Minutes 13 November 2001.Html

Submitted by Anonymous (not verified) on

13 November 2001

cvs version: $Id: minutes_13_November_2001.html,v 1.2 2001/11/19 16:18:42 gunter Exp $
cvs last modification date: $Date: 2001/11/19 16:18:42 $

Present: J.Apostolakis, M.G.Pia, L.Urban, G.Folger, S.Sadilov, G.Cosmo
Video: M.Maire, M.Verderi, M.Asai, D.Wright, T.Sasaki, H.Kurashige, Y.Morita
Phone: S.Tanaka, H.Yoshida

Regular Items

Minutes of previous meeting

The minutes of the previous meeting were approved.

Pending actions

The open actions items were reviewed. See separate list of action items.

Milestones 2001, Part 1

Status of WG development objectives for Geant4 release 4.0

A summary of achieved working group milestones was reviewed.

  • Visualization (S.Tanaka)
    • Dtree functionality implemented, available as ASCII printout, and using GAG; Dtree output to an XML file in preparation.
    • implemented HepRepFile driver to interface visualization with the WIRED event display.
    • Implemented mode accumulating events on display in addition to default refresh mode displaying single events.
    • completed restructuring visualization commands.
    • removed environment variables for building of and use of vis drivers for drivers independent of external packages.
  • Event/Run (M.Asai)
    • methods to store events implemented
    • new random number storing method
    • HepMC is pending, there is hope for an example in the Geant4 4.0 release
    • the modification to the stacking manager allowing the user to increase the number of stacks without modifying Geant4 source code is implemented.
  • Electromagnetic physics - standard (M.Maire)
    • review of multiple scattering mostly finished.
    • subcutoff delta rays for muons implemented.
    • re-analysis and re-design can only be done next year.
    • Integral approach suppressed as result of an evaluation.
    • cut by logical volume implemented by cut by material, see below.
    • store/retrieve of physics tables implemented in all processes.
    • removed production cuts in photoelectric, Compton, gamma conversion, positron annihilation effects
    • L.Urban: evaluation of high energy (> 100TeV) started. Some testing and corrections done, but more to come in January 2002.
  • Interfaces (H.Yoshida)
    • G4cin nothing to report
    • evaluation of JAS
    • Dtree implemented
    • Reduce number of environment variables; remove variables not related to external package
  • Intercoms (H.Yoshida)
    • new commands, i.e. loops, aliasing, counts for repeated commands implemented.
    • G4editors in progress ????
    • (M.Asai) preparation of html documentation files for all commands in progress. See action item 15-1.
    • Question: Can G4UItcsh be extended to CYGWIN.
      Answer: This depends on system header files, and ???kyoshi?? will try to do this possibly using CYGWIN
  • Materials & Generic Processes ( M.Verderi)
    • no major milestone
  • Geometry (G.Cosmo)
    • magnetic field per volume implemented.
    • implemented reference-counted touchables and submitted to STT.
    • first implementation for time dependent field included.
    • added three extended examples for magnetic-field, showing: settings of different steppers, electric-field and time-dependent fields, fields per per volume
    • added tool for checking geometry overlaps based on the BaBar implementation. Implemented kernel commands in a messenger.
    • implemented kernel command for resetting navigator at run-time in the new geometry messenger.
    • implemented reflection of volumes hierarchies. Tested for plain CSG solids. Specific solids and BREPS need more testing. - updated G3toG4 tool to support reflection and a limited set of G3 MANY feature.
    • migration to STL finished
    • Comment: option to reset navigator at run time for exchanging different geometry setups needs changes in run category.
  • Hadronics (J.Apostolakis for J.P.Wellisch)
    • dedicated elastic scattering models for proton and neutron elastic scattering off hydrogen, based on R. Arndt's database, are implemented
    • evaporation parts of the HETC re-write are implemented, the intra-nuclear cascade is not yet available
    • Design iteration and re-implementation of longitudinal string-decay in system testing.
    • chiral invariant phase-space-decay implementation in system testing.
    • Question(M.Asai): What is the status of user physics lists for different use cases? No answer in the absence of J.P.Wellisch
  • Electromagnetic physics - low energy (M.G.Pia), full details available in slides shown
    • new features physics
      • model for fluorescence in continuous part of e/h ionisation
      • PIXE
      • improvements in various processes
    • new designs and several design iterations done; code implemented, tested, and tagged.
    • work on SPI includes user requirements, OOAD, traceability.
    • new advanced examples implemented
    • many new unit tests, unit testing uses AIDA/Anaphe
    • migration to STL, strict ANSI compilation options
    • cuts per material implemented
  • Particles/Track (H.Kurashige)\
    • reduction of Geant4 initialization time implemented
    • save/restore of particle definition in progress
    • dynamic particles implemented, but is there interface in event generator???? (Makoto????)
    • cuts per material implemented
  • Tracking (T.Sasaki)
    • migration to STL done
    • modifications for reference counted touchables implemented
  • Persistency (Y.Morita)
    • fixed several problems related to the latest version of Objectivity (6.1)
    • Objectivity 6.1 will be supported for Geant4 4.0

The presentation analysis category started a discussion on the analysis category. MG. Pia claims the main use of analysis is internal Geant4 testing. In fact many tests in the low energy EM already make use us the AIDA interfaces when available and use LIZARD for histograms and ntuples. However other working groups used other approaches, e.g. hadronics widely uses ASCII output to remain independent of the analysis package used. A clarification of the role, contents of the analysis category was felt missing, see action item 15-5.

