Core Minutes 11/8/2005  

Present:    Joanne Bogart, Toby Burnett, Eric Charles, Jim Chiang, Johann Cohen-Tanugi, Seth Digel, Richard Dubois, Warren Focke, Berrie Giebels, Nicola Giglietto, Navid Golpayegani, Tony Johnson, Heather Kelly, Lester Miller, Igor Pavlin, Leon Rochester, Eduardo do Couto e Silva, Tracy Usher

• Upcoming GlastRelease Release: Leon's tracker work is ready to go; Heather's AcdRecon upgrades are close, should be ready later today. Tracy's work geared towards importing Bill's classification trees won't be ready on that time scale. He's not sure how to handle Filter cuts; an example would help. We should plan on reprocessing after his stuff is in. Toby has been working on an alternative, involving producing his own trees which are expected to be functionally similar to Bill's. It won't go in today's release either.

  Toby has done considerable refactoring of merit recently. In particular the static library and executables which used to be built by the package are gone. A new executable, apply.exe, has been added to GlastClassify, which will apply decision trees to a root tuple.

• Background interleaving: (Toby, for Toby and Tom) The idea is to treat interleaved background like another source: it has rate, provides a 'next event', sort of like the clock. This 'next event' is caught by FluxAlg, which will invoke a service Tom will provide to the next ROOT background event. Tentative release date is late November.

  This (background pseudo-source), fired at a high rate, could be used to monitor dead-time.

• Data handling: (Tony)

  • Upgraded single sign-on to accept either unix or Windows id
  • glast-ground portal has a bunch of new, useful links.
  • Pipeline database now has its own dedicated Glast Oracle Server. The migration went smoothly, and the new server is faster.
  • Pipeline web interface has been updated to provide more information
  • (Richard) Executables and external libraries have been moved to afs disk space. Have been making use of newer nfs servers as well, and jobs in the batch farm are now writing to local disk when space is available. Overall effect of all these changes has been much greater reliability
  • (Richard) Tony and Dan are working on the next round of Pipeline changes, expected to be more evolutionary than revolutionary.
  • (back to Tony) Data server portal should be ready by about tomorrow. Igor is adding verification of user input so that the user will know immediately if something is wrong.

• Science Tools support request: (Jim) Details are in Confluence. The three items all concern Python support. (Navid) The last one should be doable in short order.

• New cfitsio? (Seth) Other systems like FTOOLS expect a newer version than we're using. Could we upgrade? (Jim) A better solution would be to rename our library to something including version; then other packages using cfitsio would ignore ours entirely. Jim and Navid will see about implementing this.

previous

minutes index

next

J. Bogart, Last Modified: 01-Jun-2010 15:48:29 -0700