Core Minutes 31 May 2011

Core Minutes 5/31/2011

Present: Joanne Bogart, Anders Borgland, Jim Chiang, Stefano Ciprini, Richard Dubois, Dario Gasparrini, Heather Kelly, David Landriu, Leon Rochester, Tom Stephens, Tracy Usher

ScienceTools: (Jim) Since last time he added a new class, ScaleFactor, to Likelihood and a new application, lle2drm, to fitsGenApps. See Science Tools Development Notes for details.
FSSC: (Tom S.) P6_V11 data are available; starting to work on Pass7.
GR/Onboard filter: (Tracy) Not much news since last week. Changes needed to use the new FSW release (so far only on Windows) have been committed to CVS. When he gets a chance he will track down the "missing pieces" - shareables containing configuration information - as indicated in Jana's email.
Pass8: (Tracy) Pisa folk have been cleaning up various bugs. Tracy is hoping to finish up with tree-based tracking by Thursday. Alex (Drlica-Wagner) has been working on splitting the n-tuple into two pieces: a user piece and a critical system piece.

The near-term goal continues to be to get out a release suitable for generating datasets which can be analyzed for the Santa Cruz meeting. (Richard) What datasets will be needed? (Tracy) The usual: all gammas with overlays, backgrounds, calibration datasets which Johan usually produces. Nothing huge at this stage. We might need, e.g., a big background run some time after the Santa Cruz meeting as people get into the event analysis phase of Pass8 work. (Richard) wants to avoid conflicts with Pass7 reprocessing, now scheduled to start in August. He will ask Tom to do a test run in July.

(Leon) has been working on incorporating the use of diagnostic data in MC. But the diagnostic information has only been part of real data since 2009. Since it's time-varying, one can view the one bit of information (diagnostic data present or not) as a calibration, but Joanne, perhaps not interested in implementing all the calibration machinery for this, suggested hard-coding it. (Joanne) Time-varying does not a calibration make. This piece of information is not calibration, but configuration (in the sense that the flight software has to be configured a certain way to request the information). Better, one can interrogate the event data directly to see if production of the diagnostic information has been requested. In theory this could vary on a per-event basis, though in practice there have just been two epochs. For MC data I suggested making an interface (e.g., taking event timestamp as argument) for a function returning true or false. To start, just model the two epochs; change the implementation later if something fancier is needed.

(Richard) Speaking of time-dependent information, what about IRF's? (Jim) Has not been an issue so far. (Anders) At some point in the future we will have hardware failures and thus different IRFs. So time dependent IRFs will be useful. Eventually.
Core talk:
- GR, CLHEP 'n more: (Heather) GlastRelease-19-00-00 has been tagged and includes all the updates necessary for the Geant4 9.4.p01 and CLHEP 2.1.0.1 upgrades. We tentatively decided to retain the 19, rather than revert to 18. There will be a new HEAD tag soon to include update requests from Leon (for some Overlay support) and Joanne (GRBAnalysis tags).
  
  Heather plans to use Joanne's modified tagCollector today to get the first GR HEAD build for SCons going. She'll report any woes to Joanne - none are expected!
  
  ST will also be upgraded to CLHEP 2.1.0.1. The necessary updates have made it through the SCons LATEST build and with Jim's blessing will shortly be promoted to HEAD.
- AdfReader (Heather) The AdfReader unit test has been disabled (AdfReader-00-03-04) until further notice. It has been continually getting stuck in an infinite loop in our CMT windows builds. This problem is reproducible - but not immediately fixable. Nicola has been contacted. This package actually is no longer used in GR at all, but continues to be built in the LATEST builds. With Nicola's permission, we may remove this package from the list of LATEST builds.
- ft2util_2 (Heather) A code review is scheduled for next Tuesday at 10 AM if there is a tagged release by Thursday.
- glast-win04 (Heather) GR builds had been dying on glast-win04 because they were looking for a partiuclar, probably unnecessary G4 library. [This has been fixed in a new LATEST, but there are about 10 packages which fail to compile for various reasons.]
  
  (Tom S.) Will shift ST HEAD and tagged builds to glast-win04. ST LATEST is already being built there and doing well.
- GRBAnalysis (Joanne) has been proceeding according to plan. Small changes to SConscript for astro and facilities have been committed and tagged so that these packages will build in the usual way for ScienceTools and GlastRelease, but will build static libraries when part of the soon-to-be-created GRBAnalysis-scons container. GRBAnalysis-scons will have these packages and a dozen or so others from ScienceTools as well as the GRBAnalysis code itself. I also made new tags for a couple irfs packages so that, in the context of GRBAnalysis-scons, dependencies can be minimized. See this Confluence page for plans and status.
- tagCollector.py (Joanne) A new version supports the option --parent=NONE (default is HEAD) for first-time HEAD tags. It's in CVS and, for now, is installed in the $GROUPSCRIPTS directory as tagCollector_new.py.
Announcements
- SCons meeting Tomorrow (Wednesday), 10:30 AM PDT

minutes index

J. Bogart, Last Modified: