Minutes of the Geant4 Technical Steering Board Meeting,
2 November 1999
Attendance:
CERN : John Allison (chairman), Katsuya Amako, John Apostolakis, Guy Barrand,
Gabriele Cosmo, Andrea Dell'Aqua, David Jacobs, Michel Maire, Youhei Morita,
Maria Grazia Pia, Florence Ranjard, Johannes Peter Wellisch.
SLAC: Marc Verderi.
TRIUMF: Peter Gumplinger.
KEK: Makoto Asai, Takashi Sasaki.
By telephone: Satoshi Tanaka, Hajime Yoshida.
Apologies for absence: Gunter Folger, Hisaya Kurashige, Peter Truscott.
minutes: F.Ranjard - 22 November 1999
Spokesperson nomination
Manuel Delfino, IT division leader and CERN representative at the Collaboration
Board informed the TSB of the status of discussions in the CB:
-
The CB has set up a search committee which has returned its first conclusions.
-
There is now an intensive exchange of mails to come to a conclusion which
is taking longer than anticipated because the CB wants to consider the
wider implications. He apologised for this delay but stressed that this
was because the CB wanted to reach a good conclusion.
-
The CB is concerned about the take off of Geant4 in some experiments.
-
The CB encourages the TSB to be open in their activities and build on the
principles of good design, implementation and documentation.
-
Manuel thought that CB members would have been interested in receiving
the minutes of the last meeting: he recommended issuing minutes shortly
after the meeting and he did not see the necessity for "internal minutes".
-
The possibility of a joint CB-TSB meeting has been considered
but it seems difficult to organize because members are world distributed
and very busy.
-
The CB should reach a conclusion in the next few days.
Approval of the minutes of the TSB held on 22-24 September 1999
The latest version was available just before the meeting. A lively discussion
took place on the necessity of the "internal" minutes and on the content
of some parts of the minutes: diverging opinions were expressed.
Finally it was agreed:
-
To approve the minutes with some changes in the STL and Migration to ISO
C++ chapters: new draft should be circulated soon.
-
To ask the chairman, after each point, to make a short resume for the minutes
of the decision taken by the TSB.
Report from Collaboration Board
John Allison reported the discussion and decisions taken at the CB meeting
held on 27 September 1999 and not covered by Manuel Delfino:
-
New members have been accepted: V.Rolando(IGD), A.Solano(IGD), N.Starkov(Lebedev),
IN2P3, Jefferson lab,
a group from Theoretical Physics Institute of the Frankfurt University,
V.Ivantchenko(Common).
-
The TSB will be asked to give a "Activity Report" report to the CB
in 2000.
-
This report should be reviewed by a committee,the composition of
which is under discussion in the CB.
-
The target audience should be "the community at large".
-
This Activity Report should be useful for the collaboration as reports
to LCB were in the past. It should not be too great a load on the TSB.
-
The request will come early next year.
-
A request was made to use Geant4 as a benchmark for processor performance;
the TSB should follow the request.
see below for more details.
-
The TSB was asked its views on licensing and commercial use of Geant4.
-
The TSB should give its views on a publication policy.
-
It was suggested that a Geant4 textbook could be written.
Report from the X-ray Space Telescope Simulations
John Apostolakis gave a status report on the X-ray Space Telescope Simulation
program started at ESTEC during the workshop to assess the vulnerability
of the CCD detector array to damage by incident hadrons.
-
Two geometries (detailed and optimized) were implemented.
-
The first conclusion was that incident protons could damage CCDs.
-
There is not yet any number published by ESA on the subject.
-
It was a very useful exercise for the Geant4 team.
Problem Reports
Bugzilla is the Problem Report system. It was found to be easy to use when
you have already acquired some experience. New comers need a short paper
explaining how to use it. There are too many categories: new, assigned
accepted closed are probably sufficient.
To improve the system the following decisions were taken:
-
A bug report should be acknowledged as soon as possible and
assigned without delay since a new report becomes visible to the world
only when it is assigned.
-
Add a state "RESOLVED/NOTG4" problems (i.e. problems with a new compiler).
-
Add a state "RESOLVED/WORKAROUND" for future re-visiting.
-
Add a link from the Geant4 home page.
-
Add a reminder to the assignee.
-
Bug reports to old versions are accepted, but TSB representatives
should forward enhancement requests to the TSB.
Next Release
Geant4 ISO/ANSI-C++ Migration & RWT types removal
Gabriele Cosmo, as required by the TSB in September,
has proposed a strategy to make this migration. Two options were proposed:
-
ISO/ANSI-C++ migration performed before release starting procedure
together with RWT types acronym substitution.
-
Substitution of RWT types acronym now. ISO/ANSI-C++ migration postponed
to immediately after the release.
After long discussion the second solution was chosen.
It does not require freezing the repository, the work will be made on a
branch. It should be synchronized with the scheduled header-files update
for the Software Reference Manual. It requires that the migration of the
hash dictionary be done before the release.
