Software Meeting Minutes - March 4, 2003

Present: Eugene Pasyuk, Raphael D., Mark Ito. Rafeal H., Eric Clinton, