytime it chose, confer on Netscape a large share in
browsers. Schmalensee, 6/21/99pm, at 88:9-11 (testifying that "AOL holds the key to
browser share"); see also Schmalensee, 6/21/99pm, at 85:12-17; Schmalensee Dir. ¶¶
541-549). This argument is unsound.
316.1. First, Microsoft successfully predated in part because it paid AOL
to distribute Internet Explorer instead of Netscape. Microsoft's and AOL's incentives to
continue a similar arrangement in the future are not diminished by AOL's acquisition of
Netscape.
- Professor Fisher testified that "if, indeed, there is any effect of the
merger" it is "that Microsoft will have to give up some of its
monopoly rents to AOL." Fisher, 6/3/99am, at 20:19-24.
<
- See infra VII.C.2; ¶ 395.2.
316.2. Second, Microsoft's demonstrated ability to engage in predatory
conduct to crush incipient platform threats in any event makes unlikely the possibility
AOL will cease distributing Internet Explorer in order to challenge Microsoft's operating
system monopoly.
- See infra Part VII.C.2; ¶ 394.1.
(4) Dean Schmalensee is wrong that predation is
implausible because other threats to Microsoft's
operating system monopoly might exist or arise
317. Dean Schmalensee's testimony that predation is implausible because,
even if Microsoft successfully eliminated the browser threat, other threats would prevent
it from exercising monopoly power (Schmalensee, 6/21/99am, at 86:7-17; Schmalensee
Dir. ¶ 553), is also flawed.
317.1. First, Microsoft's operating system monopoly is protected by high
entry barriers, and the Netscape browser threat presented an unusual risk to Microsoft's
position. The other alleged threats are less serious today and may depend on the
success of non-Microsoft browsers to develop.
- See supra Part III.D; ¶¶ 60-62.
317.2. Second, Microsoft recoups the costs of its predatory conduct by
reducing the probability that Windows will be displaced and thus increasing the value of
its monopoly; that recoupment occurs from the outset of the predatory campaign.
- Professor Fisher testified: "Microsoft is now . . . recouping in the
form of . . . increasing freedom from the threat of losing its
monopoly power. . . . It's financial recoupment will occur from
preserving the returns to the monopoly power in operating system,
returns that might have been dissipated had it not acted in the way
in which it did." Fisher, 1/12/99pm, at 31:15 - 32:7.
- Professor Fisher further testified that "of course, one cannot know
with any kind of certainty when or even whether the threats from
Java and the browser would have led to a breakdown of the
applications barrier to entry, and, therefore, to more competition in
operating systems. And maybe the answer is never. But Microsoft
didn't give it a chance to try. And it's managed . . . to preserve its
monopoly profits into the foreseeable future." Fisher, 1/12/99am,
at 32:8-15.
- Dean Schmalensee's own charts show that Microsoft's continued
dominance of the operating system market for even a short period
of time as a result of its anticompetitive conduct would result in
immense profits. DX 2763; Schmalensee, 6/24/99pm, at 22:16 -
24:19 (conceding that Microsoft could recoup the costs of a $600
million campaign through a $9 increase in the price of Windows).
317.3. Third, Microsoft's predatory conduct has deterred, and will
continue to deter, other threats from arising.
- See infra Part VII.D; ¶¶ 402-403.
VI. Microsoft Used Predatory and Anticompetitive Conduct to Impede Other
Platform Threats as Well, Thereby Further Entrenching Its Operating
System Monopoly
A. Microsoft responded to the threat that Java posed to the applications
barrier to entry by engaging in predatory and anticompetitive
conduct
318. As explained, Java technology, both alone and in concert with non-Microsoft browsers, poses a threat to Microsoft's operating system monopoly because it
holds out the possibility of an effective cross-platform middleware that can significantly
reduce the applications barrier to entry.
- See supra Part III.C; ¶¶ 57-59.
319. Microsoft recognized the Java threat and, in conjunction with its effort to
blunt the browser threat, engaged in a series of actions designed to interfere with the
development, distribution, and usage of cross-platform Java.
319.1. Microsoft developed and then widely distributed, in part through
Windows, a "polluted" version of Java that is not cross-platform.
- See infra Part VI A.1; ¶¶ 313-314.
319.2. Microsoft's purpose in creating and widely disseminating its
version of Java was to fragment cross-platform Java and thus hinder the threat Java
could pose to the applications barriers to entry.
- See infra Part VI.A.2; ¶¶ 315-319.
319.3. In addition to polluting Java, Microsoft engaged in anticompetitive
conduct designed to cripple cross-platform Java.
- See infra Part VI.A.3; ¶¶ 330-332.
319.3.1. Microsoft, through its predatory campaign against
Netscape, weakened cross-platform Java's principal distribution vehicle.
319.3.2. Microsoft used its monopoly power to force third parties to
support its version of Java and to mislead developers into creating Windows-specific
programs.
319.3.3. Microsoft induced third parties not to support cross-platform Java and to support exclusively Microsoft's Java implementation.
7. Microsoft "polluted" Java by developing and distributing a
version that is not cross-platform
320. The first step Microsoft took to diminish the cross-platform threat Java
posed to the applications barrier to entry was, in its own words, to "pollute" Java by
developing a variation of Java that is not cross-platform.
- An internal planning memorandum for Microsoft's Java development tools
for 1998 reaffirmed the objective: "Kill cross-platform Java by grow the
polluted Java market." GX 259, at 1.
- James Gosling testified: "Microsoft has made an incompatible implementation of
the Java technology that is not cross-platform, but instead is dependent on the
Windows operating system platform and Microsoft's proprietary technology . . . .
Microsoft has designed its implementation of the Java technology to impair the
ability of programs written to that implementation to run on non-Microsoft
platforms, or even to operate properly on the JVMs sold by other vendors for PCs
running Windows." Gosling Dir. ¶¶ 54-55. As Gosling further explained, this
effort was "part of the Microsoft program of making it so that people could port
into the Microsoft world but then sort of swallow the pill and not be able to get
out." Gosling, 12/10/98am, at 37:4-7.
- An internal Boeing Corporation presentation summarized the overall
impact of Microsoft's "pollution" of Java (implemented through the
inclusion of Microsoft's Java virtual machine in Internet Explorer) in
December 1997: "Internet Explorer will allow the development of Java
applications that are dependent on the Windows platform. These
applications may not run on other platforms." GX 639, at TBC 000569
(emphasis in original).
320.1. First, Microsoft refused to support standard Java native code
interfaces and developed, instead, Windows-specific interfaces.
320.1.1. As part of the evolution of Java, Sun and several of its
Java licensees developed a standard Java API called the "Java Native Interface" (JNI).
JNI enables Java programs to draw upon code unique to particular operating systems
in a way designed to minimize the cost of porting the programs to the various operating
systems.
- "The Java technology is in the process of maturing, and
occasionally a developer will encounter a situation in which the
programmer needs to write a portion of a program in a different
programming language in order to access functionality not yet
supported in the Java technology, but which may be available in
the underlying ('native') operating system or hardware. To do this,
the Java technology includes a 'Java Native Interface' ('JNI'). JNI
is a standard Java API that acts as a link between the JVM and the
platform-specific code included in an application to perform the
particular operating-system function. JNI thus gives programmers
a way to use native platform functionality with their Java-based
software." Gosling Dir. ¶ 15; Gosling, 12/9/98pm, at 50:23 - 51:10
("JNI is one that we, along with IBM and Netscape and a number
of other companies, worked to make JNI work with multiple Java
VMs."); Gosling, 12/9/98pm, at 44:16 - 45:21 (explaining that,
"while JNI doesn't solve the problem of making native code
transparent" it "goes a long way towards making that job much less
onerous, much less costly for developers, so that they can take their
. . . native code and move it from one system to the next much
more cost-effectively.").
- Sun's public statements emphasize the cross-platform significance
of the JNI approach to native interfaces. DX 2014 (Sun web site,
"100% Pure Java Cookbook:" "The Java Native Method Interface
(JNI) is not a way to make native code platform-independent; it is
a way to make it easy to port native code."); DX 1944 (Sun press
release, April 97: "The JNI standard may give a native library its
'best chance' to run in a given JVM, according to JavaSoft.").
- See also Gosling, 12/3/98am, at 9:1-5; Soyring, 11/18/98am, at 73:19 - 76:6 (explaining that JNI
is a "tremendous value" because it offers "the possibility of writing
an application once, compiling it once and using the jni interface to
access functions on different operating systems"); Fisher,
1/7/99pm, at 27:14-20 (testifying that JNI permits developers to
call native OS functions "in a way that the code written to the JNI
is, as it were, hived off, so that if you want to write for a Java
virtual machine for a different underlying platform, it is relatively
easy to port it").
320.1.2. Microsoft, however, refused to support JNI in its
implementation of Java (until ordered to do so by the District Court for the Northern
District of California).
- "Microsoft omitted from its implementation of the Java technology
a standard API called 'JNI,' that permits platform-specific software
code to interact with Java code in a program." Gosling Dir. ¶ 58.
- Sun Microsystems, Inc. v. Microsoft Corp., 21 F. Supp. 2d
1109, 1127-28 (N.D. Cal. 1998) (requiring Microsoft to
include JNI in its Java environment).
320.1.3. Instead, Microsoft developed its own native interfaces,
known as "J/Direct," the "Raw Native Interface" (RNI), and "Java/COM." These
interfaces are designed to allow developers to link directly to Windows-specific DLLs (in
the case of J/Direct), Microsoft's proprietary and Windows-specific Component Object
Model (COM) architecture (in the case of Java/COM), and other Windows-specific code
(in the case of RNI).
- Gosling testified: "Microsoft substituted its own proprietary
interfaces, called RNI, J/Direct, and @COM. By omitting the
standard API, Microsoft prevents developers from writing a
Java-based program that can run on every JVM
implementation. If the developer uses JNI, the software will
not run on Microsoft's JVM; if the developer uses Microsoft's
proprietary RNI, J/Direct, or @COM interfaces, the software
will not run on any other JVM. Moreover, Microsoft's
proprietary interfaces require the use of Microsoft's
development tools." Gosling Dir. ¶ 58.
- Robert Muglia admitted that J/Direct (and its accompanying
compiler directive, @DLL) is designed "to provide
developers with additional development tools that make it
easier, faster and more efficient for them to use the Java
programming language to build full-featured Windows
applications." Muglia Dir. ¶ 122 (emphasis added).
- As Intel's Alan Holzman noted at a meeting with Microsoft in
April 1996: "MS has changed native interfaces. Not
compliant with Sun's current native interface at the binary
level: As part of the rewriting the Java VM, MS has
completely changed the internal object model to
accomodate COM. We think they have not told Sun and this
may be an issue for Sun as well as us if Intel and MS give a
single optimized Intel Architecture Java RT back to Sun.
The current method for calling native methods in the MS
Java VM is different and not compatible with Sun's. MS will
provide include files that allow you to bridge between the
two at the source code level but not at the binary level." GX
566.
- See also Muglia, 2/26/99pm, at 81:21 - 82:16 (explaining J/Direct
and testifying: "I'm not aware of any other operating systems that
support J/Direct, although I do know that we openly publish the
specifications so that others could take advantage of it."); Gosling,
12/10/98am, at 27:17 -28:23 (making Microsoft's native
interfaces available to work with other operating systems is
"not a trivial engineering effort" inasmuch as "there is
essentially no way to implement them without
reimplementing some significant part of Windows on
whatever your system happens to be"); Gosling,
12/10/98am, at 43:20 - 44:4 (testifying that "we certainly did
have discussions about building a solution based on COM."
Gosling further testifying that: "but as I recall, there was no
real interoperability that was ultimately possible, since COM
only existed on the Windows platform."); Gosling,
12/9/98pm, at 55:24 - 56:14 (MS's RNI interface was based
on COM, and "didn't satisfy our needs for portability because
it was an interface that could only work on Microsoft's
operating system since that was the only place where . . .
COM support existed."); Gosling, 12/9/98pm, at 76:12-15
(Microsoft talks about making COM cross-platform, but
hasn't done so).
320.2. Second, Microsoft created Windows-specific extensions to Java in
its Java development tool products which, when utilized by developers, caused
programs to work with only the Windows operating system.
- Gosling testified: "Microsoft extended the Java programming language in
ways supported only by Microsoft's incompatible implementation of the
Java technology. This is analogous to adding to the English language
words and phrases that cannot be understood by anyone else. Specifically,
Microsoft did two things. Microsoft added support for incompatible
'keywords,' or additional Java programming instructions, to its
implementation. And Microsoft added to its implementation support for
'compiler directives,' or special comments inserted in Java sourcecode that
alter the behavior of the code when it is compiled into Java bytecode.
Among other things, Microsoft's compiler directives permit the Java
Virtual Machine to interact with or 'call' proprietary Microsoft Windows
APIs. Only Microsoft's incompatible implementation recognizes these
idiosyncratic keywords and compiler directives. Among other things,
some of these extensions prevent programs from being compiled by non-Microsoft compilers, and all such extensions prevent programs from
running as intended on non-Microsoft JVMs." Gosling Dir. ¶ 58.
- See also Gosling Dir. ¶ 63 ("Microsoft uses its development tools to
encourage developers to write Microsoft-dependent Java-based programs.
Microsoft bundles its Java development toolkit in its Visual Studio
development suite, which is the most popular set of Windows
development tools. Microsoft's Java-based development tools are shipped
with the company's incompatible extensions enabled by default.");
Gosling Dir. ¶ 69 (explaining that because "of the Microsoft dependencies
embedded in these toolkits, programs written using the toolkits can cause
Java-based programs to be tied to the Microsoft implementation,
undermining the cross-platform compatibility of the Java technology");
Gosling, 12/10/98pm, at 37:13 - 38:2 ("Microsoft engaged in changes to
their compiler and their development tools such that the binaries that it
produced . . . would not actually run on all the different VMs. The output
of their compiler in a large fraction of cases would only run on the
Windows platform and on the Windows Java virtual machine." Gosling
explained that "the output of their Java compiler, if you follow their
standard recommendations, what happened was you would end up with
something that was not cross-platformed. It was targeted only at their
operating system and their Java virtual machine."); Soyring,
11/18/98pm, at 49:23 - 51:9 (testifying that "use of compiler
directives and key words which are nonstandard key words that
have been added to the Java language").
320.3. Third, Microsoft declined to include the Java Remote Method
Invocation (RMI) class library, as standard Java component, in its implementation of the
Java Runtime Environment (JRE) distributed with Internet Explorer and Windows.
320.3.1. Microsoft refused to distribute RMI, a cross-platform
communications protocol, with the JRE included in Internet Explorer 4 and thus
Windows 95 and Windows 98.
- Gosling explained that "RMI is a way for a Java program
running in one place to send a message to a Java program
in another place and have something happen, and maybe
some communication comes back and forth. It's really a
communication mechanism for two entities to just send
messages back and forth." Gosling, 12/3/98pm, at 66:3-7.
- Soyring testified that Microsoft "deviates in several areas
with what we understand the standard to be as proposed by
Sun Microsystems, and for which case, we need to run test
cases to insure the compatibility of our Java applications
across different platforms. That . . . includes the omission in
the Internet Explorer and Windows 98 product of a function
called RMI, Remote Method Invocation, which is used in
many of our applications." Soyring, 11/18/98pm, at 50:16 -
51:1.
320.3.2. Indeed, Microsoft made RMI available only from an
obscure portion of its web site, where it was difficult for consumers and developers to
find.
- As an August 1997 e-mail from Microsoft's James Van
Eaton indicated, RMI was relegated to a relatively
inaccessible location: "There will be no entry in the index for
this file. They'll have to stumble across it to know it's there.
ftp://ftp.microsoft.com/ developr/msdn/unsup-ed is where I
put it on the Internet with 37 other old files in this directory.
I'd say it's pretty buried." GX 1931 (emphasis in original).
- See also Gosling Dir. ¶ 58(c) (testifying that the "only way for
a person using Microsoft's tools or its JVM to take
advantage of RMI functionality is for them to download and
install the RMI class libraries and RMI compilers from the
Internet. While some technically proficient and
knowledgeable developers may be able to find these RMI
technologies on the web, most end users with a Microsoft
JVM will never know of their existence, let alone be able to
find, download, and install them."); Soyring, 11/18/98pm, at
51:1-6 ("Granted, this piece of code, RMI, is available from
Microsoft on their web site, but it's very difficult to find, and
our customers lack confidence that it would continue to be
supported unless it's in the base product, itself, as dictated
in at least our understanding of the Sun specification for
Java."); DX 2025 (PC Magazine observed that although
Microsoft "makes the RMI code available on its Web site,
this is much less convenient to access than if it were
included as part of the VM. Java programs that implement
RMI won't work on the Microsoft environment as shipped.").
321. As a result of Microsoft's refusal to support JNI and RMI, and its
development and distribution of Windows-specific interfaces and extensions to Java
development tool products, developers were required to choose between compatibility
with the Microsoft JVM (and thus Windows), on one hand, and compatibility with
virtually all other JVMs (which could run on a variety of operating systems, including
Windows), on the other hand. Applications that use Microsoft's interfaces do not run on
JVMs other than Microsoft's or on operating systems other than Windows.