The following timetable has been approved:
-
Release Geant4 1.0 : 7 December 1999
-
Deadline for committing, tagging : 10 November 1999
-
Migration from Rogue-Wave RWT types to G4RWT : 11 November 1999
-
Deadline for migration from RW hash dictionary: 17 November 1999.
-
ISO/ANSI C++ migration : 7-14 December 1999
-
Update of headers for the Software Reference Manual: to be suspended
during the period of the migration (i.e. from 7 to 15 December 1999.
J.Allison will circulate immediately a prescription to help the migration
from RW hash dictionary .
Content of the release
-
Geometry and transportation (J.Apostolakis)
-
CGS solids: minor fixes to GSG solids.
-
BREPS, STEP:
-
the ability to obtain the parameter string of a solid from a STEP file
and assign it to the solid.
-
An up to date class diagram of BREPS, STEP and STEPinterface categories.
-
Field: a small set of modifications targeting minor fixes, more exposure
to user parameters and elimination of unnecessary warnings.
-
New/modified system tests:
-
A test covering all simple solid types (boolean, new specific types)
-
A test dedicated to integration of complex magnetic fields.
-
Electromagnetic (M.Maire)
-
standard:
-
muls: fine tuning and bug fixes for low energy e-. Inclusion of nuclear
size effect for high energy (e>100Mev, h>1Gev)
-
IonisationPlus: production version for electrons and hadrons.
-
Eloss fluctuation: use G4Poisson instead of CLHEP.
-
PAI model: bug fixes.
-
lowEnergy:
-
hadron ionisation implemented.
-
ion ionisation implemented.
-
minor bug fix in photo electric.
-
X-rays:
-
test:
-
TestEm6 for low energy hadron.
-
All tests moved to examples/extended/electomagnetic.
-
hadronic (H-P.Wellisch)
-
design iteration on multi-fragmentation.
-
isotop-production models for neutron induced production and corresponding
design changes in hadronic/management.
-
additional string model for photon induced reactions.
-
u.u. memory usage optimization in the neutron transport.
-
particle and track (H.Kurashige)
-
particles/management
-
DynamicParticle: Dynamical electric charge (together with corresponding
changes in ParticleChange)
-
IsotopeTable: New table introduced to mange excitation energy, lifetime
and decay modes for isotopes.
-
shortlived/hadrons
-
class/method name: All words of 'barion' in class/method names are renamed
to 'baryon'.
-
persistency (Y.Morita)
-
new example of hits with association from the event object.
-
setup script for objectivity 5.2 and HepODBMS.
-
support boolean solid.
-
Visualization (S.Tanaka)
-
completed version of the new visualization commands.
-
visualization drivers which co-work with the new visualization commands.
-
final version of example macros for visualization in examples/novice/N03.
-
visualization of the components of boolean operations.
-
Documentation (K.Amako)
-
define the header file format for the automatic generation of the software
reference manual.
-
write a perl script to handle the automatic generation.
-
encourage/push category coordinators to change header files.
On which platforms
ATLAS requires that the release is compiled also on HP with the new
compiler. ATLAS will provide access to the ATLAS HP cluster to the Geant4
team.
Following platforms and compilers are accepted:
-
SUN / cc 4.2
-
HP / aCC A.01.21 and aCC A.01.18
-
DEC / cxx 6.1
-
Linux / redhat5 or redhat6
-
NT / VC++ 6.0
Forthcoming conferences (CHEP2000)
John Apostolakis suggested submitting a summary paper on Geant4. The proposition
was accepted.
Publication policy
-
Everybody should be encouraged to write papers about Geant4: take CHEP2000
as a good opportunity.
-
Find the best forum to present the physics part of Geant4.
-
People writing papers about Geant4 should inform the TSB.
-
It is felt that a general reference paper about Geant4 is needed : the
point will be reviewed at the next TSB.
Geant as a Benchmark
INTEL has proposed using Geant4 as a benchmark to test their new chip.
CB has agreed as long as it is done in liaison with the TSB. S.Jarp is
the link man. TSB has agreed to go along.
Commercial exploitation of Geant4 and code license
-
Licensing problem was discussed: what to do to protect code against possible
usage for military applications. It was felt that help is needed from licensing
experts: H-P.Wellisch will follow the question with the CERN licensing
office.
-
G.Cosmo will remove "RD44" from the code for the next release.
Next meeting
it was suggested having it just after CHEP2000.
A.O.B.
-
P.Gumplinger
-
S.Giani has resigned from the Geant4 team to concentrate to the CMS Software.
-
J.Apostolakis
-
CDF would like to use part of Geant4 in their reconstruction program.
-
M.Maire
-
There is one volunteer in CMS to write a package "a la" GEANE in
the Geant4 framework. The idea should be formalized.
-
P.Gumplinger
-
Find a way to help bug tracing : the present solution is not optimal. Peter
will prepare a proposal.
-
It is suggested that TSB experiment representatives are asked every TSB
meeting whether their experiment has any request to formulate.
-
next workshop: Guy Barrand (IN2P3) has offered to organize it in the Paris
area.