Present: Alex, Claudia, Heather, Joanne, Karl, Leon, Richard, Sean, Toby, Tracy
The meeting was almost exclusively a walk-through of the new schedule, with extended discussion of several of the projects therein:
Flux services: (Toby) Changes are underway to allow external packages to define sources.
Gleam: (Toby) Just about ready for tagging. Packages it depends on are tagged. Waiting for confirmation that CalRecon and TkrRecon are in shape.
McParticle: (Richard for Riccardo) Riccardo will start work on filling McParticle in G4.
ROOT and McHit: (Heather) ROOT versions of McHit classes nearly done. What to do about conflicting conventions? ROOT uses Get.. Set.. for method names while we use get.. set.. People just have to become sensitized to the distinction between our code and imported packages. Richard remarked tangentially that in most cases we want to stay away from lots of set methods and instead just do initialization in the constructor or in a single initialize method.
Bari digis: (Claudia) Stuck at the moment because G4Generator, though it builds successfully, crashes as soon as it tries to use G4. Karl will try it out at SLAC.
Updating BFEM: (Leon) We need to be able to analyze BFEM data with current software, preferably sooner rather than later, though it's not at this point crucial. This will involve not-huge but non-negligible changes to the way the geometry is specified in the XML files, even for the flight instrument, and the use of a special TkrRecon geometry visitor (maybe to become ultimately a visitor for Recon generally). Leon has the skeleton of such a visitor in operation already. The XML work splits into two pieces, the second of which is BFEM specific. It might be possible to offload this part onto one of the I & T people since they'll need to do similar things for EM, etc. (Joanne) If so they can consult the XML geometry-writing guide, now mostly complete.
Neural Net Track Fitting: (Tracy) Code has been committed to CVS, but doesn't yet work away from home.
CalRecon: (Richard) Awaiting Sasha's return to untangle a problem having to do with identifiers.
Release Manager: (Karl)
Nightly builds have been broken for a couple days. The scripts haven't changed; it could have something to do with changes in the batch environment.
Automatic Doxygen documentation generation is also broken since a new version of Doxygen was installed. The quick fix is to retreat to the older version, but we also need to understand what the problem is.
The mother package approach is now looking somewhat problematic.
Need a strategy for disk usage in release area and documentation area in this era of rapid and large releases. Would help to get rid of redundant objects and static libraries (just keeping .so). Anyone wishing to hang on to a release more than a couple back should inform Karl. In turn he will send out a warning message before deleting things.
glastpack: (Alex) After looking at the LHCb approach to supporting users and developers, Alex has written a wrapper for cmt called glastpack, specialized for GLAST but retaining certain key concepts from LHCb, including the idea of tiers. He has written up some notes on relevant issues, and on how glastpack addresses them. glastpack should be available for trial within a week or so.
newtag: (Karl) This script goes most of the way to automating the tagging of a single package, making it impossible for Joanne to do things in the wrong order or type in wrong version numbers as she has been wont to do. It's still being tweaked but should be available soon.
Calibration infrastructure: (Joanne) Thanks to Alex for bringing up a MySQL server on centaurusa. Joanne did some preliminary set-up (a database called "calib", access to it for Madhup and herself). Madhup is working on making a table (first draft) to hold the calibration metadata. He is also thinking about what the new form of bad hits data should be. When this settles down it will be defined in XML. Hope to get a new calibUtil package set up and even start putting code in it some time next week.
New home page: (Joanne) Having heard no violent complaints, she will install it in Frontpage shortly.
J. Bogart Last Modified: 01-Jun-2010 15:45:06 -0700