Present: Joanne Bogart, Toby Burnett, Xin Chen, Johann Cohen-Tanugi, Richard Dubois, Riccardo Giannitrapani, Traudl Hansl-Kozanecka, Heather Kelly, Michael Kuss, Sean Robinson, Leon Rochester, Alex Schlessinger, Tracy Usher, Karl Young
Gleam status: (Toby) As of last night curing nightly build the Gleam checkout package HEAD didn't get checkout out properly; may now be fixed now. Attempt to isolate and automate ROOT-specific build steps has so far been unsuccessful (had a scheme which built everything properly in both environments, but at the cost of breaking cmt recursive checkout on Linux), but Toby has a possible fix which he will try out forthwith.
Release Manager: (Alex) Last week's highest-priority planned improvements have occurred:
more nfs space for saving build products of nightly build and web-accessible afs space for output
nightly building of package heads as was already being done for checkout packages, all results and status accessible from the nightly build page.
See this summary of Release Manager status for more details and the new to-do list.
Random report: (Karl) Changes needed to get all uses of randoms changed over to the Gaudi service are in progress. A search for all such uses turned up only a modest number, so this shouldn't take too long.
System tests: (Richard) See his comprehensive report on this facility, covering motivation, design, and description of prototype implementation for CalDigi single crystal.
How will we do comparisons to a standard? We need a way to introduce some flexibility, such as increasing tolerances on differences when we have reason to believe the new (improved) version might have legitimately crept away from the old standard. And we need a way to update the standard.
Exposure: (Sean) See this description of ongoing work. Given a time, can find out where the LAT is and which way it's pointing, an essential part of the exposure calculation. The information will be generated at regular intervals. Sean will look into implementing Richard's suggestion of defining a special timer event type to generate the entries (rather than initial thought of using a special kind of particle).
Core week: Not far off now, close enough to have a web page of its own. During the meeting we came up with two more items to add to the already-substantial agenda: Sean's work on exposure described above and a joint discussion with fly folk on relation between our stuff and their stuff, their use of our stuff for simulation, etc.
previous | minutes index | next |
J. Bogart Last Modified: 01-Jun-2010 15:48:05 -0700