- Gosling summarized the impact of the "choice" that Microsoft's refusal to
support JNI and its simultaneous provision of Windows-specific native
interfaces provides Java developers: "By omitting the standard API,
Microsoft prevents developers from writing a Java-based program that can run on
every JVM implementation. If the developer uses JNI, the software will not run
on Microsoft's JVM; if the developer uses Microsoft's proprietary RNI, J/Direct,
or @COM interfaces, the software will not run on any other JVM." Gosling Dir.
¶ 58.
- See also Gosling, 12/10/98pm, at 40:18 - 41:5 ("If you had used JNI to interface
with native methods, you would find that your native method interface would not
work at all with their virtual machine, and you would be faced with the task of
reengineering their native methods to use one of the three Microsoft-provided
native method technologies. And, of course, once you have gone through that
work of reengineering your native method code to use the Microsoft native
method code, you would find that you had landed in the trap because these native
methods using J/Direct or COM or RNI don't run anywhere else. So, once you
sort of slid in there, you have gotten yourself mired."); Gosling, 12/10/98pm, at
38:3-9 ("if you used some other development tool to develop a Java program, . . .
essentially, any Java program that you developed using native methods using the
standard development tools would not work on the . . . Microsoft virtual
machine"); Gosling, 12/10/98am, at 23:20 - 24:13 ("The situation that
Microsoft's actions put us in is one where developers, by and large, can
end up having to test twice: once on a certified VM to see that they will
work across, you know, Solaris, OS/2, Apple, or whatever, and then they
separately have to test for Microsoft because of the choices that Microsoft
has made.").
- Soyring testified that Mirosoft's hiding of RMI leads to a "lack" of
"confidence that it will continue to be supported," resulting in "a reluctance
to use RMI on other platforms because it would break the cross-platform
compatibility goal they are trying to achieve." Soyring, 11/18/98pm, at
89:4-13.
322. Microsoft, having developed its Windows-specific Java implementation bundled it
with Windows and Internet Explorer to ensure its wide-spread distribution.
- A January 1996 Microsoft presentation describes as a "Response Summary" to
cross-platform Java: "Increased IE share, Integrate with Windows." GX 52, at
MS 003270.
- IBM recognized that "Internet Explorer is Microsoft's primary weapon to kill
100% Pure Java." DX 1894, at IBM 60968.
- Gosling testified: "Microsoft has used its ubiquitous operating system to flood the
market with its Microsoft-dependent implementation . . . both in the version
available at retail stores, and in the version it licenses to PC manufacturers to
install in new PCs. These are powerful methods of distributing Microsoft-dependent JVMs, and they assure that Microsoft-dependent JVMs will be present
on a very large percentage of personal computers." Gosling Dir. ¶¶ 54, 66.
- Increasing Internet Explorer's share, Robert Muglia agreed, was a response both
to Netscape Navigator and to Sun's cross-platform threat in "the sense that the
Internet Explorer feature of Windows has APIs in it that are competitive with
Java." Muglia, 2/26/99pm, at 53:25 - 54:1.
- Microsoft's Ben Slivka testified that, because of the enormous market share of
Windows, shipping software with Windows could "create a de-facto standard."
As such, Slivka considered distributing Microsoft's implementation of Windows
as one of Microsoft's "levers for Java." Slivka Dep., 1/13/99pm, at 748:12 -749:6.
- See also Soyring Dir. ¶ 28 (Microsoft's tying arrangement also advantaged
"Microsoft's implementation of Java technology").
323. By distributing its Windows-specific version of Java widely, and forcing
developers employing it either to sacrifice the use of Windows code or Microsoft
interfaces on the one hand, or compatibility with multiple Java virtual machines on the
other, Microsoft encouraged a fragmentation of the Java platform that undermines the
prospects for success of cross-platform Java.
- IBM's John Soyring testified that customers "have expressed concern
about Microsoft's implementation of Java"; specifically, that "it deviates in
several areas with what we understand the standard to be as proposed by
Sun Microsystems, and for which case, we need to run test cases to
insure the compatibility of our Java applications across different platforms.
That includes the admission from the Microsoft implementation of a
function called JNI, which is an acronym standing for `Java Native
Interface.'" Soyring, 11/18/98pm, at 49:23 - 50:23.
- As Sun's Graham Hamilton wrote: "We visited Microsoft today to hear the
technical details of their plans to 'put java into Windows 95.'" "We asked
how Java ISVs will get at all the wonderful Microsoft APIs from within
Java. We used the new game APIs as an example. The answer is that
Microsoft will either provide Java class wrappers to C interfaces, or maybe
provide an automated way of accessing Win32 APIs via OLE/COM. This
will be done on a case-by-case basis as needed and/or requested.
Clearly this is a very disturbing answer. This basically means they will be
encouraging Java ISVs to write to a broad set of APIs that are only
available on Win32." DX 2053.
- Franklin Fisher summarized the impact of Microsoft's conduct with respect
to native interfaces: "Microsoft took actions to impede the cross-platform
potential of Java by developing an interface called J/Direct. Any
application that uses 'J/Direct will run only on the Microsoft virtual
machine.' The default way of writing applications and applets for
Microsoft's virtual machine causes some of those applications and applets
not to be able to run properly on non-Windows platforms or even on non-Microsoft virtual machines running on Windows." Fisher Dir. ¶ 208
(summarizing Microsoft documents).
- James Gosling testified: "By forcing computer manufacturers to bundle both a
browser and Microsoft's JVM, Microsoft has hindered the ability of other
browsers with compatible JVMs from getting distribution on Windows PCs."
Gosling Dir. ¶¶ 68.
8. Microsoft's purpose in polluting Java was to reduce the threat
that cross-platform Java posed to the applications barrier to
entry
324. The purpose of Microsoft's simultaneous exclusion of JNI and its
development and promotion of Windows-specific alternatives was to create
incompatibility and "kill cross-platform Java," not to enhance performance of Windows-specific programs.
- An internal planning memorandum for Microsoft's Java development tools
for 1998 reaffirmed the objective: "Kill cross-platform Java by grow the
polluted Java market." GX 259.
- A fall 1996 presentation on "Microsoft API Strategy" used by Muglia
summarized Microsoft's "Java Approach" as including the following plan:
"Use our Windows VM to own Java.* and Beans . . . . MS ability to lead
Java developers is largely driven / limited by IE's share . . . . How does
Microsoft win?" "Requires leadership in browser market share, developers
target Microsoft API extensions." GX 470, at MS6 5006862, MS6
5006877, MS6 5006887 (emphasis added); see also Muglia, 2/26/99pm, at
58:4-23.
- Microsoft's Todd Nielson summarized the strategic objective behind
J/Direct in an e-mail to Gates and Chase on August 25, 1997: "So we are
just proactively trying to put obstacles in Sun's path and get anyone that
wants to write in Java to use J/Direct and target Windows directly." GX
256.
- In an August 1996 e-mail concerning "Java APIs," Microsoft's John Ludwig
similarly emphasized: "Subversion has always been our best tactic. Don't fight
awt--enhance it and support it better than anyone else. Don't fight jbdc--adopt it
and move on. Subversion is almost invariably a better tactic than a frontal assault.
It leaves the competition confused. They don't know what to shoot at anymore."
GX 1327 (emphasis added).
- Ben Slivka and others continued to emphasize the same theme in an October 1996
presentation articulating Microsoft's "Java Strategy": "1) Drive MS Java VM and
classlibs (w/Win32 extensions!) to broad installed base . . . Do encourage
fragmentation of the Java classlib space: multiple 2D standards, AWT vs. IFC,
etc." GX 518.
- Steven McGeady testified that Microsoft's Java strategy was "part of the
embrace-extend model, which was, I believe, by this time they felt that it
was unlikely that they could keep Java from happening at all, but they
wanted to have it happen in a way that was incompatible. They wanted
Microsoft's Windows Java to be incompatible with Sun Java. That would
essentially have defeated the fact that someone could write a Java
application and have it run on a cross-platform basis. If Sun's Java had
one model and Microsoft's had a different model, then Java applications
wouldn't work everywhere, and they wanted that to--they wanted to keep
that from happening . . . . They wanted them--they wanted Java
applications written for Windows to just run on Windows, and Java
applications conversely written for non-Windows platforms to not run on
Windows." McGeady, 11/9/98pm, at 71:23 - 72:18; see also McGeady,
11/10/98am, at 10:19 - 11:7 ("And whether or not it had any benefits for
end users, it would have made the Java applications written for Windows
incompatible with other platforms.").
- Gosling recalled Microsoft executives agreeing in a February 1997
meeting that fragmenting Java would harm developers: "All of us who
attended the meeting -- including Microsoft -- unanimously agreed that
unilaterally extending the Java programming language would hurt
compatibility among Java tools and programs, would injure other tools
vendors, and would damage customers' ability to run a Java-based
software product on whatever platform they wished. One of the Microsoft
representatives in attendance admitted that unilateral language
extensions would be detrimental, and said that Microsoft 'wouldn't be
cowboys' by unilaterally introducing such extensions in their
implementation of the Java technology." Gosling Dir. ¶ 61.
- Slivka wrote in October 1996: "If Sun and we disagree on this and
diverge in more areas, as long as Netscape doesn't buddy up with Sun,
that is a super outcome for us (more fragmentation)." GX 1910. Slivka
explained in his deposition that the "this" in "if Sun and we disagree on
this and diverge in those areas" meant language extensions. He also
acknowledged that it would be a super outcome, A. "Because it would just
confuse Java developers about which Java platform they should write for.
Q. That was a good thing from Microsoft's perspective? A. Correct." Slivka
Dep., 1/13/99PM, at 753:17 - 754:7; See also Slivka Dep., 1/13/99pm, at
735:13-23 ("Q: And did you see the fragmentation of the Java class library
space as being useful because it was a means to the end of defeating
'Write Once, Run Anywhere'? A: All this comes back to Windows and the
threat, you know, Sun's very direct threat to our Windows platform, and
the success of Windows on the client. So, this seemed like if the library
space fragmented, the 'Write Once, Run Anywhere' . . . would be a lot
less probable.").
324.1. Microsoft's own, contemporaneous documents show that its
purpose was, not to increase the value of or demand for Windows, but rather to "kill
cross-platform Java."
- An internal planning memorandum for Microsoft's Java
development tools for 1998 reaffirmed the objective: "Kill cross-platform Java by grow the polluted Java market." GX 259
(emphasis added).
- In October 1996, Slivka and other executives articulated "Microsoft
Java Strategy" as including: "Let Java class library space fragment
so that 'write once, run anywhere' does not happen . . . . Don't
encourage new, cross-platform Java classes, especially don't help
get great Win32 implementations written/deployed." GX 518
(emphasis added).
- In May 1997, Gates sent an e-mail to John Ludwig and Ben Slivka
stating, "I think supporting JDK 1.1 is fine and I am hard core about
NOT supporting JDK 1.2. I really needed to understand where we
were going to draw the line because I am so afraid of the slippery
slope." Slivka responded: "JDK 1.2 has JFC, which we're going to
be pissing on at every opportunity. This summer we're going to
totally divorce Sun. Apple's announcement yesterday that they will
encourage Rhapsody developers to write in Java and call native
Rhapsody services was a mere shadow of what we're going to be
encouraging developers to do." GX 253.
- In January 1997, Gates sent an e-mail to Maritz and other top
Microsoft executives in which he commented on Java as follows:
"To avoid middleware taking over an operating system you have to
make sure the integrated services are different from the
middleware -- otherwise the middleware approach has no
disadvantages and it wins. I think the path we were going down of
building on AWT was a sure disaster -- it was creating a situation
where pure 100% Java applications would look just as good as
pure Windows applications which we have to avoid." GX 351.
- In April 1997, Jim Allchin wrote: "The comment about AFC
concerns me since I think you are saying that whatever
functionality is added by SUN you will add in a compatible way.
They have you on a treadmill. I don't understand how this is a
winning course. I would explicitly be different -- just to be different.
Further, I believe the highest priority is to create and expose NEW
Windows APIs (not clone) -- absolutely callable from java. I would
add more to Windows and less cross platform. This is clearly a
hard tradeoff that will be made each day, but I believe it is not a
question of how much resource is applied between Windows and
cross platform." He continued that "without something to pollute
Java more to Windows (show new cool features that are only in
Windows) we expose ourselves to more portable code on other
platforms -- especially if the Pure Java logo push wins in any way.
Either we need to pollute to Windows or we need more other
languages to offer viable alternatives to Java (or both). I would feel
so much better if I felt we weren't on such a middleware clone SUN
strategy. This is where" you and "I are apart." GX 1334, at MSS
0003551 (emphasis added).
- Fisher Dir. ¶¶ 111-113 (collecting quotes and documents); Fisher
1/7/99pm, at 50:11-19 (commenting on GX 518: "well, what
Microsoft did with Java, as I testified this afternoon, was essentially
to engage in a strategy which discouraged or subverted
cross-platform Java. And this is quite consistent with it. Don't
encourage cross-platform Java classes, and it would doubtless
help the Java if applications written in cross-platform Java had
great Win32 implementations. Win32, as I assume we all know, is
32-bit Windows, which refers to Windows 95 and 98 and Windows
NT.").
- See also GX 58 (Slivka April 1997 list of "issues/concerns" to Mr.
Gates, stating that "I want to make sure I understand your
issues/concerns. Here's a start, can you please add anything I'm
missing?" He questioned "2. How do we wrest control of Java
away from Sun? 3. How do we turn Java into just the latest, best
way to write Windows applications?").
324.2. If Microsoft had wanted to increase the value of Windows, it
would have supported JNI; supporting JNI would have reduced development costs for
Windows programs because it would have enabled developers to write simultaneously
for both Windows and other platforms.
- As Maritz described Intel's perception of Netscape in April 1996,
"they see Netscape as a 'friend of the PC'" (GX 288) suggesting
that ensuring compatibility with Navigator's Java implementation
would promote PC-based (and therefore Windows-based)
application development.
- Gosling testified that it was unnecessary to exclude JNI in order to
achieve the goal of providing developers the ability to develop
Windows-specific Java programs: "We believe that it is perfectly
fine for them to be allowed to make a choice between building a
native method in a portable way and building a native method in
some other way. In our case in San Jose, we never tried to
represent that we thought it was a bad idea for Microsoft to put
their own native method mechanism in there. What we argued was
that it was bad for them to exclude the portable native method
interface that was a part of the standard. And their argument had
been that it was difficult, but it has only taken them a few weeks to
actually release a new version of the tool. So it really wasn't all that
difficult for them to do." Gosling, 12/10/98am, at 54:8-20; see also
Gosling, 12/10/98am, at 33:2-14 ("Microsoft could easily have
accomplished its goals of building an access to the underlying
Windows platform, without violating the Java standards--without
violating the Java specification. In fact, as the system was initially
delivered to Microsoft, there was already a native method interface
. . . called 'NMI' which worked perfectly well. It had a number of
issues, as far as developers were concerned, with portability
amongst different platforms, but, nonetheless, it allowed people to
build native methods that would access whatever underlying
platform facilities people needed to expose. And that was all done
within the context of the Java standard.").
324.3. Indeed, no other JVM vendor or operating system vendor has
taken steps to support Microsoft's Windows-specific native code interfaces, and most
have supported JNI.
- Gosling testified: "So, for doing native method interfaces outside of
browsers, this has worked quite well, and JNI has been fairly
healthy. As far as I know, essentially all of our licensees have
implemented JNI, although they're not, by and large, browser
vendors." Gosling, 12/9/98pm, at 97:24 - 98:6; see also Gosling,
12/9/98pm, at 98:18-20 ("As I understand it, essentially all of our
licensees are in pretty good shape when it comes to native
methods.").
324.4. Microsoft's argument that it developed its own native code
interfaces and refused to support JNI because it was excluded from the Java standard-setting process is not consistent with the evidence.
324.4.1. Microsoft was never interested in promoting cross-platform Java.
- In February 1997, Ben Slivka and other Microsoft employees
discussed "potential issues for Java Language meeting."
Paul Gross counseled against such a meeting, noting, "I
believe that our true goal, controlling the future of Java, will
be totally transparent and mostly unacceptable to all Java
OEMs. I believe that we currently have the most control and
at some point will have total control." GX 1326.
- Gosling testified that Microsoft's communications concerning
native interfaces indicated "that Microsoft was completely
uninterested in pursuing a technical investigation of a portable
solution." Gosling, 12/9/98pm, at 57:12-25; see also Gosling,
12/9/98pm, at 67:1-8 (testifying that "when the time came to
actually form a working group, since Microsoft had already
said they weren't interested in doing anything that was a
cross-platform design, there was not a lot of point"); Gosling,
12/10/98pm, at 52:24 - 53:12 ("their version of friendship
was just take what we give you, and all of these things are
essentially poison pills that break cross-platform portability . .
. the proposals that they made were fairly consistently things
which tied people into the Windows platform in ways that
broke all of the interoperability goals.").
