Subject: notes on JAS for SDA - August 3, 2004

Hi all,
   Here are my notes from the JAS discussion on Tuesday.  If I'm missed
or mis-represented something, please send email.  thanks to everyone for
their participation.
               Jean



 Attending:
   Patty McBride, Paul Lebrun, Krzysztof Genser, Dave Capista, Paul
 Derwent, Jean Slaughter, Suzanne Panacek, Elliott McCrory, tim
 Bolshakov, Saul Lipton, Tony Johnson, Max Turri, and Victor Serbo.

 Notes:

     Jean and Saul's transparencies can be found in the Fermilab beams
 Division document database - Document 1283.  the link is:
http://beamdocs.fnal.gov/cgi-bin/DocDB/ShowDocument?docid=1283&version=1

(I will be adding these notes and pdf versions of the ppt files a little
later this afternoon.)

    Jean said that the SDA(shot data analysis) group had decided to
 explore using JAS at the urging of Paul Derwent.   Since the SDA
 framework for extracting SDA and datalogger data was already Java, it
 would be very convenient to do plotting and fitting directly rather
 than using intermediate files to another program like Origin7 or ROOT.
 Also, the JAS ituple explorer looked like a user-friendly and free
 interface for users that currently use EXCEL.

     Saul Lipton was the chief guinea pig and he summarized his
 experience, which was basically positive.  He had a number of specific
 suggestions that the experts then commented on.

     Paul Derwent said that people with an HEP background ( i.e. experience
 with PAW or ROOT) immediately liked the data model.  Other
 people were less comfortable with it, and would need a learning curve.
 For now, JAS is meeting his needs for plotting and fitting.

     The JAS experts answered our the questions and reacted
 positively to our suggestions.  Victor described the use in BaBar for
 viewing ambient ( slow controls) data.  They volunteered to come out and
 help with using plugins, etc. or one of us could go there.

     Details:
       1) The memory usage issue needs to be investigated.  Saul should
 work with the JAS experts to try to understand what is going on.
 Garbage collection should happen automatically if everything is closed
 properly. Clicking the memory meter icon in JAS will force a garbage
 collection.

       2) the JAS IDE will be getting a new editor (NetBeans) but serious
 programmers should use other IDEs like JBUILDER.

       3) documentation suggestions:
           a) Something for Ituple explorer, especially for cuts.
           b) Plotter options suggestion from Saul has been implemented.
           c) add a reference in the tutorial to the release notes for
              using JBUILDER
           d) the forum was very effective for Saul.

       4) Random symbol/color assignments in ituple explorer is annoying -
 one would like to be able to easily configure this.

       5) Plotting things from 2 different ituples with the same
 structure.  JAS doesn't know they have the same structure.  Possible
 solutions are a script in JAS or there may be a merge option.

      6) Tim would like one big jar file.

      7) Saul asked how to save a manipulated ituple, with cuts and added
 columns.  It's in the file menu, at the bottom under "configuration".
 This question led to a suggestion to save configurations, i.e cut
 definitions, to be reused on other data.

      8) Future plans for JAS
           a) binary data format - something completely documented ( i.e
 not ROOT).
      9) Future support - Tony said they are gaining users.  If we use JAS
 it helps them.  There is a GEANT4 tutorial CD that has JAIDA analysis
 examples which has generated a lot of interest. ( Today I found a WWW
page on the JAS site with a list of users.- AJS)

     10) possible future directions for SDA use of JAS:
           a) It's clearly a useful option for SDA programmers and
                links/hints should be provided via the SDA home page.
           b) plugins as an alternative to ReportWriter
           c) Ituple with recomputed emittances