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.

  Action Items 

• 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)

previous

minutes index

next

J. Bogart Last Modified: 01-Jun-2010 15:48:16 -0700