324.4.2. Mr. Muglia's testimony (Muglia Dir. ¶¶ 61-69, 108), that
Microsoft attempted unsuccessfully to work with Sun on native code interfaces and that
Sun's internal documents somehow suggest a plan unfairly to exclude Microsoft, rests
entirely on selected documents furnished by Microsoft counsel for the purposes of
litigation and on statements in these documents that Mr. Muglia read out of context to
imply the opposite of their intended meaning.
- Muglia acknowledged that his argument was based on a
selection of Sun documents provided to him by Microsoft's
attorneys, not any independent or comprehensive review.
Muglia, 2/26/99pm, at 90:3-19.
- Muglia relied in part on a series of September 1996 Sun e-mails titled "Java Interfaces and Beans" (DX 1918) to
support his suggestion that Sun simply "ignored" Microsoft's
input (Muglia Dir. ¶ 67). But the Sun e-mails do not suggest
any unfair exclusion. John Kannegaard initiated the
exchange by informing other Sun executives that "MS
grumpy that we aren't just going with their specs. We should
tell them why." In response, David Bowen noted that
although "we are not going to adopt the MS solution," Sun
engineers had been working with Microsoft and, based on
that work, "believe that it will not be hard for them to support
our new proposed API on top of their COM interface."
Similarly, David Connelly noted: "Consistently, Microsoft's
idea of working with us is to offer us some specs clearly
aimed at the Windows 32 platform, then balk when we don't
just accept them. This was clearly true for JAR and,
perhaps almost everything else on the list." DX 1918.
- Muglia testified that Sun executives had recognized and
expressed concerns about limitations on Sun's rights to
insist on Microsoft's compliance with its specifications under
its Technology License and Distribution Agreement (TLDA),
cited internal Sun e-mails (DX 1956) in support of this, and
suggested that top Sun executive Jon Kannegaard
"concurred that the TLDA posed problems for Sun and
solicited approaches to those problems." Muglia Dir.
¶ 68. Muglia ignored, however, that Kannegaard explained in no
uncertain terms in his January 1998 deposition in Sun
Microsystems, Inc. v. Microsoft Corp., that the document
contained a typographical error and that in fact "I did not reach
such dire conclusions" as those expressed in DX 1956.
Kannegaard Dep., 1/19/97, 184:1-2 (DX 1956A) (under seal).
- Muglia also relied on DX 2174, yet another series of internal Sun
e-mails, to support his contention that Sun deliberately excluded
Microsoft. Muglia Dir. ¶ 68. He ignored, however, that
Kannegaard specifically testified in his deposition that Sun
never undertook any of the actions discussed in DX 2174:
"This is really an amazing e-mail message, isn't it? We
didn't do that." See also Kannegaard Dep., 1/19/97, at
195:1-4 (DX 2174).
- Sun documents also contradict Muglia. See GX 1357
(Kannegaard e-mail messages to Microsoft, 10/2/96: "I
looked into complaints you've heard and relayed from
microsoft on how we ignore them, surprise them etc. In
particular, they complained about the JRI design, a design
that did not go their way. We conducted a fabulous process
and came to the right conclusion. They just don't like it."
"You can look at this yourself. If you do, you will find that the
Microsoft proposal was seriously considered but, in the end,
rejected for legitimate technical reasons. You will find that
IBM, Apple, and SMLI all reached this conclusion without
any help from us.").
324.5. Similarly, Microsoft's refusal to distribute RMI, as well as its
making RMI difficult to find on its website, made no business sense except as part of
Microsoft's campaign to kill cross-platform Java and maintain its monopoly power.
324.5.1. Microsoft's refusal to distribute RMI increased Microsoft's
costs.
- Gosling testified: "In packaging it separately, . . . they did
incur an extra cost for themselves because they had to
create the separate package rather than just delivering them
together. RMI is a very small fraction of the size of the base
tool, but what it has done for developers is it's made using
this cross-platform communication technology much more
awkward because then their customers have to go through
this other step of installing this patch to actually make it
work." He continued: "And what they effectively did was
create a separate product. A separate set of binaries that
you could download separately. And it would have been a
pretty straightforward thing, and from my point of view, a
very natural thing to have left the RMI classes with
everything else. It certainly wouldn't have made the product
any larger. It wouldn't have changed their testing problem at
all because they had to test the separate product in
combination with this, anyhow. It was simply a decision to
package it separately and to incur the costs of building two
packages instead of one." Gosling, 12/10/98pm, at 41:12 -
42:24.
324.5.2. Microsoft's contention that Microsoft "elected not to ship
RMI with our Java implementation because we think it is inferior technology" (Muglia
Dir. ¶ 113) is unsupported by any contemporaneous evidence.
- Gosling, 12/10/98pm, at 42:14-24 ("Q: What advantage, if
any, was there to Microsoft in doing that? A: It certainly
made developers' lives harder. Q: Any other advantage?
A: Well, they had a set of communication technologies,
themselves, which they have been trying to promote which
are, you know, not standard cross-platform communication
technologies. And they, in some ways, conflict with RMI, but
that's a business conflict. It's not a technical conflict. It's
perfectly possible to run DCOM and RMI together.").
325. Microsoft contends that its pollution and fragmentation of Java were
intended merely to provide developers the choice of writing superior Windows-specific
Java programs, without harming the prospects of cross-platform Java (Muglia,
2/26/99pm, at 27:23 - 28:24, 36:24 - 37:8; Schmalensee Dir. ¶ 149). But this assertion
is also inconsistent with the evidence.
325.1. First, Microsoft's professed commitment to improving the Java
programming environment is belied by contemporaneous documents showing that its
intent was to "subvert" and "kill" cross-platform Java.
- As its top executives discussed in contemporaneous e-mails,
Microsoft wanted its Java implementation to "be different just to be
different" (GX 1334, at MSS 0003551 (Allchin)) and to "subvert"
Java (GX 1327 (Ludwig)). See generally supra ¶ 330.
- As Gosling explained in response to the Court's questions about
Microsoft's suggestion that it has simply improved Java, that
Microsoft's "version of better is tied to the Windows platform and
preventing interoperability with other platforms." Gosling,
12/10/98pm, at 62:3 - 63:24.
- Similarly, Soyring testified that "I can only comment on what I read
in the press and reading Microsoft press releases, is that they do
talk frequently about their support of Java. They talk about having
the highest performance Java implementation, though it's a
Microsoft specific version of Java. They talk about providing Java
application development tools, which is the Microsoft Visual J++.
So, they use those very frequently as evidence that they are an
active proponent of Java. It just happens to be misleading, and
many of our customers recognize that because of the proprietary
extensions or deletions or omissions that are in their Java
implementation." Soyring, 11/18/98pm, at 55:17 - 56:8.
- A March 1998 unreleased draft of a Symantec public statement put
Microsoft's illusory creation of "choice" succintly: "By breaking away
from pure Java, there is a perception that MS is seeking to cloud
the issue; throwing up new whiz-bang technologies which have
little practical use in 90% of applications." GX 2078.
325.2. Second, Microsoft's contention that it was simply providing
developers and consumers options is belied by the numerous steps it took to deny
consumers Java implementations not controlled by Microsoft, even when those
implementations would run well on Windows.
- McGeady testified that "Microsoft was very upset that we were
working on Java at all, and in particular, they didn't want -- we had
a very highly optimized virtual machine that ran very fast on Intel
architecture, and they did not want that VM to become part of
Netscape Navigator." McGeady, 11/10/98am, at 6:9-16; see also
McGeady, 11/10/98am, at 10:19 - 11:7.
- Engstrom made clear to Intel that Microsoft would prefer it to stop
helping Sun on Java APIs, "especially ones that run well . . . on
Windows." GX 235.
- Microsoft's Todd Nielson -- the executive in charge of Microsoft's
Developer Relations group -- summarized Microsoft's strategic
objective in August 1997 as "just proactively trying to put obstacles
in Sun's path and get anyone that wants to write in Java to use
J/Direct and target Windows directly." GX 256.
326. The testimony of Microsoft witnesses concerning its conduct relating to
Java is not credible.
326.1. Mr. Gates' testimony is contradicted by the contemporaneous
documents.
- Compare Gates Dep. (played 12/2/98am), at 20:14-18 ("Q: Is one
of the things that you're focused on trying, in Mr. Slivka's words, to wrest
control or get control, if wrest is a word that you don't like, of Java away
from Sun? A: No."), with GX 265 (Gates, August 1997: "Do we
have a clear plan on what we want Apple to do to undermine
Sun?"); GX 897 (Gates, January 1998: "There is a big question of
what we should do with JAVA on Mac -- whether doing more work
and working closely with Apple could help us in this 'battle'. I think
it can and would hate for Apple to have to go back to the SUN
camp.").
- Compare Gates Dep. (played 12/2/98am), at 21:11-16 ("Q: It is
your testimony, then, sitting here, that Microsoft was not at all
motivated by a desire to limit the extent to which Java could be
used to develop applications programming that could be used on
platforms other than Microsoft's Windows? Is that your testimony?
A: Yes."), with GX 290 (Gates February 1997 message concerning
Microsoft support for AMD's 3DX work: "If Intel has a real problem
with us supporting this then they will have to stop supporting Java
Multimedia the way they are. I would gladly give up supporting this
if they would back off from their work on JAVA which is terrible for
Intel.").
- Compare Gates Dep. (played 12/2/98am), at 32:2 - 33:21
(testifying, regarding a document written by Slivka, that "I don't
know if he's referring to pissing on JFC or pissing on JDK 1.2, nor
do I specifically know what he specifically means by 'pissing on.'"
And testifying that "he might mean that we're going to be clear that
we're not involved with it, that we think there's is a better
approach"), with GX 253 (Gates' e-mail message to which Slivka
was responding: "I am hard core about NOT supporting JDK 1.2.").
326.2. Mr. Muglia's testimony that Microsoft wanted to support Sun's
version of Java and not fragment Java is contradicted by Microsoft's own documents
and is internally inconsistent.
- Muglia's effort to explain away Mr. Gates' statements that "I am
hard core about NOT supporting JDK 1.2" (GX 253) by testifying
that "exactly what Bill meant by the word 'support' isn't really clear"
is incredible. As the Court noted, "I don't think it could be read any
other way" than to make clear that Gates "doesn't like the idea of
supporting" JDK 1.2. Muglia, 2/26/99pm, at 43:23 - 48:10.
Muglia's subsequent effort to further explain that Microsoft drew a
distinction between "supporting" JDK 1.2 on its website, while not
distributing it with Windows (Muglia, 2/26/99pm, at 110:7 - 111:6),
implicitly conceded that GX 253 accurately stated Microsoft's policy
of seeking to impede cross-platform Java.
- Although Muglia suggested that "killing cross-platform Java," as
expressed in GX 259, was merely the erroneous impression of a
low-level employee and not Microsoft's strategy (Muglia,
2/26/99pm, at 99:9 - 100:3), that testimony is belied by numerous
internal documents and Muglia's own testimony. See, e.g., GX 351
(Gates: "I think the path we were going down of building on AWT
was a sure disaster -- it was creating a situation where pure 100%
Java applications would look just as good as pure Windows
applications which we have to avoid."); GX 518 (Slivka: "Let Java
class library space fragment, so that 'write once, run anywhere'
does not happen"); Muglia Dir. ¶ 72 ("By encouraging developers to
write Java programs that make native calls to Windows APIs,
Microsoft is growing the market for Windows-specific, or 'polluted'
Java. If most developers choose to write Windows-specific Java
programs, instead of slower, less functional cross-platform Java
programs, we will have won the competitive battle and 'killed' Sun's
strategy of using cross-platform Java to 'attack' Windows.").
- Muglia also sought to explain away GX 1334 (Allchin, "we should
be different - just to be different") by arguing that it concerned
Microsoft's Application Foundation Classes (AFC), which Microsoft
abortively sought to develop as a cross-platform alternative to
Java, ignoring that this was a distinction without any meaningful
difference: "Q: And do you understand that what he is saying is
that he is recommending to Mr. Slivka that he not try to proceed by
adding functionality in a compatible way, that he should be different
just to be different? A: No, not exactly. Again, you have to
understand the context of what this paragraph has in it. It says
AFC. This was technology Microsoft was building for our
customers, and I -- as I interpret this, Jim is saying, 'Don't focus on
just the things that Sun is doing. If you do that, they have us on a
treadmill. We need to do things that are different.' And he uses
the emphasis phrase 'just to be different.' And as I interpreted that
when I read it, it meant so that we can have differentiation versus
our competitor, Sun. In other words, not just providing the same
old thing that our competitor is providing. We need to be different.
THE COURT: Why is that different that what Mr. Boies just said?
A: Well, Mr. Boies wasn't referring -- I want to be clear, your
Honor, that what Mr. Allchin was saying when he said this was in
reference to one of Microsoft's technologies known as AFC."
Muglia, 2/26/99pm, at 33:24 - 35:3.
- See also Muglia, 2/26/99pm, at 36:24 - 39:21 ("Q: Now, my
question is regardless of where the fragmentation of Java came
from, did Microsoft believe that it was in Microsoft's interest to have
that fragmentation occur? A: Yes, in essence that's true. In this
case, remember I was saying fragmentation is really a word, and
particularly with Java, meaning choice, different choices for
developers. And it is in our interest to have more choices for
developers, yes. Q: Well, sir, are you saying that the reason that
you wanted fragmentation was to just give developers more
choice? Is that your testimony? A: No, not entirely. Q: Okay.
Part of the reason that you wanted to have fragmentation of the
Java platform is because that would discourage developers from
writing to the cross-platform Java technology; correct, sir? A:
Potentially, yes. Particularly if, in fact, the fragmentation is coming
from our competitors, say, Sun and Netscape, each promoting a
different variation of Java. If there were different implementations
out there, yes, that might discourage developers from choosing our
competitor's platform. Instead they might prefer to choose our
platform, Windows.").
3. Microsoft engaged in anticompetitive conduct to exclude
cross-platform Java
327. Microsoft not only developed and widely disseminated its polluted version
of Java, but also engaged in anticompetitive conduct to promote the use of its JVM and
to exclude cross-platform versions.
a. Microsoft used predatory means to weaken the principal
distribution vehicle for cross-platform Java, Netscape
328. Microsoft impeded cross-platform Java by severely weakening Netscape,
the principal distribution vehicle for cross-platform JVMs.
328.1. Microsoft's predatory campaign to weaken Netscape, and to gain
browser share, hindered the distribution and usage of non-Microsoft JVMs.
- See supra Part V.
- Gosling testified that, to the extent that Microsoft "did a fair amount of
economic damage to Netscape, I mean, by essentially setting the price to
zero, our biggest source of damage in Netscape as a distribution channel
was when they got to the point where they couldn't fund the engineering
efforts adequately to keep up with us." Gosling, 12/10/98pm, at 66:16-21;
see also Gosling, 12/3/98am, at 50-51 (although Netscape has decided to
stop shipping a VM with Communicator, this was a result of the fact that it
just couldn't keep up with the engineering work of keeping their VM up to
date, as a result of which it has decided to redesign the browser in a way
that will work with separately distributed third party VMs).
- Gosling further testified: "I understand that evidence may be offered in
this case that will show that Microsoft has taken actions that have limited
the distribution of Netscape's browser. To the extent that Microsoft has
taken such actions, those actions have made, and will continue to make, it
more difficult for the Java technology to achieve the broad distribution
necessary to ensure its success as a cross-platform technology. The
preservation of a competitive market for Internet browsers is critically
important to preserving the viability of the cross-platform Java
technology. Microsoft's actions injure not only Sun, but the entire
community of Java-based developers, computer users, and potential
operating system competitors of Microsoft. I believe that a competitive
browser market would significantly assist Sun and other cross-platform
Java-based developers in getting cross-platform Java technology
distributed to end users through the distribution of non-Microsoft Internet
browsers . . . . The ultimate success of the Java technology is dependent on
broad distribution of JVMs that comply with the specifications for the Java
technology. Without enough users able to run Java programs, there will
be little incentive for developers to write Java-based software; without
enough Java-based software, the Java technology will be unattractive to
users." Gosling Dir. ¶¶ 45, 64-65; see also Gosling, 12/3/98pm, at 32:10-20 ("We still believe that the browser is an incredibly important channel to
get compatible Java virtual machines out there. And just because
Netscape was unable to do the engineering to deal with the compatible
Java virtual machine, that left us in something of a pickle. So, we decided
that it was important because that channel was still very important, to try
to work with them, despite their difficulties, . . . on a way where they
could do less work and yet still have a . . . compatible Java virtual machine
integrated effectively in their browser."); Gosling, 12/10/98pm, at 32:18-33:7 (testifying that, because of the difficulty of downloading JVMs,
impeding Netscape "pretty severely restricted the channel for us").
- Gosling testified that "the overwhelming majority of JVMs that have
been distributed to end users to date were distributed with Internet
browsers, specifically Netscape's Navigator/ Communicator, and
Microsoft's Internet Explorer. No other distribution mechanism for
JVMs has reached even a small percentage of the distribution
achieved through the distribution of Internet browsers." Gosling
Dir. ¶ 42.
