Core Minutes 1/6/2004
Present: Ursula Berthon, Joanne Bogart, Toby Burnett,
Xin Chen, Johann Cohen-Tanugi, Seth Digel, Richard Dubois,
Berrie Giebels, Navid Golpayegani, Traudl Hansl-Kozanecka, Yasushi Ikebe, Heather Kelly, Pat
Nolan, James Peachy, Dirk Petry, Sean Robinson, Leon Rochester, Robert
Schaefer, Tom Stephens, Tracy Usher
-
Gleam post-DC1: (Toby) Updates
of several externals are imminent:
- Root 3.10. Toby has been using it on Windows, where it is a big
improvement over the current production version, 3.04. It needs to be
installed on SLAC Linux Heather is
keeping track of everything that needs to be checked out before it becomes
production.
- CMT v1r14. We need this for the Release Area features, which should
simplify life for users considerably.
- Gaudi. Required for Visual Studio 2003 compilation
The GlastRelease released version we have now should be adequate for
remaining DC1 work. Leon has a TkrRecon patch which he will tag, but it is
not needed for DC1. Toby has promoted most new tags to GlastRelease HEAD; we
can now return to the more usual procedure in which package owners are
responsible for promoting new tags to HEAD.
Although Navid has completed work on the new version of
TagCollector, he was unable to install it because SLAC Security took away his
ability to write to the appropriate web-accessible space. They like to
personally anoint all who have this power. Should be possible to arrange this
soon for Navid [shortly post-meeting, this has been accomplished. ed.]
- Xerces 1.7 to 2.4: (Joanne) This
upgrade is far from transparent. Packages using Xerces classes directly
will need considerable modification. It's probably worth it, but people
should be aware of the coming disruption and the need to verify that
everything still works once the changes are made. A status page à la Heather
would help.
-
ScienceTools report: (James, Richard)
Toby and James have taken joint responsibility for ScienceTools architecture.
They're just getting started looking what's there and what's needed. Any
proposals for new packages should be directed to them for evaluation, naming
suggestions, etc.
(Toby) He has recently added a package,
tuple,
whose job is to provide a simple means for applications to loop over rows in
a table (FITS or ROOT).
-
DC1 - what went wrong: (Richard)
Although the Data Challenge was basically a success, there were plenty of
problems:
- The Windows builds of ScienceTools didn't work. First there were
compile problems; once these were fixed it still didn't run.
Have to get ReleaseManager Windows build going
- There were problems with ACD ribbons
- There was a mad rush over the week-end to get things together before the
meeting.
Don't
start such meetings on Monday
- Too much manual processing to create final datasets, builds and tar/zip
files
- Late discovery of problems with orbit and sources
Need better testing and diagnostics for this part of the code
- Insightful Miner worksheets were unstable
- We never got OPUS, CALDB, backgrounds, or seasoned IRFs going.
- Using the GSFC GRB package for the last 6 days of data took some doing;
there were several problems to be ironed out.
- Likelihood doesn't react well to bad input, like filename of a
non-existent file.
- The table of MC source id versus name isn't enough help to figure out
which sources are which.
- The failure rate among batch jobs was lowered from 10% to just a few
percent by increasing the size of an array (in there for a somewhat kludgey
purpose). Although for at least one failure type it was possible to rerun the
job in the debugger and reproduce (sometimes) the failure, debugging the
optimized code is at best bizarre.
See if similar failures can be reproduced with optimized code.
-
Network Monitoring: (Richard) SLAC has
a program of monitoring network response of various kinds for collaborating
sites. We should get this going for all our collaborators' sites (especially
Goddard, but others as well).
Some statistics already exist.
- Document effect of MC pruning options (Riccardo) (8/5/2003) It's
documented in the code; Tracy will look into putting the information in a
more accessible place. (9/2/2003)
- Reorganize RootAnalysis package so structure is more standard.(05/13/03)
- Write up, and perhaps present at a future meeting, hints on debugging.
(02/18/03) Some combination of Toby and Tracy will provide something for Windows.(7/22/03) Linux
document has been linked to how-to pages. (10/14/2003)
- Overhaul documentation altogether by about November (Heather and ??) (9/2/2003)
- Resolve calibration/timestamp issue. (11/20/02) Online folk have not yet thought about
putting timestamp information into EM data, but, thanks to Richard's
inquiry, may now do so. (2/25/03) We can probably live without a realistic
system for EM, but we will put it on the agenda for the DC1
workshop.(5/6/03) At the DC1 Workshop Steve Ritz promised to find an
authoritative reference for timestamp format and pass it on. (7/17/03) Now
that events have "real" timestamps, Joanne needs to reassess
situation with Calibration (9/2/2003)
- We need a web page on the workings of the Release Manager
(5/6/03) In progress (Navid) (9/2/2003)
Done! (1/6/2004)
J. Bogart Last Modified:
04-Aug-2004 15:41:52 -0700