[Halld-offline] Minutes, GlueX Offline Meeting, Nov. 17, 2010

[Mark M. Ito]

Find the minutes posted at 
http://www.jlab.org/Hall-D/software/wiki/index.php/GlueX_Offline_Meeting%2C_November_17%2C_2010#Minutes 
and as text below:

GlueX Offline Meeting, November 17, 2010
Minutes

      * IU: Ryan Mitchell, Kei Moriya, Matt Shepherd
      * JLab: Craig Bookwalter, Eugene Chudakov, Mark Ito (chair), David
        Lawrence, Dmitry Romanov, Sascha Somov, Simon Taylor, Beni Zihlmann


Announcements

    We noted, for the record, the new meeting time: every other Wednesday
    at 1:30 pm.

Review of minutes from the last meeting

    We reviewed the [29]minutes of the November 2 meeting without
    significant comment.

Calibration Database Update

    [31]CCDB design, before
    [33]CCDB design, after

    Dmitry showed a couple of diagrams highlight a recent change he has
    made to the system configuration. He plans to use [35]jQuery to help
    handle the web interface to the database.

    Matt suggested that a specification for the API be circulated for
    comment. Others may have their own ideas about how the database should
    be used; those viewpoints should be collected and appropriately
    incorporated into the design. Dmitry and Mark will work on fleshing out
    this section of the [36]specification document.

Python encapsulation of C/C++ Code

    Craig gave an overview of his experience doing this with CLAS analysis
    software. See [38]his slides for details. He used the [39]SWIG package
    to auto-generate the glue code to make the compiled code available from
    a Python script. Since the compiled code has been worked over by others
    for some time, errors introduced in using it were often in his own
    private code. He has found debugging those errors in the context of a
    scripting language much easier than doing so in say C++ or in the ROOT
    framework. Rapid development, more work done, by leveraging the
    software development work of others in a much simpler programming
    environment.

    There were caveats of course. The auto-generation can have difficulties
    for certain language constructs. We thought of examples where this
    approach might be used to wrap existing API's:
      * HDDM
      * EVIO
      * the new calibration database routines

    For the first two, a wrapped interface could be used to do simple skims
    of data sets. For the last one, command line interfaces and web
    interfaces would both be much easier in a scripting language.

Parallelism Workshop

    David reported that the Parallelism Workshop will be held January 6 and
    7. The JSA Initiative Fund will contribute to the cost. See the
    [41]conference webpage for details.

Action Item Review

    We went over the [43]resolved and [44]outstanding items. Folk were
    encouraged to address tasks assigned to them.

    Retrieved from
    
"http://www.jlab.org/Hall-D/software/wiki/index.php/GlueX_Offline_Meeting%2C_November_17%2C_2010"

References

   29. 
http://www.jlab.org/Hall-D/software/wiki/index.php/GlueX_Offline_Meeting%2C_November_2%2C_2010
   31. http://www.jlab.org/Hall-D/software/wiki/index.php/Image:Ccdb01.png
   33. http://www.jlab.org/Hall-D/software/wiki/index.php/Image:Ccdb02.png
   35. http://jquery.com/
   36. 
http://argus.phys.uregina.ca/cgi-bin/public/DocDB/ShowDocument?docid=1541
   38. 
http://www.jlab.org/Hall-D/software/wiki/images/3/38/Craigb_python_encapsulation.pdf
   39. http://www.swig.org/
   41. http://conferences.jlab.org/Parallelism2011/
   43. http://tinyurl.com/2beqwlk
   44. http://tinyurl.com/25g8af6