Present: K Amako, J. Apostolakis (chair), M Asai (note-taker), G Cosmo, G Folger, I Gudowska, A Heikinen, V Ivantchenko, M Maire, M G Pia, G Santin (deputizing P Nieminen), T Sasaki, P Truscott (by phone), JP Wellisch.
Main location: INS Laboratory, Catania, Italy
Notes version 1.0: J Apostolakis, 12th October 2004
Revisions: 1.1: M Asai, 12th October 2004
1.2: J. Apostolakis, 14th October 2004
Agreed to drop gcc 2.95.x, VC6 in release 7.0, in order to use newer features of Standard C++ and avoid conflicts with newer compilers.
Revised design for long-lived particles
Regarding the proposal to change the design of particles, and migration:
- The first prototype of revised particles, removing static and changing long-lived particle classes from singletons to helper classes. This design would require an assessment of its impact. Hans-Peter agreed to assess for hadronics.
Action (JP Wellisch, others?): Report on impact analysis for first prototype design of particles that eliminates classes for each particle type.
- A second, revised, design was proposed during parallel sessions of the workshop, which only removes static class members and keeps all singleton particle classes. It was agreed to pursue this option as a working prototype.
Action (M Asai, H Kurashige?): Create second prototype of particles, keeping classes for particle types.
- A move to this was postponed until after release 7.0, because the revised prototype implementation will not be available in time for the relevant part of the release sequence.
Decision of which design to choose awaits the results of the impact estimation and new prototype . Change is foreseen to occur after release 7.0.
Migration to cmath and sstring
Migration to use of <cmath> was agreed, replacing <math.h> for the kernel. This leads also to recommend to users to utilize <cmath> also.
Action (category coordinators, for 7.0 release): migrate code to use <cmath> in place of <math.h>
It was agreed to study the effort required to migrate to use of <sstring> and the impact on user code. This migration was seen as less urgent. Migration to <sstring> will be decided after evaluation
- sstring will likely impact global/G4cout, UI category, and all messengers including users' ones.
Action (UI, global, other category coordinators, Jan 2005?): evaluate impact of migration to <sstring>.
Migration to CLHEP 1.9 or 2.0
Issue: the impact of a migration to CLHEP 2.0 for the 7.0 release 7.0 and related items.
- this is weakly coupled with need for migration to cmath, sstring
- key concern is the use of namespaces for units.
Code in source (“cc”) files could be migrated with the "using" directive, at least for an interim period. It is not clear what the extent is of changes needed in header files (“hh” and “icc” files), and on the best method to utilize for these. It was agreed that the decision will be taken by Category Coordinators, after an evaluation by Gabriele. Note that this is urgent, and must come in time such that changes are compatible with the release sequence.
Action (Gabriele, 22nd October): provide first impact analysis
Action (Category Coordinators, 7th November): provide agreed proposal/decision on move to CLHEP 2.0 for release 7.0.
Renaming of methods of G4ParticleChange
It was agreed that renaming will become mandatory in release 7.0. Advanced users, who have created their own processes, will be offered assistance including the renaming script we utilize.
Action (Soft management, end-October ) : Check which current code uses the old signatures.
Action (Categories with processes, for start of release sequence): Revise affected classes.
A reorganization of directories and a move of the Geant4 CVS server to a new repository were discussed in a workshop parallel session, and presented for first comments at the final ‘parallel summary session’. The TSB asked Gunter to distribute the proposal to Category Coordinators for impact analysis, discussion and creation of a (and then to TSB)
Action: G Folger to distribute CVS reorganization proposal. Target date: end-October.
Summary of agreed changes / migrations for Release 7.0
- Particle Change method names
Changes under consideration for Release 7.0
- units (CLHEP 2.0),
- cvs reorganization, including directory reorganization
G.F. to distribute the proposal to CC for discussion and preparation (and then to TSB)
The TSB welcomed the offer of CENBG to host the next Geant4 Workshop in September 2005 (likely in first half of the month) near Bordeaux. Members were requested to check for potential conflicts of dates with other conferences.
Action (all): Inform M. Maire of potential date conflicts.
- Agreed to create a working definition of a contribution for the purpose of identifying authors and copyright holders.
- Agreed to continue actions to identify copyright holders (current and past)
Action(John Apostolakis): send out an email to CC requesting first draft list of contributors in each category.
It is important to include all authors and copyright holders in the discussion as soon as possible. The need to identify the criteria for a contribution, in order to recognize contributors is urgent. It is highly desirable to identify authors in Geant4 7.0, and move towards agreeing a license in time for this release.
Noted that the TSB’s key current role for the license is to enable and provide focus for the discussion of the license by contributors and license holders.
Forming two new working groups, focusing on the novice/extended examples and the advanced examples respectively.
- The TSB requested the proposing persons to forward the draft mandates and other information on the candidate working groups to the TSB, in order to consider this at next TSB meeting.
Action (Makoto, Maria Grazia, by November 1st): Distribute draft mandate & details for example Working Groups to TSB.
New collaboration members from INFN (2) and IN2P3 were proposed. Provisional acceptance was agreed, pending confirmation – expected in absence of issues/questions to be submitted within the next two weeks.