The Technical Design of the D0 Parameterized Monte Carlo Simulation (PMCS)

                            PMCS Working Group

Outline

1. PMCS Goals
  A. PMCS should be a repository of the best smearing routines known to D0
  B. PMCS should be fast for use in fast monte carlo efforts such as FMC++ and
     the new CMS.  
  C. Smearing routines in PMCS should be easily tunable. 

2. PMCS Requirements
  A. The design of PMCS should not preclude its use by D0 collaborators.
  B. The design of PMCS should facilitate multi-user development. 
  C. PMCS should be flexible to meet the needs of users who might want to 
     switch off certain parts of the simulation to increase speed.
  D. PMCS should support D0 standard output formats. 

3. PMCS technical design
  A. PMCS will consist of a number of largely independent framework packages. 
  B. PMCS output will be standard D0 chunks.
  C. PMCS input will be a modified MCKineChunk that contains pointers to 
     particle types 
  

1. PMCS Goals

1.A There is a desire within D0 and good sense dictates that the best 
parameterizations of the D0 detector should be kept logically in one place.
People should not have to re-invent the wheel, and non-experts will have 
access to the same smearing routines that experts use. The output of PMCS
should include : 
  i) smeared physics objects : em, muon, tracks, jets
  ii) smeared global quantities : Met, recoil, etc 
  iii) smeared detector quantities : toy cal, toy sili, etc.
  iv) access to generated quantities

1.B PMCS is designed to be a fast simulation of the detector response.
The design of PMCS should not preclude its use in conjunction with fast 
monte carlo efforts such as FMC++ SUSY search or CMS.  When used in 
conjunction with such monte carlo efforts, it is understood that they
are responsible for the physics events and PMCS is responsible for detector
simulation.  The inputs of these simulations may be varied, however.

1.C The smearing routines in PMCS should be easily tunable. A mechanism
should exist to produce side by side RECO and PMCS output for easy tuning.

2. PMCS Requirements

2.A PMCS should be written in an open enough framework to allow general
usage by D0 collaborators.  While we cannot meet all needs in a finite 
amount of developement time, we are open to specific early requests by 
interested parties.  For example, it is a requirement of the FMC++ group 
to acheive at least 8 events per second and provide generator output 
along with smeared output. 

2.B PMCS should be designed in a way to permit development by many people.
Each developer should have control over his own piece of PMCS, and development
should be allowed to proceed asynchronously.  

2.C PMCS should be flexible to meet the needs of users who might want to 
switch off certain parts of the simulation to increase speed.  This ideally
is controlled dynamically by framework package instantiation and rcp file. 


2.D PMCS should support D0 standard output formats.  Thus, em particles 
should appear in an em particle chunk, etc.  Thus, PMCS can interface to 
existing D0 analysis and display software. 


3. PMCS technical design

3.A PMCS will consist of a number of largely independent framework packages. 
We envision that there will be a single package to process the incoming
MCKineChunk, 

  B. PMCS output will be standard D0 chunks.
  C. PMCS input will be a modified MCKineChunk that contains pointers to 
     particle types