- See also Gosling, 12/10/98pm, at 31:14-23 (Navigator provided a means
to get the Java runtime environment out to consumers, "so that the Java
APIs are now available everywhere that Navigator is, and so people
writing Java applications can write them and run them on those
platforms"); Gosling, 12/10/98am, at 68:25 - 69:3; Gosling, 12/10/98am,
at 71:25 - 72:5; Gosling, 12/10/98am, at 77:15-17 (the web "as a channel,
is just absolutely drowned by the fact that the Microsoft VM is bundled in
the vast majority of all computers built today."); GX 254 (June 1997 e-mail from Gates to Maritz).
- Professor Franklin Fisher testified: "Microsoft, recognizing that
Netscape's browser was the primary distribution method for Java,
sought to eliminate Java by eliminating Netscape's browser as a
viable alternative." Fisher Dir. ¶207 (collecting Microsoft documents).
328.2. Microsoft intended its predatory campaign against Netscape and
other browsers to hinder cross-platform Java.
- A January 1996 Microsoft presentation made clear that the
"integration" of Internet Explorer with Windows was a crucial
element of a strategy for responding to Java: "Response
Summary: Increase IE share, Integrate with Windows." GX 52, at
MS7 003270.
- Robert Muglia agreed that "Increasing Internet Explorer's share is a
response both to Netscape Navigator and to Sun's cross-platform
threat" "in the sense that the Internet Explorer feature of Windows
has APIs in it that are competitive with Java," meaning Microsoft's
JVM. Muglia, 2/26/99pm, at 53:1 - 54:7.
- As Microsoft's John Ludwig pointed out in a September 1996 e-mail concerning Microsoft's relationship with Sun: "The only thing
that will bring them to the table is our continued success at growing
IE share, and growing our VM share thru IE and Windows
distribution, and leveraging that to create a successful position in
Java class libraries." GX 2519.
- An August 1997 memorandum from IBM's John Thompson to Sun's Scott
McNealy, Netscape's James Barksdale, and others, emphasized that
widespread distribution and usage of Internet Explorer was of serious
concern to the future of an "open" Java platform: "We must clearly
explain to our customers that Internet Explorer is Microsoft's primary
weapon to kill 100% Pure Java. Customers who want an open Java
platform must vote with their browsers!" DX 1894, at IBM 60968.
- As an executive with the investment capital firm Kleiner, Perkins,
Byers & Caulfield informed Netscape after a meeting with
Microsoft: "M/S supports Java the programming language, but is
directly opposed to Java the 'OS'. . . . Shipping IE 3.0 with every
Win95 desktop from Q4 1996 on will drive leadership for IE and
ActiveX against Navigator and Java." GX 1233.
- See also Soyring Dir. ¶ 28 (testifying that an "additional effect of
Microsoft's including Internet Explorer with Windows is the advantage it
gives Microsoft's implementation of Java technology"); Fisher Dir. ¶ 209
("Microsoft did not seek to 'kill cross-platform Java' merely by developing
its own version of Java and marketing it on the merits in competition with
cross-platform Java. Instead, Microsoft used its monopoly powers over
PC operating systems to induce and require industry participants to accept
J/Direct (and IE which included J/Direct) instead of Java (and Netscape's
browser which included Java).").
b. Microsoft used its monopoly power to force widespread
distribution and usage of its Windows-specific version
of Java
329. Microsoft did not simply use anticompetitive means to hinder cross-platform
Java's distribution; Microsoft also used its operating system monopoly to coerce firms to
distribute and support its Windows-specific Java environment.
(1) Microsoft entered into exclusionary "First Wave"
agreements with ISVs
330. Microsoft entered into dozens of exclusionary "First Wave" agreements
with ISVs with the purpose and effect of hindering cross-platform Java. These
agreements conditioned early access to information about Windows 98 and Windows
NT -- information of vital competitive significance to ISVs -- on ISVs' agreement to make
Microsoft's Windows-specific JVM the default JVM for any of the ISVs' Java programs.
-
- redacted -
GX 970, at RSC 618 (Rational Software --
DevlinF); GX 2071, at DOJ0000003 (sealed) (Symantec -- Eubanks).
- redacted - GX 2400-2497 (all sealed) (additional
First Wave agreements).
- Gordan Eubanks, moreover, acknowledged that "early access to
select specifications and beta releases of Windows NT 5 and SDKs
on an ongoing basis via CDs and secure FTP site," both of which
are conditioned in the First Wave agreement on Symantec's
agreement to make Microsoft's JVM its default JVM, is both
"important" and "necessary" to Symantec. Eubanks, 6/16/99pm, at
16:4- 17:19.
- Soyring testified to the nature and effect of Microsoft's agreements: "A
second aspect is from a commercial software developer program,
commercial software vendors who want to develop and sell an
application that carries the logo or the mark design for Microsoft
Windows 95 and Windows NT. In that contract, I understand, is
the provision that requires that the vendor, if they use Java as a
language, write it to the Microsoft Java and ship the Microsoft Java
virtual machine with their application. That poses, then, a fear
amongst our customers that that Java application may only run on
these Microsoft platforms and, thus, not be cross-platform capable
as advertised by the other Java components." Soyring,
11/18/98pm, at 53:21 - 54:7.
330.1. Microsoft was enjoined by a Federal court from enforcing those
agreements in 1998.
- Sun Microsystems, Inc. v. Microsoft Corp., 21 F. Supp. 2d 1109, 1127-28
(N.D. Cal. 1998) (enjoining Microsoft from conditioning "the right to
use the 'Designed for Windows95(98)/NT' logo on the exclusive
distribution of Microsoft's Java virtual machine").
330.2. Neither any witness nor any contemporaneous evidence describes
any legitimate business purpose for the restrictions included in the First Wave
agreements
(2) Microsoft misled developers into using
proprietary extensions to develop Windows-specific programs
331. Microsoft misled developers so that they unwittingly wrote Java programs
that turned out to be Windows-specific, even when the developer intended to create a
cross-platform program.
331.1. Until ordered to do so by a federal court in 1998, Microsoft did not
warn developers that using its extensions would bind their Java programs to Windows.
- Gosling testified that "it's my understanding that there is absolutely
no warnings in it about, you know, the fact that you're generating
platform-dependent code, and I don't even know a way to turn it off
using that tool." Gosling, 12/10/98pm, at 49:12-15; see also
Gosling, 12/10/98am, at 62-63 ("if your sole source of information
was the Microsoft product, you would not understand any of this"
because the disablers are buried several menus deep); Gosling,
12/10/98pm, at 48:22 - 49:4 (describing VJ6 "I'm not aware of any
warnings that appear in the documentation. When you actually run
the application, if you are sufficiently clever, you can actually find
some switches you can pull which say do or do not use the
Microsoft extensions. But even those don't explicitly say, you
know, . . . 'if you don't do this, you won't be portable.'").
- As Professor Fisher explained: "The problem is, if you do this in
Visual J++, at least until quite recently, you would not be warned of
a rather salient fact, which is, if you do that, you would be writing in
a way that would make . . . the program you were developing
runnable okay on Microsoft's JVM for Windows, but it would not
permit you to port the program easily to other . . . platforms."
Fisher, 1/7/99pm, at 28:20 - 29:5.
- Sun Microsystems, Inc. v. Microsoft Corp., 21 F. Supp. 2d 1109,
1127-28 (N.D. Cal. 1998) (enjoining Microsoft from: "Selling or
distributing, directly or indirectly, any software development tool or
product containing or implementing computer program code copied
or derived from any Sun copyrighted program code for the Java
Technology; as that term is defined in the TLDA, including SDKJ
2.0, SDKJ 3.0 and VJ 6.0, ninety (90) days after the date of this
order unless such product: . . . (2) has the default mode in the
compiler configured such that (a) Microsoft's keyword extensions
and compiler directives are disabled and (b) has the compiler mode
switch such that it enables, rather than disables, such keyword
extensions and compiler directives, and (3) includes a warning
which appears when a user elects to use the extended mode of the
compiler (either when the user accesses the compiler from a DOS
command line or when the user checks a box provided during
execution of the compiler software) and which warns the user
(a) that use of Microsoft's language extensions will result in
compiled code which may not run on all compatible virtual
machines, and (b) that future versions of Microsoft's development
tools may be prohibited by court order from incorporating keyword
extensions and compiler directives not contained in Sun's Java
Language Specification; however, nothing in this order prevents
Microsoft from removing the mode switch, keyword extensions and
compiler directives from its Java software development tools and
distributing or selling such resulting implementations.").
331.2. The absence of warnings about the impact of utilizing Microsoft's
extensions created a significant potential for developer confusion and the creation by
them of Java programs that were not cross-platform, but rather Windows-specific.
- Gosling testified: "You start with they're essentially a monopoly in
the operating system, and there is the fact that they bundle their
virtual machine . . . with the operating system, so that the operating
system rides along in that channel to this incredible volume of
customers. And then you take the fact they have made changes to
the Java language . . . . So, if you want to target the Microsoft VM,
which tags along on this incredibly powerful distribution channel,
you pretty much have to use Microsoft's tools." Gosling,
12/10/98pm, at 44:19 - 45:18; see also id. at 43:20 - 44:9.
- As Gosling testified: "I certainly had developers talk to me who said that
they had been developing with Visual J++ and then were surprised to
discover that the software that they had developed was not portable. And
besides this, the issue wasn't so much . . . the code that they themselves
directly wrote. But a lot of it had to do with what libraries did they use,
because if they used a library that was tainted with Microsoft-specific
extensions, then their applications wouldn't run. And while the developer
might have been aware when they wrote their own code whether or not
they were using these things--they weren't necessarily aware, because a lot
of the stuff was automatically generated by the tool, but whenever they
incorporated libraries from other places, you know, they might have been
sort of acquiring this problem, and it wasn't at all obvious when this
problem would show up. And one of the unfortunate aspects of these
extensions is that they are kind of an all-or-nothing thing. Namely, if
there is just one place in the application that has got one of these
Microsoft extensions, then that totally blocks portability." Gosling,
12/10/98am, at 58:15 - 59:11; see also Gosling, 12/10/98pm, at
39:1-16 ("There were, in general, no warnings that said, 'Oh, you're
about to be nonportable.' There were some knobs that you could
twist that were buried fairly deeply. And so, the natural thing . . .
would be if you use their tools, boom, you are locked into their
virtual machine.").
- An internal Boeing Corporation presentation in March 1998
described the likelihood of confusion as follows: "Microsoft
continues to create Windows-specific Java implementations.
Windows Foundation Classes (WFC) push the Microsoft Java
Virtual Machine (JVM) into a 32-bit Windows-only environment.
Developers using Microsoft programming tools may have difficulty
knowing when they are creating Microsoft-JVM-only compatible
Java programs." GX 635, at TBC 000428.
- In a November 1996 e-mail discussion with Ben Slivka and others,
Microsoft's Thomas Reardon suggested that promoting confusion
was intentional: "As i told charlesf on the phone, at this point its not
good to create MORE noise around our win32 java classes, instead
we should just quietly grow j++ share and assume that people will
take more advantage of our classes without ever realizing they are
building win32-only java apps." GX 1332.
- Similarly, a January 1997 article in CommunicationsWeek included
in internal Microsoft e-mails to top executives reported the
comment of a senior systems engineer at GTE Data Services as
follows: "Management's fear, according to Hagerty, is the enticing
power of the Microsoft environment. 'When developers get close to
a deadline, there may be an irresistable temptation to take
shortcuts . . . that will end up in the appliaction becoming
Windows-specific.'" GX 519, at MSS 0080019.
c. Microsoft induced third parties not to support cross-platform Java
332. Microsoft used a number of strategems to induce third parties not to
support, or to limit support for, cross-platform Java.
(1) Microsoft pressured Intel not to support cross-platform Java
333. Microsoft used a variety of means, including both threats to withhold
Windows operating system support from Intel's microprocessors and offers to include
Intel technology in the Windows operating system, to pressure Intel not to support
cross-platform Java development, even when that development would have improved
the performance of Java programs on Windows.
- McGeady testified that Microsoft pressured Intel not to support Java
"repeatedly and on multiple occasions." McGeady, 11/10/98am, at 14:19-21.
- Intel's Russell Barck also testified that "Microsoft believed JAVA was the
enemy, from their perspective, and they felt we should consider it in the
same light." Barck Dep., 8/25/98, at 45:16 - 46:8 (DX 2556); see also
Barck Dep., 8/25/98, at 47:12-24 (DX 2556) ("Q: Were there statements
made at those meetings at which you were in attend, that would give you
an indication why Microsoft thought that Intel should not support JAVA? . .
. A: I believe the primary reason was JAVA ran on multiple CPUs, or
could run on multiple CPUs, and . . . it was an open-architecture platform.
Q: If a program could run on multiple CPUs, is that something that would
be detrimental to Intel? A: Not necessarily.").
- In March 1997, Jim Allchin, responding to an e-mail from Gates, wrote "I
am positive that we must do a direct attack on SUN (and probably
Oracle). I think we have to do this because we are just letting them grow
their war chest. Between ourselves and our partners, we can significantly
hurt their (certainly SUN's) revenue base. Yet, we are very silent. We are
not getting the visibility that is required. I have talked to Compaq and HP
about SUN and I know we can leverage them much more than we are
today. We need to get Intel to help us. Today, they are not." Allchin
repeated this statement at the end of the e-mail: "Finally, as I mentioned
above, I think we HAVE to attack SUN (and Oracle). We need to do this
at a company level. I just saw a billboard! from SUN in downtown Seattle
over the weekend. SUN has a huge presence that we must stop." GX
356.
333.1. At a meeting between top Microsoft and Intel executives on
August 2, 1995, before Microsoft had settled on its Java strategy, Bill Gates threatened
to withhold Microsoft's support for Intel's next generations of microprocessors if Intel did
not cooperate with Microsoft on platform and Internet issues.
- Gates made clear at the August 2, 1995 meeting with Intel that
"Microsoft would not support" Intel's "next processor offerings if we
did not get alignment between Intel and Microsoft on platform
issues" and communications issues, i.e., Internet issues.
McGeady, 11/9/98pm, at 14:14 - 15:4. Intel's Ronald Whittier's
minutes of the meeting similarly report: "Gates would not agree to
let processors/OS's programs to progress unencumbered by
platform/communications program issues . . . MS: Very sensitive to
what Intel might do on the client side. Example: JAVA, a show
stopper." GX 279, at MS CID 00077-78.
- McGeady, described by Whittier as the executive responsible for
Java efforts at Intel in 1995, testified that Gates made clear that if
Intel "started supporting Java, that would be a, quote, show stopper
in the relationship. That would be another one of these things that
would cause the show to stop, meaning the cooperation,
presumably the cooperation, between Intel and Microsoft that we
needed" in "support of our microprocessors, both the MMX
microprocessor extensions as well as the P7 and other
microprocessors." McGeady, 11/9/98pm, at 67:14 - 68:7.
- See also Maritz, 1/27/99am, at 27:22 - 29:6 (agreeing that
Microsoft tried to stop Intel "from putting" its "resources into
supporting work by either Netscape or Sun" and that Java was a
"show stopper").
333.2. In the fall of 1995, after Microsoft had settled on the strategy to
utilize Java as a Microsoft-specific technology, Microsoft enlisted Intel's help in
preventing the "Java component model" from being established and convinced Intel not
to disclose its Java work publicly at that time.
- In a meeting on November 9, 1995, Paul Maritz told Intel
executives, including McGeady, that Intel and Microsoft should
cooperate to keep a "Java component model" from being
established. GX 564, at 477MSCID00276 (McGeady notes); GX
285 (Landsman's notes) (Microsoft's concern is not the "Java
language, per se," which Microsoft will support as "just another
language" but the "Java component object model.").
- McGeady explained that Maritz's request was "a part of the
embrace-extend model, which was, I believe, by this time they felt
that it was unlikely that they could keep Java from happening at all,
but they wanted to have it happen in a way that was incompatible.
They wanted Microsoft's Windows Java to be incompatible with
Sun Java. That would essentially have defeated the
purpose--defeated the fact that someone could write a Java
application and have it run on a cross-platform basis. If Sun's Java
had one model and Microsoft's had a different model, then Java
applications wouldn't work everywhere, and they wanted that to --
they wanted to keep" Java applications from "running anywhere in
a compatible way. They wanted them -- they wanted Java
applications written for Windows to just run on Windows, and Java
applications conversely written for non-Windows platforms to not
run on Windows." McGeady, 11/9/98pm, at 69:22 - 72:18.
- Frank Gill, McGeady's superior at that time in 1995, recorded
Microsoft's request, as well as Intel's response -- an agreement to
"not publicly endorse Java" -- in an e-mail written the day of the
November 9, 1995 meeting: "Re: Java, we told them that we felt
Java is on the way to becoming an Internet standard and felt a
need to optimize around" Intel architecture "and particularly
optimize our media components for the Java environment. They
see this as supporting their mortal enemy . . . and argue SUN is our
enemy as well. Paul describe to me as 'this is equivalent to us
optimizing for powerpc'" a competitive chip. I "left it with him that
have not starting anything until we talked today, we will not publicly
endorse Java, and I needed internal review before authorizing this
work to commence. So, bottom line, this is BIG DEAL to them."