Topics affecting several working groups

Cuts per material

The implementation affected several categories, see above. Also affected are examples/extended/electromagnetic; the changes however imply changed method names/signatures, and therefore this intermediate implementation is not explicitly exposed to users.

Question (M.G.Pia): what is the time-scale for the long term solution, i.e., region dependent cuts?
Answer: The discussion of this has started is and is well advanced, this can however only be solved in 2002.

Reference counted touchables

This was implemented mainly by R.Chytracek. The implementation is generic, i.e. it is available as a class template and available from global. Extensive checks for performance were done; in extreme cases we loose about 6%, and in the more realistic, but still very demanding case of the Atlas accordeon the loss of performance is less than 5%.

Release schedule

G.Cosmo reported on the progress of the preparations for the release Geant4 4.0. Tag for categories in the first group have passed system testing, and we were testing tags for the second category now. Testing tag of the third category was to start on Monday 19-November.

The release will be based on CLHEP version 1.7, and Objectivity 6.1. The version of AIDA needs to be clarified, see Action item 15-6.

Documentation review

D.Wright asks to notify him with a short description on commits to the documentation, stating if new features were documented, or if this was in reaction to comments from the reviewers.

D.Wright is concerned about the progress for implementing changes asked for by the documentation review:

  • Physics reference manual: no updates yet
  • Toolkit developers manual: an estimated (10-25)% of the asked changes
  • Applications developers manual: ~25% of the asked changes
  • Installation guide: ~2/3 done
  • Introduction to Geant4: re-write not yet delivered.

In a comment the usefulness of the source code browser LXR was pointed out. See action items 15-9 and 15-10.

Software Process

G.Cosmo gave a short update of the current status:

  • no new design models were committed, except reverse engineered designs from visualization and analysis, provided in kind by the low energy EM working group. This is extremely disappointing given all the efforts made since long to collect design documents for all categories. Gabriele refuses to present further updates on this.
  • source text for the URD v 5.1 is available, but the latest version 6.3 only only exists as postscript.

The responsibility for updating the general URD is to be assigned within the Architecture WG.

The electromagnetic category asked which tool should be used to produce design documents -- answer Rational Rose --, and how they could be helped to access and use this. The need to improve on the access to this tool was acknowledged.

Question(J.Apostolakis): The template for new user requirements was used by two categories during the last workshop. What is the status of this?
Answer: Nothing is in the repository yet. Sergei is working on a tool allowing to better track new requirements from template to implementation.

Milestones 2001, Part 2

Comparison projects. For electromagnetic physics the Atlas comparison projects are converging. Advanced drafts for papers are expected to be available in the near future.

Geant4 applications. New extended examples were developed.

Regression suite. First steps towards this were taken by the low energy electromagnetic working group.

Evolution of documentation. The first part of the documentation review was done, and the implementation of the results is in progress, albeit slow -- see above. The work to improve the training kit has not started; G.Cosmo pointed out that nobody was following up on this. However the training kit was at least used for a school in Torino. M.Asai will be using the training kit, and offered to give feedback from this experience.

Towards the Milestones for 2002

A discussion of milestones for next year was left for the coming TSB meeting.

Report from Event-biasing mini-workshop

The mini-workshop took place at SLAC the last week of September. Use cases for event-biasing have been identified; these are 1) radiation characterization, 2)penetration studies, 3) dosimetry studies, 4) background studies, 5) Accelerator studies. Various techniques to achieve event biasing were identified, i.e. importance biasing, leading particle biasing, cross section biasing, and cuts. A URD is clearly needed, and M.Dressel will start this by filling in the user requirements for geometry.

Examples fro biasing could be build essentially without modifying Geant4 itself; however there may be constraints on this coming from experiments, e.g. LHC experiments will not want to use wrappers for processes for radiation background studies.

Nomination Architecture Subgroup

This subgroup of the TSB has the mandate to study cross category design issues inviting category coordinators to the discussion when these are affected by the proposed design changes. J.Apostolakis nominates M.Asai to lead this subgroup, and appoints H.Kurashige, G.Cosmo, and J.P.Wellisch as members.

AOB

The next Category Coordinators meeting is foreseen for January 2002, to prepare for a TSB meeting in February.

New action items:

Action 15-1 All implementors of messengers to provide extensive HTML documentation for the messenger commands. This will be used by intercoms as interactive help
Action 15-2 G.Cosmo to create novice or geometry example showing the use of G4Assembly including reflecting solids
Action 15-3 J.P.Wellisch to provide summary of features in current and future tags to be used in the documentation of new or modified features in the release
Action 15-4 M.Maire ands M.G.Pia to present use cases of histogramming and analysis in the categories of standard electromagnetic physics (MM), and of low energy electromagnetic physics (MGP) at the category coordinators meeting in January.
Action 15-5 J.Apostolakis to clarify the contents of analysis category
Action 15-6 G.Cosmo to identify version of AIDA used for release Geant4 4.0
Action 15-7 H.Yoshida,S.Tanaka to clarify the support of code written and hitherto supported by G.Barrand
Action 15-8 M.Asai, G.Cosmo to clarify use case of reference counted touchables, and then create an example demonstrating their use.
Action 15-9 F.Jones, S.Sadilov, G.Folger to announce to the collaboration the availability of LXR Geant4 3.2 with the aim to get input from collaborators and to gain experience with bigger loads.
Action 15-10 F.Jones, S.Sadilov, G.Folger to study and propose a setup of LXR for Geant4 4.0. The possibility to make available to collaborators LXR for monthly tags shall be evaluated as well.
Meeting Date