Present:

CERN: John Apostolakis (JA, chairman), Gabriele Cosmo (GC), Aatos Heikinnen (AH), Michel Maire (MM), Maria Grazia Pia (MGP), Maya Stavrianakou (MS), Hans-Peter Wellisch (HPW).

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

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

TRIUMF: Peter Gumplinger (PG)

Excused:

Gunter Folger, Florence Ranjard, Sergey Sadilov, Hajime Yoshida

Minutes: G.Cosmo - 7 July 2005

Initial items

• Maya Stavrianakou is welcome as new CMS representative, replacing Tommaso Boccali.
• Approval of minutes previous TSB
  Since the minutes were announced very late there are 4 more working days for TSB members to read them and comment, before final approval. It was agreed that draft minutes should provided within two weeks of the meeting and agendas a minimum of 10 days before the meeting.
  (Transitory note:) Correction in June 2nd AOB section: should refer to two publications which are in preparation, based on the existing examples for distributed computing environment and proton-therapy.
• News from other meetings
  • Geant4 Japanese Space User's Meeting: a report by Tasumi Koi has been circulated to the TSB about the meeting held in KEK at the end of June, with attendance from the Japan Aerospace Exploration Agency.
    HPW commented that implementation of ion-spallation has been developed and is ready for use. He posed a question on how to support the community interested in its use and suggested organising a mailing-list as a start. T.Sasaki will follow up about potential users.
  • A meeting on astrophysics activities particularly focussed on neutron-background simulation for underground physics will be held in Germany, HPW will attend. The meeting will be in German.
  • Validation workshop in Genova, the 13-20th July: attendance expected is of the order of 20-30 people.

Geant4 Collaboration Agreement

The result of the vote within the Geant4 Collaboration is the following:

• 63 active developers voting in time: 61 yes, 2 abstentions.
• 2 additional (late) developers and 5 Institute representatives: all 7 in favour.

Technical and user-relation issues

• Data libraries
  The recent issues about acknowledgement, licensing and use-conditions of the data libraries were discussed. A necessary first step for the aknowledgement is to know and document exactly which are the data sources used.
  MGP noted that there have been cases where data has been distributed privately, something that should not happen.

  HPW informed of the citation guidelines for the ENSDF data (used for the source-evaporation and radioactive decay collections), which requires specifying date and time when they have been downloaded. The solutions envisaged were checking for the log files of the original downloading or using the date of committal in the Geant4 repository and extrapolating.

  A first list of the data sources is already available in the general paper published in 2003; this list should be maintained and eventually completed. An action item (34.2) is already in place to be considered by Category Coordinators and concerning the aknowledgements to data providers. HPW noticed that the data providers should be asked for the citation guidelines, in order for us to make proper citation of the data used. A README for the G4NDL data set is in place and should be made more visible.

  We should move quickly to provide to the users the complete list of data used, with the proper aknowledgment drafted by eventually adopting the citation guidelines.

  HPW says that G4NDL cannot be distributed anymore to the public as it is; a change in the US copyrights since end of last year puts new restrictions in the licence, conditions and usage rights; according to HPW, restrictions apply on the redistributions of data (NFB, FENDL, ...) and a licence for G4NDL must be provided such that it respects all the involved data sources.
  It is still not clear though what the restrictions/conditions are, since also the distribution is made from the Geant4 site in Switzerland.
  MGP noticed that redistribution problems were easily solved in the past in the low-energy working-group with the data providers in Livermore, by clarifying the conditions and non-profit usage of the data. Since Geant4 developers are in a good relationship with the Oak Ridge data distribution center, an enquiry can be easily made to clarify the whole process. Since many of the data libraries are distributed by multiple sources, it is surprising why they don't have similar problems; different conditions of use ?
  Issues which may raise by the use of the data should not be handled privately but should be made known to the TSB and handled in the TSB. Licencing issues and data-update should not be mixed; they should be addressed separately and fully understood. Building a small task-force to investigate this issue which affects Geant4 and its users is necessary at this point.

  Action item: JA to establish a task-force for investigating the data issue.
  Action item: JA to contact CERN responsibles on potential distribution issues from the CERN web site. For the legal matters the CB should be involved.

• Open problem reports
  Several problems have been left in NEW state for a long time. Several had to do with reports on the hadronics. Sometimes bugs requiring a trivial fix have been left open for long time with no answer and fixes were not provided even in public releases. It was noted at the meeting that problems which were NEW in Bugzilla for hadronics have recently been addressed and those currently open are now set to ASSIGNED.

  Action item: HPW to provide patches including fixes to the problem reports which required a fix and submit them to STT.

• Hypernews and user queries
  It was noted the absence of support for responses in the Hypernews hadronics forum. HPW mentions that Geant4 has been bleeding expertise in the hadronics in the past months, such that now public support cannot be provided in the same equal manner as it is for collaborating Institutes.
  AH expressed concerns about the way the hadronic group handles responsibilities. More transparency in the working-group is necessary, both in terms of the development activity brought ahead in the group and the handling of the data libraries. AH also noted that the Elastic part of the Bertini code is not used at present, the code however is ready though requiring some testing; it may help in solving some of the reported issues.
  DW claims on the other side that there are enough resources to bring the Hypernews hadronic forum up-to-date; some collective effort is necessary from each collaborator in the hadronic group to look at them.
  It was also noted the importance of being able to transfer responsibilities and expertise among the developers, something that helps a lot in the turn-over of people inside a working-group.
  It was agreed that the responsibility to monitor Hypernews and ensure answers in a reasonable time is mandatory for each Category Coordinator. MM noted that in some cases users don't know which Forum to apply the query to and they end up choosing the wrong forum; GC replied that it is also role of a Category Coordinator (but also any developer who follow the evolution of a forum) to point to the right responsible (coordinator or developer) to ensure that an answer is given to the query related to his category.
• WG representation at the TSB meetings
  The presence of particularly the working-groups Coordinators at the TSB meeting is mandatory. It was noted the importance of having the agenda of the meeting well beforehand in order to first discuss relevant items in the WG if necessary and have effective discussion at the meeting.
  The representatives must anyhow be present, or at least ensure that a deputy (also if not part of the TSB) is present at the meeting. Same applies for the experiment representatives. This rule will now be enforced.
  JP mentioned that attending a meeting can be difficult for the quality of the communication; as previously proposed in the past, he is suggesting to move to VRVS which would improve the ability of people to participate.

