Core Minutes 11/11/2008

• News from ADASS: (Richard) Three 15-minute plenary Fermi Talks, all in a row, went well.

  Google Sky (see also web version) is interested in us. Main developer was there along with another developer from UC Davis. It's a competitor of Microsoft's Worldwide Telescope which is, possibly, a more feature rich app, but runs only under Windoze. Generally useful application. Can Fermi release sky map and catalog or arrange to make it available to LAT collaborators only? Tools to take sky image and divide into tiles. (Toby) Will speak with Andy Connolly about utilizing these tools with GoogleSky. Marshall has been working on this, including running in server mode which provides additional features and possibilities. (Richard) Possibly use this tool to bring up light curves, or hook up to ASP.

  VO Events Howard Williams - receives SWIFT and Fermi GRB alerts, some problem with Fermi data (being investigated), tools to convert to .kml.

• GR status: Overlays Work continues. Can now read DIGIs, perform calibration, write out in a form to be used for overlay. Discussion with Leon and they will change direction slightly: will create/store new output classes containing the overlay info, as DIGIs are not ideal for this purpose, especially for the calorimeter. Hopes to have a working version "in a few days". Preparing to gear up for "system tests" to compare results with data. (Julie) We need this by early February.

  v15r49 branch (Leon) Updated tags are being finalized. HepRepSvc has improvements so that the WIRED event display can show ghost tracks.

• Diffuse response columns: (Richard) discussed adding these columns to FT1 files with Tom Stephens at ADASS. There are two columns: galactic and extragalactic. This is a very CPU-intensive operation and we don't want to recompute this any more often than necessary. There are issues with implementing this [lifted from Richard's email]:

  where to run this

  I recommend it be in L1proc for ongoing processing; can it be done in parallel?

  we'd need a reprocessing task to backfill

  optimising the filling

  we'd save a lot of CPU time in the short run by only computing the response for diffuse events, ie have gtdiffrsp take note of CTBCLASSLEVEL - this may change when we go to events being in one class only.

  what to call the files

  special name for each reprocessing or increment L1's version numbers?

  this is a general question for any bulk reprocessings

  how to get the files into the astroserver and FSSC server

  what to call the diffuse columns

  TomS & I recommend fixed columns names and a version in the header

  (Toby) When diffuse model changes do we need to recalculate? (Richard) Yes.

• Reprocessing status: (Tom) Going well. Most recent GlastRelease, v15r49, is working and 20 runs from August have been reprocessed and in data Catalog (/Data/Test/Flight/Repro/ReproTest8). This includes diffuse model calculation. Next steps include completing the full August 2008 reprocessing (once again) and validate. Then, convert the "test" reprocessing into "production" reprocessing, which means finding a permanent and logical place to store the files (in xroot and data Catalog), and making the files visible to the astroserver and FITS skimmer. That would then conclude spacecraft alignment reprocessing project. The next reprocessing project involves the diffuse response issue discussed by Richard above.

• ScienceTools: See the latest report.

• FSSC Report: (Eric W.)

  • code freeze end of October based on Science Tools v9r8p2. Very few mods needed to generate successful builds, folded back into SLAC code.
  • Providing binaries for Linux SL4/SL5 and Mac (versions?)
  • Request from Seth about providing binary libraries for platforms not supported at SLAC

• Documentation: (Chuck) A good part of last week was spent tweaking the iLat tutorial, and we're starting to get some feedback on it. In following the HelpSoft list this morning, there is a growing demand for instructions on how to install the iLat software and set it up to run on personal computers. This demand appears to be driven by server/router timeout problems people are experiencing when working offsite. I'll start working with Nicola and Andrea to get a reliable instruction set together as soon as possible.

  Yesterday, I created a new tag for the workbook, and Berrie has updated the mirror site. I've also started work on SCons documentation, which is going well thanks to Navid's work in Confluence. Both the SCons and Joanne's GoGui documentation will soon be linked in from the splash page of the Advanced section of the workbook.

  And ... pictures taken during the core meeting held here at SLAC were finally posted to the web late last week.

  [thanks for the post-meeting update Chuck!]

• Core talk:

  • ROOT and skimmer: (Heather) reported on ROOT 5.18e with new xroot client code. First tests were discouraging. David Chamont is on the verge of releasing a new skimmer which should fix problems with reading a large number of merit files.

  • KIPAC software collection: It's available at SLAC, including HEADAS software (e.g., ftools). Heather is following up on understanding whether we can use that installation within Fermi.

  • SCons tag collector, RM (Navid) hopes to have tag collector done this week. ScienceTools builds with SCons are still broken (due to problem in pointlike) but should be straightforward to fix. Karen is working on front-end web page for Release Manager. She's making progress but still needs feedback.

  • GR and Externals (Emmanuel) Making progress on GR build with SCons but there are still some problematic packages. See GlastRelease-scons build status page for details.

    Docs for installing external libraries are being collected and maintained here in Confluence.

  • GoGui (Joanne) was at the Desert Museum near Tuscon during the meeting (wonderful place), but prior to that made a couple improvements to GoGui:

    • Package delete function added
    • Disable build, clean for "extra" versions of packages which SCons will not in fact operate on (old behavior was that build or clean would happen, but on a different version of the package being selected).

previous

minutes index

next

J. Bogart, Last Modified: 01-Jun-2010 15:45:58 -0700