GX 284.
- McGeady, 11/12/98am, at 69:14-17 ("Well, yes, actually. We had
hoped to support Java in a more public fashion . . . . MS was able
to convince Intel's management to withhold its public support for
the program."); see also McGeady, 11/9/98pm, at 73:2 - 74:22
(Microsoft succeeded in having Intel withhold public support for
Java).
333.3. Throughout 1996 and 1997, Microsoft continued to push Intel into
supporting its Microsoft-specific implementation of Java and away from compatibility
with the Sun model.
- In April 1996, Intel's Craig Kinnie reported that John Ludwig of
Microsoft "also wanted to convince me that our media class library
work was aiding the competition to the Win/IA platform and they
considered our work as competitive." GX 567.
- McGeady explained this comment: "Q: Did you have an
understanding in April of 1996 as to what that was referring to? A.
Yes. They were very upset that we were taking our optimized
audio, video, 3-d graphics software and adapting it to work within
Sun's Java framework and environment, and they wanted us to
stop. They considered it competition." McGeady, 11/10/98am, at
13:18-23; see also McGeady, 11/10/98, at 8:3-20 (Microsoft
strongly pushing toward the Microsoft model and "away from our
stated architecture and strategy, which was to adhere to the Java
standards and be compatible with Sun").
333.4. Microsoft also pressured Intel not to assist Netscape with Intel's
Java virtual machine, even though Netscape Navigator was still a leading Windows
application and Intel's work would have improved the performance of Navigator on the
Windows operating system.
- In an April 1996 memorandum, Intel's Alan Holzman summarized
requests made by Microsoft's Bob Muglia and John Ludwig:
"Muglia/Ludwig want exclusive access to Intel Architecture's VM
work: Ludwig does not want us to give Netscape our Java virtual
machine work. Netscape access to VM is VERY touchy with MS!!!"
GX 566.
- McGeady explained that Holzman's comments reflected that
"Microsoft was very upset that we were working on Java at all, and
in particular, they didn't want -- we had a very highly optimized
virtual machine that ran very fast on Intel architecture, and they did
not want that VM to become part of Netscape Navigator."
McGeady, 11/10/98am, at 6:9-16.
333.5. In June 1996, Bill Gates convinced Intel not to "wrap" Windows
APIs as Java APIs, i.e., not write Java APIs that would work well with Windows, even
though doing so would improve the performance of Java programs on Windows.
- Eric Engstrom admitted, with regard to Microsoft's efforts (GX 235)
to dissuade Intel from "helping Sun create Java multimedia APIs,"
that Microsoft's conduct had an impact on Intel's development
work: "Q: Well, in fact, you were successful for some period of time
in convincing Intel to stop helping Sun create Java multimedia
APIs, were you not? A: For some period of time we succeeded in
working roughly along the lines of this e-mail, yes." Engstrom,
2/23/99pm, at 33:19-23.
- In a meeting in June 1996, Gates told Intel CEO Andy Grove that
Intel should not "take anything resembling a Windows API and
wrap it as a Java API." GX 289 (6/6/96 Gates e-mail to Maritz).
Grove acquiesced. GX 289.
- As McGeady testified, the sole purpose for not wrapping Windows
APIs as Java APIs was to harm the prospects of cross-platform
Java. He explained that "the bottom line was we were -- we felt we
were taking our multimedia technology, interfacing with the
underlying basic capabilities of windows and providing that
multimedia technology as a integral part of Sun's Java. That, I
interpret, is what Bill was complaining about." McGeady,
11/10/98am, at 15:21 - 16:12.
- McGeady further testified that wrapping Windows APIs as Java
APIs would benefit users and developers because: "You would
have a Java implementation that was both compatible with and
would be able to run the same applications on a macintosh and on
a windows machine, and you would have a java implementation
that would run faster on a windows machine than one that was
implemented in a way that wasn't cognizant of the underlying
Windows APIs. So it would be faster and it would be more
compatible." McGeady, 11/10/98am, at 15:21 - 17:3.
333.6. Microsoft continued its campaign to pressure Intel not to support
cross-platform Java in 1997.
333.6.1. In February 1997, Bill Gates suggested to Intel CEO Andy
Grove that, if Intel wanted Microsoft not to support a competitive chip, Intel would have
to stop assisting Sun with Java multimedia.
- In February 1997, Microsoft was asked by AMD, an Intel
competitor, to support an AMD technology called "3DX."
When asked by Jim Allchin if Microsoft should support the
technology although Intel would not be in favor of such
support, Gates responded that, "If Intel has a real problem
with us supporting this then they will have to stop supporting
Java Multimedia the way they are. I would gladly give up
supporting this if they would back off from their work on
JAVA which is terrible for Intel. I have a call with Andy on
this topic coming up on Monday." GX 290.
- In February 1997, Gates wrote to Paul Maritz about "Intel
and Java." "I called Andy today to talk to him about our
sensitivities about JAVA and asking what goal Intel has in
working on JAVA. I told him we thought they were hurting
themselves but [sic] doing any development other than just
pure tuning things for Intel." GX 935.
333.6.2. In the spring of 1997, Microsoft convinced Intel to stop
helping Sun develop Java multimedia APIs, "especially those that run well on
Windows."
- In a May 1997 status update on Microsoft plans with Intel,
Microsoft's Eric Engstrom identified a primary Microsoft
objective for Intel as: "Intel to stop helping Sun create Java
Multimedia APIs, especially ones that run well (ie native
implementations) on Windows." GX 235. To achieve this
goal, Microsoft proposed to Intel that it would include Intel
technology in Windows in return for Intel's agreement to
stopping helping Javasoft write multimedia APIs. GX 235
("Intel agrees to stop helping Javasoft in any area where
Microsoft agrees to ship Intel technology as part of DirectX
media," and "augment DirectX rather than compete with it . .
. . We want them pushing DirectX rather than some bits of
Javasoft and Intel technology.").
- Engstrom testified about his May 1997 meeting with Intel
concerning Java multimedia APIs as follows: "Q: Let me be
sure I understand your testimony. Your goal was to try to
convince Intel to work with Microsoft's multimedia and not
with Sun's competing Java multimedia; correct, sir? A: My
goal was to get them to work with us. You know, I wanted to
be the best partner I could be for Intel. Q: And, in fact, sir,
you wanted more than that, didn't you? Didn't you also want
Intel to stop helping Sun create Java multimedia APIs? A:
Yes, that would be a goal of being a great partner for them,
is that, you know, they have no bandwidth for other partners.
. . . Q: Let me just be sure I understand your testimony
before we move on. In this particular case, it was
Microsoft's goal absolutely to stop Intel from working with the
competitor, someone who created competing APIs to
Windows, even if those APIs ran well on Windows; is that
what you said? A: It was my goal to get Sun -- to get Intel to
work with Microsoft for their multimedia investments, not with
Sun. Absolutely." Engstrom, 2/23/99pm, at 21:14 - 22:10,
35:2-23.
- Although Engstrom denied that he reached agreement with
Intel, he admitted that Microsoft was successful in stopping
Intel from helping Javasoft with multimedia Java APIs for
some period of time. Engstrom, 2/23/99pm, at 33:11-23.
334. Microsoft's efforts to prevent Intel from supporting cross-platform Java
made no business sense except to preserve Microsoft's monopoly power and
significantly impeded Intel's support and development efforts for cross-platform Java.
- Bob Muglia conceded that it is not in Microsoft's interests to have
developers write to "Sun's Java platform," even if the resulting programs
work well on Windows. Muglia, 2/26/99pm, at 79:8 - 80:15.
- As Steven McGeady put it, based on his meetings with Microsoft
executives, "they wanted Java applications written for Windows to just run
on Windows, and Java applications conversely written for non-Windows
platforms to not run on Windows." McGeady, 11/9/98pm, at 69:22 -
72:18.
- As Professor Franklin Fisher summarized, Microsoft sacrificed its interest
in improving Windows in pressuring Intel not to support cross-platform
Java efforts: "The more things that will run well on Windows, presumably,
the better is Windows functionality, . . . the more attractive Windows will
be to users. Here is something that suggests that Microsoft was
particularly not interested in having something that would run well on
Windows." Fisher, 1/7/99pm, at 52:19-24.
335. In light of Mr. Gates' contemporaneous writings and the other evidence, his
denials during his deposition of any effort to keep Intel from supporting Java are
incredible.
- Compare Gates Dep. (played 11/9/98am), at 62:21- 64:23 ("Q. Did
Microsoft make any effort to convince Intel not to help Sun and Java. A.
Not that I know of."), with GXs 279, 289, 290.
(2) Microsoft pressured Apple and IBM not to support
cross-platform Java
336. Microsoft sought to use its leverage as Apple's primary application supplier,
and the leverage that stems from the need for QuickTime to be technically compatible
with Windows, to secure Apple's support of Microsoft's Java strategy.
- Microsoft linked its cooperation with respect to continuing the
development of Microsoft Office for the Macintosh operating system to
Apple reducing its support of cross-platform Java. As Microsoft executive
Don Bradford wrote on January 21, 1998 to Bill Gates and others: "Mac
Office is the biggest Apple carrot. From a browser, mail client and Java
perspective, Apple has other options and will try to sit on the fence.
Negotiating from a single point, centered around Mac Office, will give us
better leverage." GX 267.
- As Bradford wrote on February 13, 1998 to Gates and Maritz, as well as
other Microsoft executives, with respect to Microsoft's August 1997
agreement with Apple: "We get our APIs deployed in perhaps the only
VM on Mac, reinforcing Java-is-a-language-not-a-platform, while
minimizing our investment." GX 268.
- On April 28, 1998, Gates wrote his executives, and again emphasized that
Microsoft should use its cooperation with Apple on its QuickTime software
as an inducement to get Apple to support Microsoft's Java strategy: "I
would love to see enough convergence of QT and our equivalent so that .
. . we use this 'convergence' to get more JAVA/IE affinity from Apple." GX
270.
337. Similarly, Microsoft sought to use its operating system monopoly to ensure
that IBM would not support cross-platform Java development.
- In October 1997, Gates wrote to Joachim Kempin "that the Java religion coming
out of the software group is a big problem . . . . Overall, we will never have the
same relationship with IBM that we have with Compaq, Dell, and even HP
because of their software ambitions. I could deal with this just fine if they weren't
such rabid Java backers. We are told they have funded Netscape to do the
browser without email and they are proposing to fund Netscape to give their
browser away. They have also funded a lot of SUN stuff . . . . I am willing to
take some risk in improving the relationship and think you should approach them
on steps for improvement. You should not position it as 'overnight we will treat
you like Compaq.' We should position it as let's do some things that are good for
both of us but which require some of the rhetoric to be lowered on both sides. On
their side I mean Java and NC." GX 257.
- As Garry Norris of IBM testified, that Bill Gates' statements were "very
consistent" with the type of treatment IBM received from Microsoft in the
1996 time frame, specifically, pressure not to offer software products that
competed with Microsoft's products. Norris, 6/8/99am, at 59:4 - 60:4.
- See also Gates Dep. (played 11/17/98pm), at 44:15 - 46:25 (Q: Mr. Gates,
let me put it this way. In October of 1997, were you trying to get IBM to
reduce its public support for Java? A: I say in here that under some
circumstance the rhetoric should be lowered on both sides and that I think
that's -- you know, that makes sense in certain circumstances . . . . Q:
And by `rhetoric,' you're talking about public rhetoric, were you not? A:
Definitely public rhetoric.").
338. Microsoft's pressure on Apple and IBM is unjustified by any legitimate
business interest.
338.1. The contemporaneous evidence demonstrates that Microsoft's
goal was to undermine cross-platform Java, not to enhance Windows (except by
defeating the cross-platform threat).
- On August 8, 1997, Apple's Avadis Tevanian sent an e-mail to
Microsoft referencing an earlier Microsoft/Apple agreement, "Java
collaboration," and requesting that Microsoft stop "trying to lock-out
QuickTime from Windows." Gates forwarded the message to Paul
Maritz and other Microsoft executives with the following comment:
"I want to get as a much mileage as possible out of our browser
and JAVA relationship here. In other words a real advantage
against SUN and Netscape. Who should Avie be working with? Do
we have a clear plan on what we want Apple to do to undermine
Sun?" GX 265.
- In an August 21, 1997 e-mail reporting on "conversations with billg
last nite," Microsoft's John Ludwig made clear what Microsoft's plan
was: "On the Java work, Bill says the ball is in our court to come up
with a win/win solution. He suspects that we should be cooperating
only at the lowest level -- native code interfaces, jits, etc., not at the
class lib level. Bill was clear that his whole goal here is to keep
Apple and Sun split. He doesn't care that much about being aligned
with Apple, he just wants them split from other potential allies." GX
255.
- As Gates put it in January 1998 : "There is a big question of what
we should do with JAVA on Mac -- whether doing more work and
working closely with Apple could help us in this 'battle'. I think it can
and would hate for Apple to have to go back to the SUN camp." GX
897.
338.2. In light of the contemporaneous documentation of Microsoft's
efforts to "keep Apple and Sun split," Gates' denial in his deposition (Gates Dep.
(played 11/2/98pm), at 14:14 - 17:6; 18:18-24) of any involvement in any plan by
Microsoft to encourage Apple to undermine Sun's Java is not credible.
(3) Microsoft entered into agreements with ISVs that
limited their ability to support cross-platform Java
339. In addition to its First Wave agreements, Microsoft entered into other
agreements to restrict distribution of non-Microsoft JVMs and native code interfaces by
ISVs.
339.1. Microsoft entered into agreements with third parties to support
exclusively Microsoft's JVM and native code interfaces. Those agreements were
enjoined in November 1998.
- For example, an unsigned License Agreement between MS and
Fujitsu for redistribution of MS Java software development kit
includes the following obligations on Fujitsu: "3. Obligations.
During the term of the Agreement, Company shall:
- redistribute
the Microsoft virtual machine for Java included in the Licensed
Software (the 'MS Java VM') as part of Company Product and not
any other virtaul machine;
- use only the Microsoft native code
interfaces (J/Direct, RNI, Java/COM) that are part of the MS Java
VM for any native code calling;
- support Internet Information
Server for server-side functionality;
- make reasonable efforts,
where appropriate, to redistrbution the Microsoft Visual J++ trial
edition with Company products."
GX 1331, at MSS 0165897-98;
see also Muglia, 2/26/99pm, at 59:19 - 62:12 (Microsoft and
AimTech agreement).
- Sun Microsystems, Inc. v. Microsoft Corp., 21 F.Supp. 2d 1109,
1127-28 (N.D. Cal. 1998) ("Since the court finds that Sun is likely to
prevail on the merits and that it may suffer irreparable harm if
Microsoft is not enjoined, a preliminary injunction is hereby issued
against Microsoft, and its officers, agents, servants, employees,
attorneys, and those in active concert or participation with them
who receive actual notice of this order by personal service or
otherwise, pending trial, from: . . . (E) Conditioning any license to
any Microsoft product on exclusive use or distribution of Microsoft's
Java virtual machine; (F) Entering into any agreement, condition or
arrangement with any third party that requires such third party to
exclusively use Microsoft's interfaces to its runtime interpreter when
invoking native code.").
339.2. Microsoft also restricted RealNetworks, the leading vendor of
Internet streaming software, from supporting Netscape and Java technology.
- The July 18, 1997, agreement between Microsoft and
RealNetworks requires, among other things,
- redacted -
GX 1369, at MS8 000635, MS8 000637 (sealed).
- RealNetworks' Jacobsen explained: "In essence, Microsoft has
rights of first discussion if we're going to go support something that
would compete with DirectDraw or DirectShow. The impact on us is
to put a speed bump in discussions with Sun or Netscape because
before we could conclude an agreement with Sun or Netscape, we
would have to go to Microsoft to have discussions with them."
Jacobsen Dep., 1/13/99, at 148:6-14. Microsoft, he further
explained, would be given the opportunity to propose an alternate
to the Sun or Netscape interface, and only if that solution did not
work would RealNetworks be permitted to support Sun or Netscape
interfaces. Jacobsen Dep., 1/13/99, at 149:21 - 151:4. If this
clause were invoked, Jacobsen testified, he expected that
Microsoft "would engage is some serious discussions in between
carrots and sticks" to convince RealNetworks not to support the
competitive interfaces. Jacobsen Dep., 1/13/99, at 150:23 - 151:4.
4. Microsoft's efforts to impede cross-platform Java facilitated
the maintenance of Microsoft's operating system monopoly,
hindered innovation, and harmed consumers
340. Microsoft's conduct reduced the likelihood that cross-platform Java could
erode the applications barrier to entry and thus facilitated maintenance of Microsoft's
operating system monopoly.
- John Soyring testified that Microsoft's Java-related conduct has had "a
retarding effect on the acceptance and the growth rate of Java in that our
clients, at least at a minimum, pause to think, is Java acceptable since it
may not be able to achieve this objective of `write once run anywhere.' A
second aspect is from a commercial software developer program,
commercial software vendors who want to develop and sell an application
that carries the logo or the mark design for Microsoft Windows 95 and
Windows NT. In that contract, I understand, is the provision that requires
that the vendor, if they use Java as a language, write it to the Microsoft
Java and ship the Microsoft Java virtual machine with their application.
