RunJob Meeting
15-March-2004
------------------------------------
Present: Greg, Dave, Amber, Jerry, Anzar, Robert, Eric, Lee, Peter


(1) Project News

  -Lothar will be talking to David Stickland of CMS CCS this week.
      -Subject: Requirements, software agreement. 
  -Had a meeting with Rick St. Denis, Wyatt, and Ruth on 4-March.
      -Trying to come up with a simple description of RunJob and 
       its use cases in preparation for April 8.  This is being 
       hashed over and will be released for comment later this week.
  -Jerry:  What is missing- 
       A "cookbook" manual for how to use things.  
         -Greg will try to get this out soon, and will distribute a 
          draft of it with the minutes.  He is working on one from 
          January.
  -GGF Report: Lots of interesting tools there, will send out in 
   minutes.
      -Petri Nets  (Anzar will look into) 
      -Workflow execution managers
           -Can use it (tweak it) as a workflow manager (Anzar) 
               -But we should consider it very carefully (Jerry).
               -It should be applied only where necessary. 
      -Configuration and constraints did not seem to be a big concern.
       This is our specialty.

(2) We will move meetings to 9:30 AM CST bi-weekly on Mondays.

(3) Prototypes for ConfiguratorReferences
   
  -Dave needs the new description stuff that I was going to write.
     -Greg can finish that by tomorrow night, Dave can have a prototype 
      by the end of the week.   (Neither of us had time!) 

(3) All other business

  -The contextBlock construct was not in ShahKar.  (Eric will take a look at 
      including this.)
  -Experiments participating in RunJob
     -"D0RunJob" will be used in reprocessing.  Peter Love is working on this. 
      Prototyping phase in April.  "D0Runjob" inherits from the ShahKar package 
      already.
      "MCRunjob" (D0 Version 6) will be deprecated and maintained by 
      Iain Bertram in the interim.
  -Eric will run PyChecker on what Dave has checked in (shReek and XMLP).
  -Recursive copy problem with Logger, was fixed by Dave using accessor to 
   LogRef, and there is now a __deepcopy__() methog in the Logger.