Software Coordinating Committee Conference Call
		       September 5, 2002 3:00 PM EDT
			  Recorder: C. DeTar

Present:

  Brower, Osborn, DeTar, Mendez, Holmgren, Simone, Mawhinney, Edwards, Watson 
  Neff, Pochinsky, 

Absent:
  
  Kostya Petrov, Chulwoo Jun,

** Action items

----------------------------------------------------------------------
Agenda:

Hartmut Neff starts at BU Nov 1 but is here this week and next. I
asked him to join the meeting so he can get an ideas how he would like
to contribute.


As ususual:

1) Update on QDP C Level 2 API (DeTar)

2) Update on QDP C++ API  (Edwards)

3) Where do we stand on I/O. 

Are we ready to divide up the task to deal separately subtopics? (I
don't have a fixed idea of who should do what, but I have put down
some names to "stimulate volunteers!")

     a) Programming interface - a QDP (level 2) data-parallel instruction set.
       (Edwards, Brower, Pochinsky)  
     
     b) File formats --- extension of NERSC system.
       (Mawhinney, Detar) 

     c)Web access/cataloging/archiving
       (Watson, Holgrem)

----------------------------------------------------------------------

    Brower:   Status of test code for QMP (Jim and Celso)
    Simone:   Celso and I are getting started.
    Celso:    Getting my environment set up and migrating profiling stuff.
    Brower:   Test suite could be developed in parallel with profiling.
    Celso:    Will be out of town next week.  Week after to proceed.

    Brower:   Update on QDP?
    DeTar:    Sent out the qdp.h file two weeks ago - 
              The LaTeX document will follow soon.

    Brower:   QLA (C)?
    DeTar:    Developing a code generation scheme via Perl
    Brower:   Can we see a sample?
    DeTar:    Hard to do that without the whole product.
              Should be ready in a couple of weeks.

    Brower:   QDP++
    Edwards:  Waiting to converge on the semantics.  

    Brower:   I/O?
    Edwards:  Gave a summary
    DeTar:    Also want to contact Edinburgh group
    Mawhinney: Craig McNeile, Alan Irving, Booth? are the players
    Watson:  They are working on metadata schema and file format
             Also finding ways of describing the binary data (format data)
             as well as physics data.  No solution for single file format
             yet.  Also looking at cataloging.  They are looking at XML 
             databases.
             It would be good for SciDAC to get involved to help assure
             that an international standard might emerge.
    Watson:  Richard Kenway agreed we would name 1 or 2 people in SciDAC
             to be engaged in the process.
    Brower:  Who?  Simone and Edwards and Watson for database and 
             cataloging, since overlaps with datagrid.
**  Watson:  Will send names to Richard
    Edwards: (Sent a gif from Chris Maynard to Simone on file formats)
    Mawhinney: For storing propagators, should we build in to
             the interface a process for storing the source and
             gauge configuration to guarantee the necessary
             information is kept?
    DeTar:   Should we be putting metadata in the binary file or
             separately?  I would advocate having one metadata file
             for holding information common to a whole set of files
             and a more minimal set of metadata to be kept in each
             file.  That would make it easier to edit global parameters.
    Watson:  Let's defer these decisions until we interact with
             Edinburgh.
    Pochinsky: Should we be allowing the user to edit any metadata?
    Brower:  Perhaps we need a tool for editing.
    Brower:  Need a proposal for a QDP extension to do I/O
**           Brower, Edwards, Pochinsky, DeTar, Petrov
    Mawhinney: In QDP interface need to distinguish between temporary
             file storage and archival.

----------------------------------------------------------------------
Conference concluded at 2:05 PM EDT.
Next plenary call Sep 12 3 PM EDT.