That poses, then, a fear amongst our customers that that Java application
may only run on these Microsoft platforms and, thus, not be cross-platform capable as advertised by the other Java components." Soyring,
11/18/9pm, at 53:13 - 54:7.
- David Clare, an executive with Novell Corporation, explained in his
deposition that the fragmentation of Java is inherently harmful to
customers: "So Java, again, provides a unique opportunity -- where
everyone's trying to sell platforms, right, we sell NetWare, they sell
Solaris, they sell -- but we are trying to ensure that there is some
consistency, you know, so that we can ultimately help our customers
because it's not going to help the customers if we fragment Java, right? . .
. . Novell is very committed to -- to maintain and preserve the Java as the
consistent implementation across different platforms." Clare Dep.,
10/29/98, at 63:12 - 63:20, 79:16 - 81:3 (DX 2613).
- In October 1997, Ken Morris, Chief Technology Officer of PeopleSoft, a
significant Java application developer wrote to Steve Ballmer and Alan
Baratz (President of JavaSoft): "We are concerned that the current
course of events may result in competing definitions of Java. This
splintering of Java would destroy one of the most powerful benefits of the
technology -- that it is based on a universally agreed upon standard.
Competing Java definitions would force every application developer,
including PeopleSoft, our competitors, other software component vendors,
and in-house systems designers, to divert valuable development
resources towards porting efforts at the expense of product innovation,
much like the situation today with the various implementations of Unix."
GX 1324.
- James Gosling testified: "One of the goals of the whole Java technology
is to save developers the cost of developing on every possible virtual
machine. One of the clear goals of the Java compatibility tests is to insure
uniformity amongst all of these virtual machines so that when a developer
develops a program, they don't have to test on this virtual machine and
that virtual machine and the other virtual machine. The situation that
Microsoft's actions put us in is one where developers, by and large, can
end up having to test twice: once on a certified VM to see that they will
work across, you know, Solaris, OS/2, Apple, or whatever, and then they
separately have to test for Microsoft because of the choices that Microsoft
has made." Gosling, 12/10/98am, at 23:20-24:13.
- Netscape's Richard Schell testified that it is "of great concern that
Microsoft had gone down its own path. Ideally, in an ideal world, we
would have cooperated and collaborated with Microsoft to ensure that we
all provided a -- a common implementation. It was not our desire that
there be multiple implementations. It clearly was, to me, clearly was
Microsoft's intention to provide at least two, and that made it sort of
difficult, but at least we would have one that was cross-platform and
hopefully could influence Microsoft to -- to comply and conform with that
cross-platform implementation." Schell Dep., 9/15/98, at 71:15 - 72:9 (DX
2587).
- Gosling summarized the importance of browser competition and
unfragmented Java to software developers and users: "The preservation
of a competitive market for Internet browsers is critically important to
preserving the viability of the cross-platform Java technology. Microsoft's
actions injure not only Sun, but the entire community of Java-based
developers, computer users, and potential operating system competitors
of Microsoft . . . . If Microsoft's broad distribution of its Microsoft-dependent implementation of the Java technology in Windows, in Internet
Explorer, and in its development toolkits leads Java developers to use
these Microsoft dependencies, the potential of this technology to reduce
the barriers to developing new operating systems will be undermined, and
Microsoft's operating system monopoly will be further enhanced. In short,
if Microsoft successfully fragments the Java technology, the cross-platform benefits to vendors, developers, and users of the Java
technology will be damaged, and any threat that the Java technology
poses to Microsoft's dominant Windows operating system will be
neutralized." Gosling Dir. ¶¶ 65, 73-74.
- Soyring also testified that Microsoft's fragmentation of Java reduces the
possibility that cross-platform Java will be able to serve as a foundation
for renewed operating system competition: "Microsoft licensed Java
technology from Sun, and has released its own Java implementation.
Microsoft tuned its version of Java for Windows, inhibiting the potential for
application developers to write applications once and have them run on
many different operating systems. Wide use of Microsoft's version of
Java negates the potential that Java might undermine Windows
application advantage." Soyring Dir. ¶ 28.
- Professor Fisher summarized the impact of Microsoft's conduct: "Now,
they didn't eliminate for all time any possible threat, but they took actions
in both the browser area and the Java area to insure that what at least
they saw as a threat didn't materialize very fast. . . . It's also the case that
their actions toward Java have been such that, at least for the time being,
Java, combined with browsers or not, doesn't seem very likely to become
an alternative platform. It's been substantially impeded." Fisher,
1/6/99am, at 35:13 - 36:1; see also Fisher, 1/12/99pm, at 22:4-9
("Microsoft has, I think, also impeded the development and spread of
Java. Whether that has successfully blunted the Java threat as opposed
to merely slowed it down, I don't think one can say at this time, but I think
it certainly has done a good deal in that direction.").
341. Microsoft's argument that its conduct with regard to Java has had little
competitive impact, because the prospect of cross-platform Java is illusory cannot be
reconciled with the evidence.
341.1. Microsoft's suggestion that Sun's objective is not to encourage
competition in operating systems -- but rather to eliminate operating system competition
altogether and favor its own operating system or "put Java directly on the chip" (Muglia
Dir. ¶ 32; Muglia, 2/26/99pm, at 10:5 - 11:22; Martiz Dir. ¶ 244) -- is inconsistent with
the evidence and not relevant to assessing the competitive implications of Microsoft's
conduct.
- Gordon Eubanks, former CEO of Symantec, a significant Java tools
and runtime vendor, emphasized the cross-platform, not Sun-specific, potential of Java: "One of the great things with Java is
that when you create a Java application, it will run on any machine
that has a Java virtual machine. That was one of Sun's
contributions. Bill Joy was really the spirit behind this, one of the
great scientists at Sun. And he is right about this. If you create an
applet, it can run on any machine. And, of course, that is perfect
for the Web because you don't know when you're downloading
something, what machine it's going to be running on." Eubanks,
6/16/99am, at 68:11-20.
- Maritz sought to portray Java as a threat specifically in concert with Sun's
JavaOS, but could do so only by describing Java as a cross-platform
threat, with JavaOS as merely one example of an alternative operating
system whose success could be facilitated by the success of Java: "And
the the second point is the point that I spoke about earlier with
respect to middleware, that one of the ways that threatens
Windows is it provides a means to move the applications off the
Windows platform onto another operating system. And, in
particular, Sun Microsystems is developing an alternative operating
system that I have labeled there the 'Java OS.' And I believe, in
fact, there is some cooperation between Sun and IBM to develop
that Java OS operating system. "Q: And what exactly enables the
replacement of Windows by Java OS, as you suggest by your
diagram? A: The point here is that you can develop a version of
that Java runtime that will run on the other operating systems. You
have one version of it that runs on Windows, and another version
that runs on the other operating system, or another operating
system. But the applications, in either case, just see the
middleware layer. That makes them easy to move to the other
operating system, thereby enabling the other operating system to
compete with Windows." Maritz, 1/28/99am, at 60:9 - 61:2.
- Gosling also made clear that Sun's goal was to ensure that Java
implementation "could run on top of many other platforms: on top of the
windows platform, on top of the mac platform, and so forth." Gosling,
12/2/98pm, at 47:3-9; see also Gosling 12/2/98pm, at 67:21-25 ("the way
that you phrase that question was as though our goal was to take over one
hundred percent of the entire universe. And actually, I think we would
have been happy with two orders of magnitude less than that."); Gosling,
12/2/98pm, at 68:15-21 ("One would never actually do something
like that" put Java directly on the chip, "I mean, the Java virtual
machine was designed as something . . . that's easy to adapt to
many kinds of hardware. It's actually a rather poor design for any
specific piece of hardware, so one would never actually get rid of
the virtual machine completely."); Gosling, 12/2/98pm, at 77:5-7
("taking the Java virtual machine and doing a pure silicon
implementation of it would be, actually, a nutty thing to do, to use a
technical term."); Gosling, 12/2/98pm, at 78:17-18 ("It turns out that
the more of the virtual machine you put into silicon the slower it
goes . . . what they were trying . . . to do is not achieve high speed
but small size.").
- Indeed, as Dr. Gosling explained, Java could stimulate competition
among CPU chips, not by displacing operating systems, but by
freeing microprocessor vendors of limitations inherent in designing
for particular operating systems: "I mean, one of the things you
have to understand about the Java architecture is that it makes
applications that are distributed by developers independent of the
underlying platform. Independent of not only of the operating
system, but of the specific chip that's there." Gosling, 12/2/98pm,
at 61:9-63:14.
341.2. Microsoft's suggestion that cross-platform Java is flawed (Muglia
Dir. ¶¶ 10, 17-18) is also inconsistent with the evidence and not relevant to assessing
the competitive implications of Microsoft's conduct.
341.2.1. The evidence, including evidence from Microsoft, shows
that cross-platform Java presented a significant platform threat.
- See supra Part III.C; ¶¶ 57-59.
341.2.2. That cross-platform Java cannot yet fully replicate the
applications available for Windows means simply that the technology is evolving and
reflects that Microsoft has succeeded in retarding its development.
- As Gosling testified that: "statistically speaking, in both of
these cases of source compatibility and binary compatibility,
the number of problems--if you survey a large number of
applications, the number of problems is statistically very
small." Gosling, 12/3/98am, at 49:18-23; see also Gosling,
12/3/98am, at 32:14-18 ("While we are not in a state of being
perfect in achieving that goal, we are significantly better than
we were in the past, and we are continuing to get better."
Gosling also explained that Java has improved significantly over
time: "As people's implementations of Java runtimes have gotten
better and better, you know, . . . the need to do native methods
purely for performance has become much, much less frequent than
it was . . . . There exists a small number of cases, although with
the latest generations of things, actually going to native methods
can actually slow you down." Gosling, 12/2/98pm, at 6:3-15; see
also Gosling, 12/2/98pm, at 7 (noting that DX 2045, an article
complaining about Java performance, reflects the state of affairs
several years ago); Gosling, 12/2/98pm, at 12:3-14 (noting, with
respect to DX 1926, an article stating that Java has poor
performance, particularly on Win16: "I've spent a lot of time
with corporate customers around the world. There are
hundreds of thousands of people developing Java programs,
and doing so quite happily and quite successfully. And, you
know, for a lot of people, the actual raw performance of Java
is not often the number one concern. You know, it's just is
when you and buy a car. You don't just go and buy the
fastest car. You may buy the car that is the most reliable, or
you might have--buy the car that's got the most comfortable
seats. Yes, there have been performance issues. They're
getting better.").
- Dean Richard Schmalensee acknowledged that "many
companies are aggressively pursuing the promise of Java."
Schmalensee Dir. at ¶ 144.
- Executives at Symantec also recognized the viability of
cross-platform Java, particularly compared to the difficulties
of writing platform-specific code, notwithstanding the
immaturity of Java: "Admittedly we're now at the stage
where it's 'write once, debug everywhere.' However, that's
still far more productive than 'write once, re-write
everywhere' for the C++ and VB/VJ++ developers." GX
2078 (March 1998 Symantec e-mails including draft of public
statement ultimately not released because, although
"confirmed that everything . . . in this statement is true and
can be proved," "it's better to say nothing than risk the blast
from MS.").
341.3. Microsoft's suggestion, that AOL will partner with Sun to develop
a Java-based version of Navigator, is inconsistent with, and itself refutes, its argument
that Java is inherently flawed.
- Muglia pointed out: "Several press reports indicate that Sun will
rewrite Netscape Navigator entirely in Java . . . . Press reports
have quoted Sun management as saying that part of the
agreement will be the development by Sun of a 'pure Java' version
of Netscape Navigator browser, which will then be distributed by
AOL to all of its customers."). Muglia Dir. ¶¶ 20, 97.
- Similarly, Dean Schmalensee appeared to focus on AOL's potential
distribution of Java virtual machines as of competitive significance:
"It's my understanding that AOL has indicated it's going to
distribute . . . a very large number of Java virtual machines . . . as
part of its client software." Schmalensee, 6/21/99pm, at 95:16-23.
B. Microsoft engaged in anticompetitive conduct to induce Intel to
abandon or restrict platform-level software
342. Microsoft used anticompetitive conduct to pressure Intel into abandoning or
restricting Intel's own platform-level software efforts.
- Microsoft repeatedly objected to Intel's efforts to develop
platform-level software
343. Intel, the world's leading supplier of microprocessors (Maritz Dir. ¶ 308),
also developed software in order to improve the performance of personal computers
and thereby to stimulate demand for its microprocessors.
- Intel Vice-President Steven McGeady testified that the Intel Architecture
Lab's "agenda at that time was to make PC's sing and dance, make them
audio-aware, video-aware, in support of our move to get more people to
buy PCS for their homes." McGeady, 11/9/98pm, at 18:13 - 19:15, 41:2-6
(same); see also Maritz Dir. ¶ 313 (Intel's software development is
typically aimed at stimulating demand for newer and faster Intel
microprocessors).
344. Intel was eager to develop such software because Microsoft's operating
system had not kept pace with advances in microprocessors and other hardware,
especially in the audio/visual or multimedia area.
- Microsoft employees admitted that Microsoft had "completely missed the
boat on developing a compelling state of the art media subsystem for
Windows95." GX 563.
- Microsoft's Eric Engstrom testified that Microsoft support for game
software had stagnated for 10 years. Engstrom, 2/23/99am, at 39:10-14.
- Microsoft's Carl Stork wrote that Intel did not "want to rely on us to meet
our commitments. They have a list of commitments we have missed." GX
921 (Carl Stork 5/12/95 email).
- Microsoft Senior Vice-President Brad Silverberg wrote that "certainly we
have been remiss in not advancing the hw platform faster." Intel was
"understandably impatient with our pace." GX 921, at MS98 0168652
(Brad Silverberg 5/10/95 email).
- Microsoft Vice-President David Cole wrote about Intel: "These guys are
tired of waiting for Windows releases to make advances in hardware.
They feel the need to write system extensions to do this. We don't want
em to." GX 921, at MS98 0168652 (David Cole 5/10/95 email).
345. Despite Microsoft's admitted failures, when Intel attempted to introduce its
own software to improve the performance of the PC, Microsoft pressured Intel to
withdraw or drastically scale back every Intel software initiative that encroached on
what Microsoft considered the "platform" level.
- McGeady testified: "It had been made very clear to us on multiple
occasions when we tried to establish these application programming
interfaces . . . . Every time we tried to establish one of those in the
marketplace, we got a fight from Microsoft." McGeady, 11/9/98pm, at
35:17 - 36:1, 41:21 - 43:17. McGeady also testified that Intel "had for a
number of years been trying to innovate . . . in system software. We had
a long series of initiatives aimed at mostly multimedia optimization on the
platform . . . . After Microsoft sort of fought us one by one on those, they
finally got frustrated and just told us we had no business writing software
at that level. They owned software down to the metal. That was
Microsoft's position." McGeady, 11/10/98am, at 17:4 - 18:7.
- An Intel presentation by Craig Kinnie, dated May 4, 1995, reported:
"Microsoft firmly believes that the largest developer of Pentium Processor
based platforms has no business developing platform level software." GX
920; see also McGeady, 11/10/98am, at 23:8 - 24:25 (Kinnie was pointing
out that "he was being pressured and Intel was being pressured at other
levels not to develop any software that would exist at the same level as
the operating system, [at] platform level").
- Before the NSP initiative, Intel had developed a technology called video
display interface (VDI). VDI would have improved the performance of
digital video on the PC platform. McGeady, 11/10/98pm, at 14:4-9.
Microsoft called each and every one of the "manufacturers and persuaded
them not to adopt the software and deliver it into the marketplace."
McGeady, 11/10/98pm, at 14:10-16. This delayed introduction of the
technology approximately a year and a half until Microsoft could
incorporate it into Windows 95. McGeady, 11/10/98pm, at 14:10-22.
346. Microsoft sought to induce Intel not to offer "platform level" software (and
pressured third parties not to install it) because Microsoft viewed such software as a
potential threat that might eventually "commoditize" the PC operating system and
thereby reduce Microsoft's monopoly power.
- Paul Maritz explicitly acknowledged that Microsoft's objection to Intel's
NSP software was the establishment of "middleware APIs." GX 1309
(7/28/95 Maritz e-mail).
- Steven McGeady explained that Intel needed to expose APIs and device
driver interfaces (DDIs) in order to allow applications developers to take
advantage of new capabilities in the hardware. Microsoft did not want
applications developers to reduce their dependence on operating system
interfaces in favor of interfaces provided by others. McGeady, 11/9/98pm,
at 36:1 - 39:3.
- Bill Gates made clear to Intel that "MS doesn't want to relinquish control
over APIs" and that he believed that "'Today's API is tomorrow's DDI.'"
GX 283 (Russell Barck and Frank Ehrig memorandum summarizing
11/7/95 Intel/Microsoft meeting); GX 282 (Barck's contemporaneous notes
of Gates' remarks at 11/7/95 meeting record that Gates would "rather not
allow others to promote API.").
