Hi,

Here's my minutes from today's meeting. One thing we forgot to decide
on is when the next meeting should be. I'll be away next week so I'm
inclined to skip a week and make the next one June 14th. There should
be real progress to report by then, I hope.

--

Callers: T. Berfeld, B. Cole, M. Heffner, D. Lange, M. Messier,
S. Johnson, R. Raja, S. Seun, R. Soltz

Online: Harvard has taken devilery of the loaner PC board. Andre
L. will be setting it up next week while Mark is on vaction. Both
Andre and Mark should be added to the p907daq listserv mail list.

Monte Carlo: Raja has still somework to do to fix the MC
geometry. Hit/Digitization for most detectors is in place. Tasks
remaining are:

1) RICH digitization: Sharon has the code in place but is working on a
   bug that prevents the geantinos (used for Cherenkov light tracking)
   from finding the mirrors. May be problem with geometry layout.

2) TOF - the code for the TOF hits needs to be put into the "dd_geant"
   framework. Tim and Raja will work together on this.

3) A design is needed for the beam Cherenkov's and they need to be
   added to the MC.

Offline framework: There was much discussion of getting started on an
offline framework. Proposals are:

1) We will plan to use ROOT I/O for the event data as it has become a
   standard used by many experiments.

  - Mark will think about a proposal for a data model based on root
    I/O and send it to Brian C. to get input based on E910 experience.
    Expected plan is that there will be three stages of data
    processing, RAW data files will be processed to create files with
    RAW and reconstructed data combined in the same file. From
    these, data summary files will be produced. It is thought that the
    data volume for the experiment is low enough that this duplication
    should be OK.

  - Mike H. will make some tests of the memory and speed overheads for
    writing root objects out of the DAQ.

2) The MIPP software will be C++ based. Mark will start an offline
   software web page and post a set of coding conventions there for
   comment. The MC will stay GEANT3 (Fortan based) and other existing
   code may be wrapped to avoid significant re-coding.

3) We will plan to use PostgreSQL (http://www.us.postgresql.org/) as
   the database package. Work to build the interface to this package
   needs to be done. Brian C. will post a summary of the interface
   used by E910. As a test case, one of the geometry tables used by
   the MC will be migrated to use PostgreSQL.

4) We will plan to use SRT (http://www.fnal.gov/docs/products/srt/)
   for the build system, as well as package and release
   management. This is used by many experiments:
   Raja reports that D0 uses it:
     http://www-d0.fnal.gov/software/cmgt/cmgt-example.html) And David
   David L. reports that BaBar uses it:
     http://www.slac.stanford.edu/BFROOT/www/Computing/Environment/Tools/SRT/

These proposals will be adopted after the collaboration has had
sufficient time to review them.

Other things to do: 

- Steve J. will investigate toolkits to use for the event display.
  The detector coordinators will be contacted (by Peter B., if he
  agrees) to supply information about what data their systems will
  produce so that the DAQ/offline data objects can start to take
  shape.

Existing software:

- Brian C. will post the E910 TPC code so people can become familiar
  with it. This code will be mirgrated into the 907 repository as
  progress on the 907 data and package model evolves.

- Raja will send Mark and Sharon a pointer to the SELEX RICH
  (fortran) reconstruction code. Depending on complexity, the plan is
  to either re-write or wrap it in C++.