Attendees:  Richard Dubois, Heather Kelly, Tony Johnson, Leon Rochester, Tom Stephens, Tracy Usher

Started with Leon providing a demo of his work to fire up Wired from a Gleam job:

Leon has provided merit display by adding a widget in the event display that when clicked will provide the ntuple variable contents.  (How do we avoid this in the future?  If Heather provides a way to indicate an input merit file - is that enough?  Or is there additional work to be done?)

All agreed rotating the display was fast.  What seemed slower was going to the next event.  Possibly due to Corba client?  Also note that ROOT files are opened/loaded for every event, Tom wonders if that's slow.  Tony points out that all the steps are possibly slow. (Tony* Note there are separate issues depending if one is working with a local gleam server or a remote gleam server (the default way of running WIRED right now). In the latter case the opening root files, processing events with gleam and fetching events is all slow – way slower than the 2 seconds that Tracy is complaining about. When running with a local gleam server it is not clear that these issues are relevant).

There are two upcoming new hires - but it will take at least a month for one to come up to speed.

Tony mentioned that there are a number of JIRAs from Elliott concerning things like background colors, keyboard shortcuts...

Tony also pointed out that we could avoid Gleam entirely and just read in ROOT files.  There was a general desire to retain the connection to Gleam as its useful to debug geometry as well as reconstruction issues.

Leon asked about the future for Fred?  Heather replied it's actually looking better now that we know we can upgrade to later versions of Ruby.  Previously we thought we were stuck at Ruby 0.8.2 - but that is no longer the case.

We agreed that the next step is trying to work out the trivial issues, and then find someone to dive into the others.

To Do List

  • Look into next event speed issues LWRD-10@JIRA
  • Provide text box to input run id / event id to load random events LWRD-11@JIRA
  • Provide text box and interface to FluxSvc as Fred does to modify source used for simulation LWRD-12@JIRA
  • Provide point coordinates as the old old GUI does LWRD-13@JIRA
    • Should be possible at least for standard views
  • Provide link to updated Wired that Tracy is now using and info concerning missing plug-in that had to be installed, or improve JAVAWS to fix issues LWRD-14@JIRA
    • Includes instance tree - but that still needs to be populated for Gleam?
  • Provide reasonable way to fire up Wired from Gleam using appropriate JO settings LWRD-15@JIRA
  • Allow replay of events LWRD-16@JIRA
  • Turn off assumption that network is required to read in events
  • Update wired to allow input ior file
  • Make sure last path used for ior is remembered by Wired LWRD-17@JIRA
  • Update workbook to explain Wired usage
  • From Elliott:
    run through an Run# Event# file and produce a jeg for each event with the views selected (e.g., one jpeg with x-y, x-z, y-z views). One should also be able to setup the detail of what detector elements would appear. LWRD-18@JIRA

List of things we were working on with Heather:

  • Provide support for input ROOT relation files
    • This should already be done as of GR v17r16
      The openFile method is expecting files in the order:  mc, digi, rec, relation, gcr -- where you can insert "" for a file to skip.  For the relations to fully work - you would need to read in mc, digi, and recon - otherwise, some of the relations may be unable to be populated on the TDS.
  • Handle switching from MC to real Data within the same Gleam session, where the types of input ROOT files will change
    • Also should be complete as of GR v18r4p2 - could add this into Pass7 branch
  • Interface for input merit file
    • This is the one that still needs to be done!!

Glast Wired outstanding issues  (${entries.size()} issues)

T Key Summary Assignee Reporter P Status Resolution Created Updated Due
Refresh

List of Wired Concerns from Tracy July 13, 2010 - here for reference

1) It is very very slow
2) It does not support random event access from the controls on the interface
3) The interface for displaying properties of objects is somewhat cumbersome, also not clear how to associated objects
4) Myriad annoyances, e.g. it doesn't seem to ever remember the last folder I visited for opening the hepeventserver.ior file to connect to Gleam.
5) ... unfortunately, a few items I managed to forget before going back to Fred.
I think speed is a show stopper for me.
Also, I don't understand how we simpletons can find and play with the plugins for Fermi ehancements?

  • No labels