- Microsoft's Paul Osborne wrote: "Microsoft doesn't want Intel to be in the
system software business for the very same reason we don't want the
operating system to become a commodity." GX 921, at MS98 0168650
(5/15/95 Paul Osborne email); Maritz, 1/27/99am, at 15:12-17. Paul
Maritz explained that "commodity" means "in the software business, when
you have lots of competitors, each with roughly the same product, then
the value of your software is diminished. So by 'commodity,' we mean
here where the operating system wouldn't have the same value because"
there are reasonable alternatives. Maritz, 1/27/99am, at 16:9 - 19.
2. Microsoft engaged in predatory conduct designed to block
Intel from distributing its platform-level NSP software
a. Microsoft viewed Intel's platform-level NSP software as
a potential threat to its operating system monopoly
347. One of the platform-level technologies Intel developed was Native Signal
Processing (NSP); NSP was designed to offer consumers "substantially improved video
and graphics performance." McGeady, 11/10/98pm, at 13:24.
- Intel designed NSP to allow the PC platform to process multimedia data
(i.e., audio, video, 3D, graphics), often in real time, in order to improve
and accelerate the performance of multimedia content on PCs. McGeady,
11/9/98pm, at 18:9 - 19:15, 45:13 - 46:17.
- In addition to NSP's performance benefits, NSP also provided an interface
that would have allowed additional innovation in both hardware and
software by third-party vendors. McGeady, 11/9/98pm, at 46:18 - 47:20.
- Microsoft understood Intel's two goals with NSP to be: "(1) to grow the PC
market; and (2) to let Intel advance hardware faster than Microsoft would."
GX 921, at MS98 0168653 (5/10/95 Stork email).
348. Microsoft believed that NSP had the potential to be a significant software
platform that eventually could pose a threat to Microsoft's monopoly power in PC
operating systems.
- NSP exposed APIs that developers could use. Maritz, 1/27/99am, at
12:2-17.
- Maritz acknowledged that NSP had the potential to become a significant
software platform. Maritz, 1/27/99am, at 11:23 - 12:22. This testimony is
corroborated by Microsoft e-mail: "What makes Intel unusual is they
actually stand a chance of being successful" at establishing platform level
software. GX 921, at MS98 0168650 (5/15/95 Paul Osborne email).
b. Microsoft blocked platform-level NSP through predatory
conduct
349. In order to stem the NSP threat, Microsoft took steps to prevent Intel from
shipping NSP.
349.1. First, Microsoft told Intel that it objected to NSP's platform-level
software and made clear that, if Intel persisted in distributing NSP, Intel's relationship
with Microsoft would suffer.
- In April 1995, three Microsoft executives, Carl Stork, Marshall
Brumer, and Eric Engstrom, met with Gerald Holzhammer and
other Intel executives and discussed NSP. The Microsoft
executives expressed displeasure at Intel entering "'their' OS
space." The Microsoft contingent also said that Microsoft needed
"to own ALL driver software 'to the metal', i.e., silicon." GX 563.
- Microsoft intensified the pressure to drop NSP over the next few
months. In May 1995, Microsoft vice-presidents Paul Maritz and
Brad Silverberg, along with Microsoft's Stork, met with Intel
executives to discuss Intel's NSP program. The Microsoft
executives complained that Intel was shifting the software
boundary with its NSP project by writing software that Microsoft
considered to be part of its operating system space. GX 275;
McGeady, 11/9/98pm, at 23:3 - 26:23. In Microsoft's view, NSP
made Intel a competitor in Microsoft's operating system space. GX
275; McGeady, 11/9/98pm, at 26:25 - 27:11.
- Stork summarized Microsoft's reaction to NSP in his report of the
May 1995 meeting. GX 921, at MS98 0168653. Stork noted that
Kinnie of Intel "frankly admits that NSP is a system software
platform that they want people to write to, and that lets them put in
new function without dependency on us." Stork concluded: "The
bottom line is that Intel wants to enable new hardware function in
our OSs and to set 'underware' system standards, without being
dependent on MS' inclination or execution." Stork recommended
that Microsoft should not allow others to offer platform level
software, even if that meant a slower rate of innovation. If Intel
wished to enable a new function, Stork wrote, Intel's only "winning
path" should be to convince Microsoft to support the effort and
sometimes "to accept the outcome that the timing isn't right" for
Microsoft. GX 921, at MS98 0168653 (5/10/95 Stork email).
- Bill Gates explained to Intel's Andy Grove that Intel's attempts to
compete with, rather than follow, Microsoft in software were
unacceptable: "The problem we have is that we have to sort of
choose in software related issues which company will lead and
which will follow. In chips its very clear. In software you have a
group that won't allow us to lead and has all the prestide (sic) and
profits of Intel to drive them forward." GX 277 (5/25/95 Gates
email). Gates later elaborated that Microsoft is the "software
company here and we will not have any kind of equal relationship
with Intel on software." GX 278 (7/7/95 Gates email to Microsoft
executives).
- In early July 1995, Gates reported that he had tried to convince
Grove "to basically not ship NSP." GX 278 (7/7/95 Gates e-mail to
Microsoft executives); Maritz Dir. ¶ 320. Gates predicted that Intel
would exert less pressure to ship NSP in 1995 but that it "will take a
major effort for us to convince them to back off from this for 1996."
GX 278 (7/7/95 Gates email to Microsoft executives).
349.2. Second, Microsoft succeeded in blocking Intel from distributing
NSP by pressuring OEMs not to install it.
349.2.1. OEMs provided the only channel through which Intel
could distribute NSP because NSP could not feasibly be installed by end users.
- McGeady testified that NSP "was a piece of system
software. It wasn't an application like a normal end user
would install on their personal computer. It really needed to
be installed by the PC manufacturer at the time they loaded
the operating system for everything to worth smoothly, and
so it was very important to us that that channel of getting
those PC OEM's to adopt this technology" was available.
McGeady, 11/9/98pm, at 20:9 - 20:22. The "two viable
ways" to distribute NSP were thus through the Microsoft
operating system or through OEMs. McGeady, 11/9/98pm,
at 21:4-12.
349.2.2. Microsoft choked off this essential channel by using its
monopoly power over PC operating systems to pressure OEMs not to presinstall NSP.
- On May 31, 1995, Stork articulated the "plan of record"
based on his conversation with Maritz: "meet with compaq
next week, and the most important partners the week
following" in order "to keep evangelizing" the OEMs not to
install NSP. GX 923 (5/31/95 Stork email).
- On June 6, 1995, Maritz reported the plan's success to
Gates: "We should continue to do what we have done which
is to tell the OEMs and ISVs that we don't agree with them
using or installing NSP. Thanks to the Internet that
message is already out. No major OEM is now going to
install NSP. In that sense it is already dead." GX 924
(6/6/95 Maritz email).
349.2.3. Microsoft succeeded not only in inhibiting OEMs from
shipping NSP, but also in making OEMs reluctant to adopt other Intel innovations
without Microsoft's express permission.
- On October 18, 1995, Gates reported to Maritz and Jim
Allchin: "Intel feels we have all the OEMs on hold with our
NSP chill. For example, they feel HP is unwilling to do
anything relative to MMX exploitation or the new audio
software Intel is doing using Windows 95 unless we say its
ok." GX 281 (10/18/95 Gates email).
- Steven McGeady testified that it was "common knowledge"
that Microsoft was having a "chilling effect" on OEMs and on
the acceptance of Intel's Internet technology. McGeady,
11/9/98pm, at 32:18 - 33:17; see also McGeady,
11/12/98pm, at 36:11-17 (discussing GX 281).
349.3. Third, Microsoft, in addition to wielding these sticks, held out to
Intel the carrot of including some components of NSP in subsequent releases of
Windows, if Intel stopped promoting NSP's interfaces -- in other words, if Intel stopped
positioning NSP as a platform.
- Maritz testified that Microsoft and Intel agreed that Microsoft would
accelerate work with Intel to incorporate NSP functions into
Windows 95, provided that Intel limited NSP distribution for
Windows 3.1 to a "few specific hardware manufacturers" to whom
Intel had existing commitments. Maritz, 1/27/99am, at 37:1-8;
Maritz Dir. ¶¶ 317-321.
- On May 31, 1995, Stork reported that Maritz had proposed to
Intel's Ronald Whittier "that if they would hold off on pushing OEMs
to install NSP," then Microsoft "would hold off on selling against"
NSP. GX 923.
- McGeady testified: "They wanted us to stop delivering this stuff,
back it out of the marketplace, and wait until we could give it to
them to integrate into future operating system, the timing of which
was undefined." McGeady, 11/10/98pm, at 44:13-22.
350. As a result of Microsoft's pressure, Intel abandoned NSP as a platform-level technology and agreed to Microsoft's offer to incorporate some aspects of NSP
into Windows, thereby eliminating NSP's potential to threaten Microsoft's operating
system monopoly.
- On July 28, 1995, Paul Maritz wrote that Intel had "given up on our
original major objections which were that they were trying to establish
middleware APIs" and jam "random, unrelated stuff" under NSP. GX
1309.
- Steven McGeady explained Intel's motivation for abandoning NSP: "Intel
did fail to introduce NSP into the marketplace because, as a primary
cause, Microsoft, in particular Bill Gates, told Andy Grove that Microsoft
did not want NSP in the marketplace" and because "Microsoft helped us
late in our business interests by threatening to withhold support for other
microprocessors in the meantime." McGeady, 11/10/98pm, at 81:6-23.
McGeady further testified that, although some components of NSP were
given to Microsoft for incorporation into later releases of Windows, the
"key capability of NSP, which was realtime multimedia management below
the operating system was never released," nor were some other key
components, such as its 3D rendering system. McGeady, 11/10/98pm, at
76:5-22, 11:24 - 13:9 (although Microsoft did incorporate some NSP
components in later operating system releases, key capabilities that "NSP
would have brought to bear still are not present on the PC today").
- Rob Sullivan of Intel testified to the same effect: "Microsoft opposed it on
every front. . . . In that time frame, we were looking at multiple billion-dollar investment streams for Merced and MMX . . . . The successful
launch of those products was critical [to] our core strategy. It just wasn't
worth it . . . to try to go and fight this issue. We had products in the
marketplace . . . without operating systems support . . . having borne the
cost of creating those products. We gave in." Sullivan Dep. (played
11/12/98pm), at 47:10 - 48:12.
- An October 1995 Gates email to some of his top staff also corroborates
McGeady's testimony: "Intel feels we have all the OEMs on hold with our
NSP chill. For example they feel HP is unwilling to do anything relative to
MMX exploitation or the new audio software Intel is doing using Windows
95 unless we say its ok. . . . Andy believes Intel is living up to its part of
the NSP bargain and that we should let OEMs know that some of the new
software work Intel is doing is Ok. If Intel is not sticking totally to its part of
the deal let me know." GX 281.
3. Microsoft used its monopoly power to ensure that Intel did not
resume developing or supporting platform-level software
351. In addition to pressuring Intel into abandoning NSP, Microsoft used its
operating system monopoly to ensure that Intel did not resume developing or
supporting rival platform-level software.
351.1. At an August 2, 1995, meeting at Intel's headquarters with Andy
Grove and other high-level Intel representatives, Bill Gates again attacked Intel's
software efforts and expressly threatened to withhold Microsoft's support for Intel's next
generations of microprocessors if Intel continued to develop or support rival platform-level software or assisted Netscape or Sun in such efforts.
- Gates expressed outrage that Intel had 600-700 engineers, in his
view, competing with Microsoft (McGeady,11/9/98pm, at 10:20 -
11:5), and that Intel used its profits from microprocessors to write
software that competed against Microsoft's. McGeady, 11/9/98pm
at 13:11 - 14:3; GX 279 (Intel Vice-President Ron Whittier's 8/2/95
minutes of the meeting: "Gates issue: fundamental problem with
'free' software from IAL cross-subsidized by processor revenues.");
GX 280 (8/28/95 McGeady memorandum: "On August 2 1995, in a
meeting of Intel and Microsoft executives, Bill Gates told Intel CEO
Andy Grove to shut down the Intel Architecture Labs."). Gates was
upset that Intel was "making investments in software of any sort."
McGeady, 11/9/98pm, at 11:1-5. Gates "felt that anything" Intel
"did in software was competitive or would harm Microsoft and felt
that we shouldn't be using our profits from our micro processor
business to write software that compete with them." McGeady,
11/9/98pm, at 13:11 - 14:3; DX 1805 (Fred Pollock memorandum
re 8/2/95 meeting: "We then had about 1-hour 'interlude' of Bill
Gates bashing IAL.").
- Gates went on to make clear that "Microsoft would not support"
Intel's "next processor offerings if we did not get alignment between
Intel and Microsoft on platform issues" and on communications
issues, i.e., Internet issues. McGeady, 11/9/98pm, at 14:14 - 15:4;
GX 279 (Whittier's minutes: "Gates would not agree to let
processors/OS's programs to progress unencumbered by platform,
communications program issues."). At this time Intel was seeking
Microsoft support for its planned MMX processors and its 64-bit
microprocessor (then code-named "P7"). McGeady, 11/9/98pm, at
15:9-22.
- In addition to setting limits on Intel's software efforts, Gates raised
"Internet issues." GX 279. Gates cautioned that Microsoft was
"very sensitive to what Intel might do on the client side. Example:
JAVA, a show stopper." GX 279. By "client," Gates meant
"browser." Maritz, 1/27/99am, at 27:12-21.
- Whittier's minutes of the meeting reflect Gates' explanation of what
Microsoft expected from Intel: "On the 30/70 use of 3rd party
technologies, Intel using Netscape in Windows environment is not
a problem (provided we do not set up the 'positive feedback loop'
for Netscape that allows it to grow to de facto std.)." GX 279. As
McGeady testified, Microsoft agreed to Intel's internal use of
Netscape browser as stand-alone application on Windows, but
anything Intel did to encourage applications developers to use the
Netscape browser as a platform, including any Intel/Netscape
development or technical work, would be a problem with Microsoft.
McGeady, 11/12/98pm, at 19:5 - 20:20; GX 279 (Whittier's minutes:
"BG: Supporting certain third party deals will be problem . . . we
need to consider in the context of their (pervasive) internet program
to assure we are not unknowingly stepping on one of their key
strategies!").
- Gates also told Intel to devote its Internet resources to a high-end
web server (GX 279, at 3), a product that would have little impact in
the web client market. McGeady, 11/9/98pm, at 12:13 - 13:10.
- At this time, Intel had been engaged in an effort to work with Unix
vendors to increase compatibility among the various versions of
Unix. McGeady, 11/12/98pm, at 33:6 - 34:4. At the August 2,
1995, meeting, Gates objected to any effort by Intel to increase
compatibility among Unix vendors: "Gates was very concerned that
Intel was back in the business of 'unifying Unix.' He does not like
us trying to unify Microsoft's competitors." DX 1805 (Fred Pollock
memorandum regarding the 8/2/95 meeting, noting that Gates'
objection arose in the context of discussions regarding Intel's next
generation of microprocessors); GX 279 ("UNIX: big flap -- MS
wants lots of UNIXes.").
351.2. Intel found Bill Gates' threat not to support Intel's next generations
of processors "credible and fairly terrifying" because its processor business is
dependent on Microsoft's support.
- McGeady explained the import of this threat: "It was clear to us
that if those processors didn't run Windows, they would be useless
in the marketplace, so the threat was both credible and fairly
terrifying. . . . The Windows operating system commands a very
large share of the operating system market. If our processors
aren't supported by that operating system, very few people would
buy them. They wouldn't run the software most widely used by our
customers." McGeady, 11/9/98pm, at 15:19 - 16:5.
- McGeady further explained that, if Intel "didn't tow (sic) the line on
at least some of these software programs, that Microsoft would
continue bad-mouthing not only Intel software, but perhaps more
specifically fail to support MMX. The effect of this would be slow or
no adoption of it by the PC manufacturers, and the result of that
would be a failed chip introduction from us, and a big problem."
McGeady, 11/9/98pm, at 44:2-11. See also McGeady, 11/9/98pm,
at 43:2 - 45:12 (describing the impact of Microsoft's withholding of
support for MMX, a technology in which Intel had invested
approximately $500 million).
352. Microsoft's use of its operating system monopoly to threaten Intel had its
intended effect: Intel did not resume platform-level NSP and recognized that it should
seek alignment with Microsoft before supporting or offering alternative platform-level
technologies.
- McGeady testified that Intel withdrew NSP "because the cumulative value
of the various threats that MS could bring to bear was a greater risk than
we were willing to take." McGeady, 11/12/98pm, at 48:23 - 49:3.
McGeady further testified: "For some time we continued to look for a way
to ship the overall integrated version of the NSP technologies. We were
never able to do that." McGeady, 11/12/98pm, at 36:1-3. Although some
components of NSP were given to Microsoft for incorporation into
Windows 95, the "key capability of NSP, which was realtime multimedia
management below the operating system was never released, and other
key components, including the 3-D rendering system, was never released
. . . ." McGeady, 11/10/98pm, at 76:18-22.
