|
Participants: Steve, Armin, Stefano, Ketevi,
Giorgos, Francesco, Michela, Gianpaolo, Diana,
Alan, Natasha, Ada, Stefania, Gabriella
MuonIdentification
reorganisation (Alan)
Motivation - The work in progress on low energy
Muon Identification (Theodota, Natasha, Monica)
needs to get into the repository release structure. The current 27 classes
in MuonIdentification makes
it a rather large package already.
Possible solution:
After a short presentation of the possible solutions, we define as the most
reasonable the following:
A Flat structure under Reconstruction/MuonIdentification containing the
existing package subdivided
the following subpackages:
muidAlgs (or muidRec)
- Athena reconstruction Algorithms and strictly related to the algorithm
code
containing most of the present
algo part.
softMuAlgs (or softMuRec) - Athena Algorithm
and strictly related to the algorithm code
containing the new algorithmical
part
The naming pattern following the post-fix Algs is in agreement
with Moore and the InnerDetector
While the Rec post-fix is used in other combined reconstruction packages.
Both are reasonable.
muidEvent - Object classes
muidStatistics -
Containing algorithms for truth comparison, ntuple filling
muidUtils -
Classes that are of common use from the algorithm.
This proposal will be circulated, and if there are
no strong opinion against it,
the migration will start in 3-4 weeks from now (after 7.2.0 release is out)
Fake Muons rejection
with Moore (Francesco)
Francesco reports on the work on fake muons performed
in Naples. PowerPointTransparencies
The samples used contain event from DC1 data and initial Layout (prompt
muons).
How to unify this sample?
Initial layout, at present, has no TRT in. There is evidence of systematics
in the pixel.
On the other side DC1 layout does not reproduce fully the present layout
that will
have a different granularity in the ID.
The most reasonable approach seems to be a full study on DC1 layout. Datasets
for prompt
muons have to be produced (?).
CSC digitization tests
(Ketevi)
During all the meeting we often refer to the tests
on CSC digitization performed
by Giorgos, Ketevi, and Jean Francois.
Results of this test can be found here
under (CSC digitization tests)
From these plots you can see that CSC digitization is
OK,
while CSC phi-strip hits are not properly digitized.
In order to have the fix Serguey(?) has to be contacted.
Validation for 7.0.1
We discussed about the ongoing work on validating release
7.0.1.
We decided that a first presentation including the first results on
release validation will be presented at the Atlas software validation team
meeting (monday october 13th).
See web page here
regarding the ongoing work for different subdetectors
and physics channels.
Since at the time I am writing this minutes it is already
tuesday.
I took the freedom to insert a few comments knowing what happened after
that:
The results collected before monday were shown in the
presentation
that you can find here (thanks to Michela
and Monica).
The results shown did not include pp->J/Psi-->mu(6) mu(3)
that are now available here (thanks
to Natasha).
In the phone meeting we pointed out that if we want
to follow up
on a validation effort we should put some efforts on automatic checks.
In the monday's meeting David Rousseau was pointing out that there is
such a kind of effort already. The work involve J. Couchman and it
is supposed on my understanding to produce a few plots on pre-defined
samples of events that will then be linked via a web-interface.
I think we should try to define among us the most interesting plot we would
like to see and try to evaluate with the developers of the automatic tool
the reasonability of them. The next validation meeting will be at the beginning
of november (november 3rd?) by that date I think we should have clear ideas
among us and a first attempt to contact the developer.
I do not know if everybody of you know what I am going
to say.
The physics community is pointing towards having
the release 7.0.1 as a reference version for physics studies,
in the next months several
changes are foreseen in the software and the idea is to use 7.0.1
(or 7.0.2 reopened for taus) for physic study until 8.0.0 will be out.
Test Beam analysis(Stefano)
Stefano reports on a problem with lineFitter for chambers
having s<>0.
For solving the problem the knowledge of the s value is needed. This
value is available in MuonDetPosition, and Ketevi will add the access methods
to
MuonDigitPosition.
The possible choice of accessing the conversion service from DigitToHit for
the
Test Beam data will be inserted in the Moore package.
Persistence (Giorgos and
Alan)
Giorgos (and Alan, after Giorgos has lost the connection)
reports on the work on
persistence.
The main problem comes from the fact that we have problems in persistence
of a
pointer to a vector of pointer (that is part of our track class.) The problem
has been referred to
the POOL and we can now foresee two possible solutions:
1- The POOL (team) solves the problem and we can persistify the object as
it is.
2- We modify the objects to a vector of pointers (this has to be done from
Moore/MuID/ipat).
Geantino map based material
service (Armin)
We agreed on releasing a Geantino-map based material
service for the Moore track
fit. The service is in debug/test status
To run the new service, it has to be ensured that
data files, jobOptions are in the (TestRelease/...) run
directory, and option "MooMakeTracks.material_scenario=3;"
has to be chosen in the jobOptions file.
Milestones (Ada)
For the definition of milestones there will be mail
circulation
in the next days.
|