  Action item: JA to investigate usage of VRVS for meetings.

• Muon physics bug in hadronic physics-lists
  There's an issue arisen from Hypernews; it concerns the ordering of some of the muon processes in some of the hadronic physics-lists which is apparently buggy; it is affecting the AlongStepDoIt() (missing registration of an AlongStepDoIt() for the pair-production, so that it is never called). Effects on lateral distributions were verified, but no real noticeable effect for calorimetry in general. HPW suspects the problem is in the builders, a fix is available and provided to the user for response and verification. The fix is not yet in the repository, it will be introduced soon (next week), so it can be verified also by LHC experiments. It was noted that the same bug is also present for the electrons, but no real effect was observed, the fix will be anyhow introduced also for electrons.
  GC noted that also effects to performance at run-time must be verified.
  AH pointed out that it is good to add a test that verifies the fix and shows what are the differences.
• Validation
  Validation of the physics is an issue, it should be done transparently specifying the data used for validation and the physics-list. Papers on journals should be published as soon as possible to be used as reference.
  HPW noted that the physics-lists are 'educated-guess' because no guarantees can be given and should be considered as such.
  MS mentioned that a system is being put in place in CMS for validating every release of the software with physics quantitites. Something similar exists also in ATLAS.
  MM noted that first-level validation for hadronics and electromagnetics exists (at level of simple example with macros dedicated to specific physics tests), however, there are difficulties if the hadronic physics-lists machinery is used. The advanced-examples working-group is focussing on validation as well and can be very useful to complement the validation made by other groups or experiments.
  AH noted that information about the physics and speed performance of each model is important and valuable and should be made available.

  Action item: HPW and MM to assure that the physics-lists and their constituent parts that are distributed rely on the components that are provided and tested by the providers of the underlying physics components in a one-to-one correspondance.

• Updates to physics-lists
  It was expressed concern about the fact that updates and fixes already available should be provided in reasonable time as development goes on.
  The new particles design requires physics-lists are changed to remove the processes as data-members of the physics-lists; they should be defined in the ConstructProcesses() method. It was agreed that the hadronic physics lists will be corrected as soon as possible to allow for the new implementation of particles to be submitted to testing. It was also noted that the kernel should provide a way to gracefully detect unproper initialisation of processes, in order to avoid similar mistakes from the users side. Proper documentation should be provided to inform users about this initialisation issue.
  HPW mentioned the case where static-data invoke for a particle-definition; will this be a problem ? To verify.
• Configuration of physics-lists
  If using the hadronic physics-lists some of the specific parameters (for configuring some cuts for example) cannot be used. The requirement is to be able to configure EM and HD physics-lists. There's no transparency on the way how to configure the physics. MS noted that there should be additional way to configure the physics other than through the UI only.
  It would be very useful to have a task force which takes care in improving the transparency in the way physics is configured; physics coordinators should be part of this. The problem should be first of all to clearly understand and verify if a straightfoward software solution for properly tuning EM and HD physics together is possible. Moreover, the documentation in this area should be clear in the user's guides.
  It was also noted that configuration issues exist and have been reported since long time. These should be solved! It is not currently possible to use shared-libraries of the physics-lists if granular libraries of the kernel are in use, nor it is possible to use DLLs on Windows with the physics-lists; moreover the usage of EXTRALIBS should be reserved to the users. It should be considered to have the physics-lists (or just the builders?) properly packaged and included in the kernel.

  Action item: to physics coordinators to add chapter documenting design and usage of the physics-lists in the user's guide.

Regular items

• Membership
  • C.Jarkskog membership endorsed by TSB.
  • Provisional OK for proposed new run&event members: expected contributions to scoring and tallying noted. Period for final questions is one week, before acceptance is provided.
  • Provisional OK for EM low-energy members in INFN-Legnaro. Expected contributions to validation noted. Period for final questions is two weeks, before acceptance is provided.
  Action item: JA to check if a formal matter arises regarding new institution.
• Reports from experiments and institutes
  No reports.
• Pending actions
  
  • 34.3: a problem with firewalls was identified affecting several space Institutions. Investigating to trying to fix it
  • 34.5: JA will make formal announcement to geant4-announce.
• Issues from Technical Forum
  HPW informs of progress about the issue on the data raised at the last Technical Forum: found people who will provide the missing data (only Hg is missing for now). They will be soon made available.
  A potential date for the next Technical Forum is July 19^th

Geant4 Workshop

JA will make formal announcement to geant4-announce in the next days.
Collaborators are encouraged to start registering for the Workshop !
It is also time to start thinking about the program of the Workshop, since this will also help in attracting people to the registration. A preliminary discussion will be started at the workshop in Genova in 10 days.

AOB

• Posting errors in G4Hypernews: it is suggested to report to the web master for errors which occur on posting.
• A potential Category Coordinators or a wrap-up meeting will be set up to address remaining points on the agenda which were not discussed at this meeting.