Present:

CERN: John Apostolakis (JA, chairman), Gabriele Cosmo (GC), Michel Maire (MM), Maria Grazia Pia (MGP), Maya Stavrianakou (MS), Hans-Peter Wellisch (HPW), Gunter Folger (GF), JohnHarvey (JH), Florence Ranjard (FR), Sergey Sadilov (SS), Nikolai Starkov (NS), Vladimir Ivantchenko (VI), Andrea Dell' Acqua (ADA)

KEK: Katsuya Amako (KA), Takashi Sasaki (TS)

SLAC: Makoto Asai (MA), Dennis Wright (DW), Joseph Perl (JP)

TRIUMF: Peter Gumplinger (PG)

Ecole Polytechnique: Marc Verderi (MV)

Excuses:

Aatos Heikkinen,

Minutes: F.Ranjard - v1.0 30 September 2005

Revisions: J. Apostolakis

Initial items

• Approval of minutes previous TSB
  No remark. Minutes are approved.
  
• News from other meetings
  • No CB meeting since the last TSB meeting.
  • The Collaboration Agreement has been sent for signature in July: SLAC has signed it. CERN is ready to sign.
  • HPW notified participation in two meetings: GC requested notification of talks on behalf of Geant4 at meetings with a mail to TSB.
    
    

Issues from Technical Forum held on 27.9.05

• update of shower parametrization in ATLAS presented: request to discuss potential for further performance improvement (is being followed up by JA).
• requirement from BaBar to keep system testing running on Solaris (SLAC to assist with 20% of a FTE)
• requirement from LHCb to be able to use a different version of HepMC with a release of Geant4 : accepted - proposed solution to move classes using HepMC to an example.
• still a requirement from LHCb: limit the amount of printout to the minimum in production.
  

Development and Next Release

• report on planned migrations
  
  • sstream: It can be done in parallel (GC), Category Coordinators can commit. No longer need to compile under gcc 2.95, but should compile on WindowsXP, Visual C++ 7.1 .
    
  • Migration to CLHEP 2.0.x
    • agreed to proceed with evaluation of migration to CLHEP 2.0.2.1;
    • a requirement is that toolkit and user code must work also with CLHEP 1.9.2.1;
    • migration done in such a way that users are not forced to use namespace.
      
    • CLHEP now provide a new header file that enables its users to avoid namespaces for units
      • Choice rests with Geant4 whether to protect the global namespace or to continue to put units in global namespace. ( This issue is most relevant in case of header files with many units. )
    • Accepted options:
    • Preferred solution: allow only source files to have "using" statements, but not include files
    • Agreed/accepted solution: allow each unit that is needed in header file to be defined for this file ( using CLHEP::units::mm; using CLHEP::units::GeV; )
    • Potential interim solution: allow full 'using CLHEP::units' in header file
    • Choice will remain whether to put units in global namespace in case of header files with many units
    • Evaluate the work needed for the migration.
    • Action (Category coordinators): evaluate first report and migration strategy and provide feedback on issues affecting move during the Geant4 8.0 release cycle.
      

• Visualisation code migration proposal
  some work is required from other categories but users could also be concerned due to the changes in G4Trajectory (draw method)

• Progress with expected development
  • Particle table (no longer static particle definition)  : Users need some documentation to use it correctly. Physics lists code use data members which prevent any change by users.
    
  • Run, Event and Detector response: merge of "envelope " and "region" concept
  • Geometry and Transport: Overlap detection on construction (placement of volumes) , work is ongoing , prototype planned for October reference tag.
  • EM physics : energy deposit dependence on step size and cuts is significant for particular materials and setups (HEP: LIC, LHCb reports; medical). The origin was traced to the correlation of angular and lateral distributions in multiple scattering and the choice of step limitation in electron multiple scattering.  To address this issue a revised implementation will be made available in the October reference tag.
    • the old behavior will still be available.
    • Users must be made aware of the changes in physics results and CPU time: results will be independent of step size and cuts (for broader ranges of their value) but at a particular size or cut value CPU time will increase (in several cases, but not all).
  • User Interface: a prototype Python interface is nearly ready, and will be reported at the Bordeaux workshop.
    
  • Documentation
    • Reorganisation of manuals is ongoing
  • Advanced Examples
    
    • new examples on the way.
      
      

Collaboration web

• Regarding the move to the new Geant4 Web site scheme , Dennis and Gunter will collaborate to implement the change. A first step is to identify how to redirect key old pages to new locations.
  • Action: GF will prepare recommendation whether/how to redirect old pages to new site.
• Once the redirection issue is resolved (but no later than the week after the Bordeaux meeting)
  • Action: DW to announce to the collaboration the move to new web site design, giving a 2 week period for remarks and suggestions.
• GF will make the switch of top page and site, once urgent issues, which have been identified, are addressed.
• A new policy was agreed that important information should be available on at least one of the Collaboration web sites and not only from any private site(s).
• The possibility to providing mirrors of central collaboration site(s), with copies in different continents or geographic regions was raised.
• TSB expressed concerns about:
  • private web pages distributing key issues and about geant4.com site which gives the appearance of a commercial site associated with the Geant4 site - while in fact it is not endorsed or approved by the collaboration.
  • category collaborator lists  not maintained up to date.
    

Regular items

• CHEP 2006:  proposal of abstracts should be circulated to the TSB.

Collaboration Agreement

• signed by SLAC
• CERN is preparing to sign.

AOB