Core Minutes 1/26/2010

• Snow days at SLAC: or more accurately, rain and lightning days. SLAC was shut down to most employees Tuesday and Wednesday last week after both the primary and back-up power lines went down. (Richard) There was an ISOC post-mortem at which our response to the outage was discussed. We did well at restoring processing soon after power returned (as SLAC did well at restoring services as quickly as possible) but communication — ours to FOC and especially SLAC to us — could have been better. The back-up ISOC on campus worked well enough; could be better documented. (John) The outage caused no particular problems for FSSC.

  (Richard) Berrie will mirror the collaboration email list as well as the new alternate email list Pat set up during the outage. It contains alternate email addresses for those whose primary email address is at SLAC. Chuck is in the process of making a couple Google sites, one protected and one open, so that there will be an off-site back-up for critical information.

  More suggestions on how to respond to extraordinary circumstances are welcome.

• From Heather: Still officially on leave, she nonetheless forwarded some items to Richard:

  • rhel3 builds will be terminated this Thursday, Jan. 28.
  • Eric encountered a ROOT problem while attempting to build TMine. (Anders) This is a documented problem. It can be avoided by rearranging code (or by using an older version of ROOT). (Tracy) In a somewhat related matter, Eric has been working on getting TMine to build on Windows as a step in the process of making it a standard product, but he's stuck at the moment.
  • Jonathan Ormes noticed some magnetic field variables were missing fromt the ntuple. (Leon) has sent a log file to Heather.

• ScienceTools: (Jim) Working on various issues; no progress to report yet.

• FSSC: (Eric) No news

• Documentation: (Chuck) See Richard's discussion of response to the power outage above.

• GR:

  • Pass 8 (Tracy) A new, Pass 8-capable version of GlastRelease will be coming out soon, hopefully before the Pass 8 meeting this Thursday. It has new AcdRecon and CalRecon code. The latter instroduces simple clustering, allowing for the possibility of multiple clusters.

    Updates to go into GR v18r1p0 [tagged on Thursday, Jan. 28] include

    1. Modified CalCluster TDS class to split the "fit" parameters (the minuit fit to find the centroid and axis) into its own location in the class, as opposed to being hidden in the"moments analysis" parameters.
    2. This requires changes to all downstream code, including RootIo
    3. Include input/output of the CalCluster <—> CalXtalRecData objects used in the event display
    4. Updates to the event display to aid in disentangling multiple clusters when running the multiple clustering mode.
    5. Includes updates to the "new" AcdRecon.

    (Tracy, Leon) An issue involving overlays has arisen: sometimes not all the energy in the overlay gets transferred. Small differences are to be expected, but the difference is not always so small. The effect tends to be worst when the overlay energy is large.

    A task force is looking into possible improvements to the onboard filter. Tracy is doing his best to keep a low profile.

  • Pass 7 (Anders) Eric C. and Maria Elena are working on outstanding mysteries.

• Core talk:

  • New SCons versions (Joanne) We're currently using 1.2.0.d20090919. A newer checkpoint release came out in December (d20091224) and yet another one this month (d20100117). The main differences are improvements for Windows but unfortunately the goodies come at the expense of a changed interface (for Windows only), incompatible with the old. 1.3.0 is expected very soon. We'll make the jump to the new interface after that — and after I resolve an outstanding problem with it in case there are multiple installations of Visual Studio and the user doesn't want the default (newest) one.

  • GR and SCons (Joanne) Major remaining issue other than those involving externals like gaudi, for which she is thankfully not responsible, is job options. There is a plan; it will take a while to fully implement.

  • ST and SCons (Joanne) The issue Toby raised last meeting, concerning installation of python files, has been addressed in code that is committed to CVS but not yet tagged.

previous

minutes index

next

J. Bogart, Last Modified: 01-Jun-2010 15:45:51 -0700