- Intel's notes of a November 7, 1995, meeting between Intel and Microsoft
executives, including Gates and Maritz, report: "PM [Paul Maritz]: Agreed
to release RL DDI under gentlemen's agreement that we [Intel] pull-out of
3DR assuming RL meets Intel desired requirements." GX 282. RL stands
for Reality Labs, Microsoft's 3D rendering engine, and 3DR was Intel's
competitive 3D rendering software. Prior to the meeting, Intel had been
seeking but was unable to obtain Reality Labs DDI from Microsoft. Intel
pulled back its 3DR and API evangelism efforts after the meeting. Barck
Dep., 8/25/98, at 36:13 - 37:21, 28:19 - 29:5 (DX 2556).
4. The effect of Microsoft's conduct was further to entrench its
operating system monopoly, hamper innovation, and deprive
consumers of the benefits of Intel's platform-level software
353. By forcing Intel to abandon platform-level NSP and other software
initiatives that Microsoft did not approve, Microsoft impeded innovation by denying to
consumers the benefits of hardware and software advances not controlled by Microsoft.
- Microsoft's Carl Stork bluntly explained Microsoft's position: If Intel wants
to enable a new function, the only "winning path" is convince Microsoft to
resource the effort. If the proposal does not fit with Microsoft's priorities,
Intel has no choice but to "accept the outcome that the time isn't right for
us." GX 921, at MS98 0168652 (5/10/95 Stork email).
- Steven McGeady testified: "The hardware vendors . . . both at the time
and now, continue to be frustrated . . . because they had more ideas
about new ways and interesting ways to do things that would have benefit
for the end user than they were able to get pushed up through the
operating system layers." McGeady, 11/9/98pm, at 47:4-10; see also
McGeady, 11/9/98pm, at 45:13 - 47:20 (detailing the benefits NSP offered
consumers); McGeady, 11/9/98am, at 36:2 - 41:8 (explaining how
alternate platform-level interfaces can foster innovation and deliver
greater performance to consumers); GX 563 (4/13/95 Holzhammer email
explaining that if Microsoft controls all driver level software, "innovation at
the HW level would grind to a halt since silicon vendors would need to rely
on MS to get driver support for their new stuff.").
- McGeady further testified: "The more competitive and diverse a software
environment - application development environment is, the more
innovation occurs and the more different options are presented to
consumers. Correspondingly, as the software environment has become
more of a monoculture around Microsoft, the rate of innovation appears to
be slowing, and the number of different and varied options presented to
the consumer is diminishing." McGeady, 11/9/98pm, at 61:24 - 62:7.
- Professor Fisher testified that Microsoft's pervasive control over
innovation harms consumers because, if "we're going to live in a Microsoft
world" that "may be a nice world, but it's not a competitive world, and it's
not a world that's ultimately consumer-driven." Fisher, 1/7/99am, at 27:11
-28:17. "Microsoft has been giving out very, very, strong signals that
innovation is fine and they will cooperate with it. They will even assist it, if
what you're doing is producing simply complements for Microsoft
products. But they will take very, very aggressive action against you if
what you're doing is producing innovations that might lead to something
that threatens their operating system monopoly. . . . It rather
discourages, I should think, people from thinking of ways to . . . innovate
in ways that would threaten" Microsoft's operating system monopoly.
Fisher, 6/2/99am, at 25:7-26:3.
- With regard to Intel, Netscape's Marc Andreessen wrote: "if they only
have Microsoft as a single channel to innovate on the PC platform, then
Microsoft controls the rate of innovation and slows things down to suit
Microsoft's interests, which is not in Intel's best interest." DX 1619
(9/18/95 Marc Andreessen email to other Netscape executives).
354. Microsoft's efforts to blunt Intel's platform-level software also eliminated a
potential threat to Microsoft's control over APIs, and thus a potential threat to its
monopoly power, and deterred other threats from arising in the future.
- Intel abandoned NSP's platform-level aspects because of pressure from
Microsoft. See supra ¶ 350.
- Professor Fisher testified: "Microsoft has given signals to the world, both
through these [its actions with regard to browsers and Java] and through
its actions as regards Apple and Intel, Microsoft cares a lot about whether
there are going to be innovations that might, in one way or another,
present a platform threat; and that if you want to make innovations in that
direction, you're going to have to deal with Microsoft in a very, very
serious way. That is also a way of blunting or preventing future platform
threats." Fisher, 1/12/99pm, at 22:10-18; see also Fisher, 6/2/99am, at
25:7 - 26:3 (expanding on same concept).
5. Microsoft's contention that technical considerations explain
its objections to NSP is pretextual, and the testimony of its
witnesses regarding NSP is not credible
355. Microsoft argued that it objected to NSP because it was originally designed
for Windows 3.1 and thus would slow the adoption of, or could at some point cause
technical problems with, Windows 95 (Maritz Dir. ¶¶ 316-322). But this argument
cannot explain Microsoft's conduct.
355.1. First, contemporaneous Microsoft documents make clear that
Microsoft's overriding objection to NSP software was its platform-level attributes, rather
than the fact it was initially designed to work only with Windows 3.1 or any technical
issues.
- Early records of Microsoft's reaction to NSP show that Microsoft's
major problem with NSP was not the technical merit of the
software, which Microsoft had not at that point evaluated, but rather
that Intel was entering Microsoft's "OS space," which conflicted with
its goal of owning "ALL driver software 'to the metal', i.e., silicon. . .
." GX 563.
- Microsoft's Paul Osborne candidly explained: "Microsoft doesn't
want Intel to be in the system software business for the very same
reason -- we don't want the operating system to be a commodity."
GX 921, at MS98 0168650 (Osborne 5/15/95 email) . For that
reason, Osborne wrote, it was important to "neutralize this situation
with Intel" regardless whether it was "Good code or bad,
competitors or not." GX 921, at MS98 0168650 (Osborne 5/15/95
email); see also GX 921, at MS98 0168652 (5/10/95 Stork email
expressing dissatisfaction that NSP would allow Intel to "put in new
function without dependency on us").
- Microsoft's David Cole did not mince words on this subject of
making advances in hardware: Intel "feels the need to write system
extensions to do this. We don't want em to." GX 921, at MS98
0168652 (5/10/95 David Cole email).
- After Intel had agreed to cease promoting the NSP APIs and DDIs,
Paul Maritz explained that he was now giving Intel "the benefit of
the doubt" because "they have given on our original major
objections, which were that they were trying to establish
middleware APIs, and that they were using NSP to jam all sorts of
random, unrelated stuff. Now they are just purveyors of big, slow
software." GX 1309 (7/28/95 Maritz email) (emphasis added).
355.2. Second, Microsoft's contention is pretextual because technical
concerns about NSP cannot explain either (i) Microsoft's threats relating to Intel's next
generation of microprocessors or (ii) Microsoft's insistence that Intel not compete in
platform software even after Intel had abandoned NSP for Windows 3.1.
- See supra ¶ 351.1.
- See supra ¶ 351.
355.3. Third, Microsoft's contention that encouraging Windows 95's
adoption justified its efforts to kill NSP is belied by Microsoft's own continued support of
Windows 3.1.
- Microsoft continues to offer Windows 3.1. E.g., GX 1188 (sealed)
- redacted -
- Dr. Warren-Boulton testified that continuing to support older
versions of Windows is "a normal profit-maximizing thing" for
Microsoft to do. Warren-Boulton, 11/19/98am, at 64:7-17.
355.4. Fourth, Microsoft's contention that it simply had technical
objections to NSP is belied by the fact that Microsoft insisted that Intel abandon NSP's
platform-level efforts even though Intel was prepared to solve any technical problems
and to offer a Windows 95 version.
355.4.1. Intel had developed by June 1995, and was about to
begin testing, a beta version of NSP for Windows 95 and, absent Microsoft's predatory
efforts to block it, likely could have solved any remaining technical issues.
- In June 1995, Intel had a running beta version that was
about to begin the quality assurance testing process.
McGeady, 11/10/98pm, at 24:5 - 25-11.
355.4.2. PC OEMs would not have installed Windows 95 and NSP
for Windows 3.1 on the same machines.
- NSP could be installed only by OEMs (McGeady,
11/10/98pm, at 25:17 - 26:4), who have every incentive to
ensure that the software they select properly functions, as
that reduces their support expenses. See supra V.C.1.b.(1);
¶ 178.1; McGeady, 11/10/98pm, at 55:3 - 56:1.
355.4.3. To the extent that potential risks may have arisen with
subsequent customer upgrades to Windows 95, there are a number of means that Intel,
the OEMs and Microsoft commonly employed to deal with potential problems with
upgrades. There is no reason to believe that NSP presented unsolvable or unusual
difficulties.
- Since OEMs would not preinstall NSP on a Windows 95 PC
(until Intel released the Windows 95 version of NSP), the
only scenario that could potentially pose a compatibility
problem is that a customer with an NSP/Windows 3.1 PC
might later decide to install a retail version of Windows 95 or
Windows NT on the PC (McGeady, 11/10/98pm, at 25:17 -
26:4). McGeady testified about several methods commonly
employed to deal with problems of this type (McGeady,
11/10/98pm, at 35:20 - 36:22; 37:7-16), that Intel and the
OEMs had arrangements for technical support for NSP
(McGeady, 11/10/98pm, at 38:15 - 39:5; 55:9 - 56:1), and
that NSP did not present unusual risks compared to other
PC software and device drivers (McGeady, 11/10/98pm, at
40:7-24).
356. Testimony offered by Paul Maritz and Bill Gates regarding Microsoft's
efforts to induce Intel to cease developing platform-level software is either unreliable or
incredible.
356.1. Mr. Maritz' testimony that Microsoft simply told Intel what was in
Intel's best interest (Maritz, 1/27/98am, at 28:17 - 29:22, 34:15 - 25) and that the
discussions regarding NSP were merely a benign part of an ongoing dialogue resulting
from the close relationship between the companies' products (Maritz Dir. ¶¶ 308-309) is
not credible.
- The threat Microsoft made to Intel at the August 2, 1995, meeting
was not a procompetitive, routine discussion between producers of
complementary products; it was, rather, a blunt warning that Intel
should not continue to pursue developments that could threaten
Microsoft's monopoly power. See supra ¶ 351.
- The contemporaneous documents explicitly show that Maritz
offered to stop "selling against" Intel if Intel held "off on pushing
OEMs to install NSP." GX 923.
- By contrast to the picture he painted at trial, the documents Maritz
authored in the course of his work make clear that Microsoft's
principal concern with NSP was the possibility that Intel might
control a new set of "middleware APIs." GX 1309 (7/28/95 Maritz
e-mail).
356.2. Mr. Maritz' testimony that, after Intel agreed to limit its distribution
of NSP with Windows 3.1 in return for certain Microsoft assistance regarding Windows
95, Intel later withdrew NSP "without further communication, to my knowledge, from
Microsoft" (Maritz Dir. ¶ 322) is inconsistent with the contemporaneous evidence of
further Microsoft/Intel communications.
- Gates wrote to Maritz and other top Microsoft executives that Andy
Grove "believes Intel is living up to its part of the NSP bargain" and
asked the recipients to let Gates know if "Intel is not sticking totally
to its part of the deal." GX 281 (10/18/95 Gates email).
356.3. Mr. Gates's testimony that Microsoft did not express concern to
Intel about Intel's platform-level software development work and that Gates did not
even know about Intel's software development work (Gates Dep. (played 11/9/98am),
at 69:8-19) is not credible in light of both Steven McGeady's testimony and the
contemporaneous evidence.
- McGeady testified that Intel had briefed Microsoft and Gates on its
Internet software effort several times and that he had personally
briefed Gates at the August 2, 1995 meeting. McGeady,
11/9/98pm, at 8:3-13; 9:7 - 10:19.
- Gates himself wrote on July 7, 1995, that Intel had an Internet
software group of 100 people working under McGeady. GX 278
(7/7/95 Gates e-mail).
- Whittier's minutes of the August 2, 1995, meeting reflect that Gates
warned Intel against supporting Netscape or Java as a alternate
platform or stepping on any of Microsoft's key Internet strategies.
GX 279.
356.4. Mr. Gates' testimony that Microsoft did not threaten to withhold
support for Intel if Intel failed to cooperate with Microsoft (Gates Dep. (played 11/9/98
am) at 70:22 - 73:10) is also not credible in light of the evidence.
- See supra ¶ 351 (summarizing McGeady's and Sullivan's testimony
and Whittier's minutes (GX 279) regarding the August 2, 1995,
meeting).
- In discussing Intel's request that Microsoft tell OEMs that some of
Intel's software is okay to install, Gates asked his executives to let
him know if "Intel is not sticking totally to its part of the deal." GX
281.
FOOTNOTES
1. GX 1129, at MSV 0005245 (Packard Bell); GX 647, at MSV 0002127 (Toshiba)
(sealed); GX 1183, at MS98 0009095 (Hewlett-Packard) (sealed); GX 1374, at MS98
0009538 (Hitachi) (sealed); GX 458, at MS98 0009146 (Gateway (sealed); GX 1375
(Sony) (sealed); GX 1060, at MS98 0009311 (AST) (sealed); GX 1373, at MS98
0009399 (Micron) (sealed).
2. See, e.g.,Warren-Boulton Dir. ¶ 103 (outlining typical terms); GX 1213, at MS6
5000388 (sealed) (AT&T Agreement, § 3.3); GX 1148, at MS6 5001000 (sealed) (
Prodigy Agreement, §§ 3.3, 3.4); GX 1134, at MS6 5000172 (CompuServe Agreement,
§ 3.3); GX 804, at AOL 0001738 (AOL Agreement, § 7.2); Myhrvold 2/9/99pm, at 77:14-19 (conceding Microsoft's contracts restricted the distribution of other browsers);
Myhrvold 2/10/99am, at 6:14-22 (admitting that his testimony that Microsoft imposed
distribution restrictions was "absolutely wrong" was itself wrong); Chase Dir. ¶ 95
("Although the particulars vary somewhat from OLS to OLS . . . [their agreements] are
similar in many ways to Microsoft's agreement with AOL.").
3. See, e.g., Fisher Dir. ¶¶ 184-185 (detailing typical terms of ISP agreements);
GX 1141, at MS6 5000007 (sealed) (Earthlink agreement, § 3.1); GX 1140 (summary of
the Brigadoon agreement); GX 1147 (summary of the IDT Corp. agreement); GX
1144,at MS6 5001130 (sealed) (SpryNet agreement, § 3.1); GX 1146, at MS6 5000924
(sealed) (Mindspring agreement, § 3.1); GX 1213,at MS6 5000388 (sealed) (AT&T
agreement, § 3.3); GX 1214,at MS6 5000953 (sealed) (Netcom Agreement, § 3.1).
4. GX 1163, at CNET 000032 (7/14/97 MS/CNet ICP agreement, section 2.3); GX
1164, at p. 4 (7/15/97 MS/CondeNet ICP agreement, section 2.3); GX 1175, at AOL-0000149 (9/10/97 MS/AOL agreement, section 2.3); GX 859, at AOL-0000123 (AOL
Summary of MS Active Desktop Agreement, dated 9/23/97, at AOL 0000149); Colburn
Dir. ¶ 42; Colburn, 10/29/99pm 51:8 - 54:24 (explaining active desktop restrictions); GX
1156, at INT 00005 (6/6/97 MS/Intuit agreement, section 2.2); GX 206 (internal
Microsoft email from Will Poole to Brad Chase, dated 4/17/97, in which Poole
summarizes terms of Intuit agreement); GX 1176, at TWDC 0710 (7/3/97 MS/Disney
agreement, section 2.3-2.5); GX 874, at TWDC 0299-300 (Disney Online and Microsoft
Active Desktop Agreement Summary of Terms); GX 856, at MS98 0100299 (7/3/97
MS/Disney agreement, sections 2.3-2.4); Barksdale Dir. ¶ 182 (Disney prohibited from
offering Netscape compensation of any kind); GX 776 (Wadsworth Decl. ¶5); GX 1163,
at CNET 000032 (7/14/97 MS/CNet agreement, very similar to Intuit, section 2.3); GX
855, at WD 0004 (7/1/97 MS/Wired agreement, section 2.3); GX 1210, at p.4 (7/9/97
MS/Sportsline agreement; MS to put icon for Sportsline on sports channel, section 2.3);
GX 1209, at p.4 (7/25/97 MS/MTV agreement; MS to put MTV icon in entertainment
channel, section 2.3); GX 1211, at ZD 0005 (8/6/97 MS/ZDNet agreement; MS to put
ZDNet icon in appropriate subchannel, section 2.3); GX 1170, at p.4 (8/15/97 MS/NBC
agreement, section 2.3); GX 1174, at MS98 0100073 (sealed) ( , section 2.3); GX 1159, at TM 000057 (6/26/97 Hollywood Online
agreement, section 2.3); GX 1164, at p.4 (7/15/97 MS/CondeNet agreement, section
2.3); GX 1165, at TWDC 0372 (7/17/97 MS/ESPN agreement, section 2.3); see also
Fisher Dir. ¶ 195 (summarizing restrictions); Warren-Boulton Dir. ¶ 115 (same).
| |