Present: Joanne Bogart, Toby Burnett, Sasha Chekhtman, Xin Chen, Richard Dubois, Alex Schlessinger, Tracy Usher, Karl Young
Gleam overview: (Toby) See his Gleam notes (new features, to-do) concerning unreleased tag v2r1p5.
We now have a meritAlg for Gleam, producing some of the standard output needed to evaluate performance. We are not that far from functionality required for a Release.
Tangled Webs We Have Woven (aka TkrRecon propagator issue): With several streams of interdependent but uncoordinated development it is perhaps not surprising that we found ourselves in a situation where nothing quite worked. Ingredients in the mix included
The G4 propagator had a couple deficiencies: incorrect calculation of radiation length (since fixed) and lack of service to find nearest edge (on the to-do list). On the other hand, the Gismo propagator hadn't kept up with all the geometry-related changes.
At this time problems have been or are on their way to being solved; Tracy and Toby will get together to iron out any remaining glitches.
Still more Gleam items:
Classes ending with "Col" should satisfy certain interface requirements so they behave like STL containers.( (Toby) The specification for std::vector is a good place to start.) The Gleam release should not be made until this is the case. The process of repair would be aided by a written description of precisely what methods, typedefs, etc., are required.
ROOT Trigger Output: Doesn't exist yet; should go in the event header.
How do we test a new release adequately? New features sometimes only need initial testing; other diagnostics should be run routinely for every release.
(Re-)introduce ability to coordinate multiple passes of subsystem recons.
IDT Face-to-face: Presenters will meet this afternoon (Wednesday) to make sure the first 3 talks dovetail properly:
Action ItemsJ. Bogart Last Modified: 01-Jun-2010 15:48:40 -0700