Present: Dan, Hans, Heather, Ian, Joanne, Karl, Leon, Toby, Tracy
HEPvis trip report: (Toby) This was a relatively small (about 20 people) meeting for people working on HEP event displays and analysis frameworks. Among the represented chunks of software were ROOT (René Brun et. al.), JAS (Tony Johnson), WIRED (Joe Perl and Mark Dönszelmann) and HippoDraw (Paul Kunz). The latter looks interesting, but has the defect (from our point of view) of having no interface with ROOT. Unfortunately there was no one specifically representing Gaudi (and naturally no discussion of IDL). Ideally AIDA = Abstract Interfaces for Data Analysis should allow users to switch relatively painlessly from one analysis platform to another, but of the above-mentioned only JAS currently conforms to it.
Core status: (Toby) See the Core status page for a rundown on the move to Gaudi v7, CMT v1r7, etc. We're moved over, but there are still some issues to be resolved.
pdrApp/userAlg: (Toby, Heather) Generally speaking the instigator and primary user of these packages, Steve Ritz, seems pretty pleased. More could be done, however.
TBIOROOT: (Ian) The input side is essentially done, but the output side is still perplexing. The inherited (so to speak) design is perhaps too big on inheritance: the one real workhorse class inherits from 4 others. It is an open question whether the current design should be scrapped, or whether eliminating some features, such as the included histograms, would simplify the problem significantly.
Generic XML geometry status: (Joanne) There is a workshop on XML and detector geometry coming up at CERN on May 22. Joanne will participate via VRVS (assuming she can drag herself out of bed in time); Riccardo plans to make a non-virtual appearance. Most likely each will make a presentation about parts of the GLAST scheme.
Code for identifier dictionaries and identifier conversion is beginning to see the light of day. A first draft of header files for most but not all of the anticipated classes has been written, as has a small part of the implementation.
Balloon: (Heather) A version of ROOTWriter exists which will handle data from the cosmic runs (in old SIU format). There is also a version for the newer data format.
Dan Flath is back in town and is working on ROOT MC classes for the balloon, probably later for other data as well.
Release/Testing: (Karl, Toby) We need some uniform, preferably automatable, procedure for unit testing. Currently-existing test programs in some packages, usually written for debugging, are not always appropriate for this purpose. For example, some depend on a GUI interface and a human to interact with it.
Now that Thomas has gone, we need a unix build czar. Karl is the natural heir; he intends to verify at least that tagged packages build properly. There may be other times, particularly after nontrivial changes to requirements files, when it would be useful to attempt to build from the head, even if not yet tagged.
J. Bogart Last Modified: 06/01/2010 15:45