Minutes of the Geant4 Technical Steering Board Meeting

29 March 2000

Attendance:

Minutes: F.Ranjard - 19 April 2000

Preliminaries

Comparison projects

The results of each project will be published jointly by the experiment/partner and the Geant4 collaboration.

Project proposals were discussed and clarification requested, as needed.   The CB will meet on April 12th to reach a formal conclusion on the proposals.

Four proposals were presented:
 

•  ESA:  Comparison of  low energy proton scattering with TRIM and  real data.
• as the project is not funded yet, the time scale will be defined by ESA,
• results could be published jointly with the Geant4 team.
The comparison  of low energy proton scattering with data is accepted as a proposal.  It requires a better
understanding of the data in order to proceed.  The comparison with TRIM is left to ESA.
More details on time line and man power are requested.  The second proposed project is withdrawn.
 
• BaBar:  Use Dimuon data from drift chambers and the Silicon Vertex Tracker to study  delta rays distribution, dE/dX in thin materials.
• BaBar manpower will be available for detector modeling, providing the comparisons and aiding in interpretation of possible discrepancies.
• Geant4 consultancy manpower will be required :
• to provide expertise in electromagnetic processes,
• to evaluate and, if needed, improve the precision of electromagnetic processes.
• time scale: the project will last 6 months starting in May 2000
• manpower is not completely defined.
The proposal was accepted by the TSB, pending the definition of the BaBar manpower.  Possible comparisons with Geant3 is left to BaBar.
 
• ATLAS: Test  Beam data taken in the e.m. barrel, hadronic end cap with forward calorimeter and hadronic barrel will be used to study e.m. and hadronic showers. The work has already started on some of the detectors.
• preliminary estimates of ATLAS manpower available:  4.5 FTE  on the e.m.calorimeter;  0.6 to 1.0 FTE on hadronic endcap; to be defined for hadronic barrel.
The timescale of the proposed project(s) was not made clear in the proposal and remains to be defined.
Other refinements needed include when the data was or will be taken,  and the type and energy range of primaries.
 
• CMS:  Comparison with data taken in the Hadron Calorimeter test beam setup.
• CMS manpower: one person involved in the analyses of the test beam data.
• Geant4  collaboration shall provide people with the following skills:
• with good knowledge of the G4 hadronic processes,
• to help in refining the geometry description
• setting up the framework needed to run and interpret the simulation.
The project needs some refinements: more precise setup, time line and manpower.

Training milestone

• Geant4 users community is heterogeneous, needs are very different.
• a very modular approach is necessary.
• several courses have already been done:
• academic style of series of lectures: 3-5 hours
• short-hand on : 1 day
• material exists, training sessions could be organized quickly after some clean up.
• training using interactivity is very useful.
• self teaching material is not the first priority.
• improvement of the navigation throughout the user's guide is welcomed.
• A first FAQ web page has been created.  Developers are requested to submit further items.   The submission procedure should be developed - for now please email submissions to Gunter Folger.

• A central repository for training materials will be created.
• It was decided to focus resources on the following items as a priority:

• short course of 1 day
• course suitable for a School,  about 5 one hour lectures plus hands-on exercises.
• academic style of series of lectures: 3-5 hours.

Software process milestone

• Q/A & optimization activity

filtering of the source code:
• to identify  coding rules violations,  metrics violations,
• to perform monitoring  and profiling of CPU,  memory usage,  memory leaks
actions:
• create a team of 2 persons for ~ 30% of their time for set up and implementing the necessary tools. Get help from external groups,  make  available the results of the filtering,  delegate to Category Coordinators the responsibility  of implementing  fixes.
• Analysis & Design software cycle

Make sure that improvements and new developments are kept in sync with the overall OOA&D.
actions:
• Category Coordinators should review and update if necessary A&D documents,  User Requirements Document,
• These documents should be available on the Web.
• Testing

The goal is to improve the System Testing activity:
actions:
• prioritize recruitment of new manpower and promote a training activity for new-comers.
• improve tests and examples: increase their scope, use statistical tests, establish clear responsibility for maintenance and integration of system tests and examples.
• improve/implement automation.

It is important to focus on areas where improvement is most needed.

The documentation of the current Software Process should be accomplished, in order to drive the Improvement.

Open discussion forum

Requirements for a user forum include the ability to structure information, to view it using threads to follow discussions and to allow moderation.   Several systems exist which allow moderation and structure,  including Hypernews and Monarch.

• There was no consensus on "moderation"
• a first list of possible Forums has been established:

installation, detector description, hits, e.m. physics, hadronic physics, environment, fast simulation and user requirements .
• Gunter Folger will collect information and augment the FAQ web page.
• If Hypernews is chosen,  SLAC will be asked to host it and for the resources to implement it for Geant4.
• time line: ~ one month.

Interaction of proposed working group milestones

There were no comments on the interaction of the milestones of the working groups.

Schedules for next releases

Two major releases are scheduled: Friday 30 June,  Friday 8 December.

New members

• ESA  : add Ramon Nartallo as Geant4 member.  Accepted.   The working group will be specified.
• IN2P3 : add  Wojciech Wojcik from Lyon Computing Center in the STT group.
• IGD : remove 2 members and add 8 ( G. Cabras, A. De Angelis, B. De Lotto, G. Fedel, R. GianniTrapani, F. Longo, F. Marchetto, F. Longo) in the low energy e.m. group.
• HARP : a proposal from the collaboration was previewed.  It was received with enthusiasm.

Policy for reference tags

• low level categories must commit their code 10 days before the monthly reference tag date,
• high level categories must commit their code 7 days before the monthly reference tag date.

Regular items: reports from user groups

• BaBar

William Lockman is the  SLAC representative.
• BaBar has developed an alternative, experimental transportation process because they were unable to satisfy their accuracy requirements using G4transportation.
• It has not required any significant modification to any other part of the G4 toolkit,
• Initial standalone G4 transportation and BaBar extension benchmarks have been communicated to the geometry group coordinator,
• Further benchmark studies will be carried out in the simulation program and communicated to the geometry group.
• BaBar made a proposal for Floating-point exception handling.

• LHCb

is developing a project in Moscow:
• Project:
• simulate test beam setup consisting of 3x3 and 5x5 cell calorimeter Shachlik modules
• Goals:
• get experience with Geant4, understand its philosophy
• produce correct geometry,
• find what representation is efficient, this will provide valuable feedback on how to implement the description of the detector in GAUDI.
• need help to identify correct approach to designing geometry.
• check how well it describes the behavior of the detector
• energy resolution and spatial resolution: comparison with test beam data and Geant3.
• measure performance of simulation and compare with G3. Iterate with studies of physics by changing cuts to reproduce physics whilst optimizing the performance.
• Resources: ~ 1 FTE
• Timescale: 6-9 months, starting in May.

Next meeting