Creator: M. Verderi, (23 Oct
2006)
Last revision: J.
Apostolakis (31 Jan 2006, 17:30 CET)
·
Decision:
o
The
December release is agreed to be a minor release: 8.2.
·
Reported
developments:
The
parallel navigation introduced in the release is proposed not to be the default
behavior for this time, but an option. It is expected to become the default in a
following release. Open issues remain regarding the parallel navigation are related
mainly to the multiple scattering process which makes significant use of the
navigator for tracking, especially to obtain an accurate safety. A new class will
enable it to obtain a valid safety for the post-step point, in which it can
propose a move, taking into account all the geometries.
·
Issues
raised:
·
Physics
lists:
o
Testing
of physics lists provided with the toolkit must be undertaken for release 8.2.
·
Action
item 1 (Physics list task force, examples coordinators): Clarify who has the
responsibility for undertaking and reporting on testing, timing studies for
physics list ‘engines’ provided in the toolkit.
·
Track
at rest:
o
A
problem has been reported with some tracks which come to a rest.
·
Action
item 2 (T. Sasaki, M. Asai): Clarify and document what part of G4 has the responsibility
to kill tracks that come to a rest? This responsibility was moved from tracking
to processes.
·
Interface
changes (collated in preparation for release 8.2):
o
EM:
plan to remove references in physics list builder-components to older versions
of processes (including multiple scattering).
The processes will be kept until a future release.
o
Interface
changes are being prepared in hadronic cross-sections, which are not to be
visible to users.
o
Current
methods to create a ghost geometry construction will be replaced and deprecated
(kept in release 8.2), and will not be avalable in the next release.
·
System
test:
o
The
system test team reports that many tests appear neglected, and suffer from a
lack of revision.
o
The
level of coverage of the system tests was questioned, and the appropriateness
of some tests. An example was given: a problem in a Lorentz transformation
which could not be detected by the current category-level or system integration
tests.
o
Can
coverage of physics be improved by using existing programs ? Consider including,
for example, part of Tatsumi's test suite contribute.
o
Vladimir
Ivanchenko reported that for release 8.1 all EM processes were exercised in EM testing.
·
Action
item 3 (physics coordinators): Verify that all processes are tested in
category-level and in system-integration testing.
It was clear that the effort
to maintain and improve system tests must shared with the relevant working
groups.
o
Action
item 3: Category coordinators are asked to review and revise the system tests
relevant to their areas, assessing whether they are up to date, whether they
test all relevant models and key functionalities.
·
CLHEP:
o
The
difficulties in installing Geant4, arising from the installation of the
underlying CLHEP package was noted. These
installation complications have been reported extensively by many users in diverse
user communities. This could potentially
be solved by including part of CLHEP in Geant4 installations and providing this
as an installation option. Investigating and potentially developing such an
option for the upcoming (December 2006) release was judged not to be possible
due to the short timescale.
o Decision:
The
new SB chair elections will not happen during the release phase.
o About next SB mandate:
A
better communication with Geant4 members and people joining the technical forum
has been identified as one of the goal of the incoming SB.
o SB chair election:
SB will nominate the candidates and the
collaboration membership will elect. A
team with mandate to propose the election rules will be formed. Joseph Perl volunteered
to contribute.
The Geant4 membership is to
be confirmed in time for the start of the election.
- Action
item 4: WG coordinators are to update the current list of contributors in their
group.
An Oversight
Board (OB) meeting took place at the workshop on Thursday, October 11th.
Petteri Nieminen (ESA) chaired. The OB
membership is now available at the OB site . J. Apostolakis reported to the
The main
Makoto suggested that an educational
application HyperNews thread be created. Hajime would be in charge of it.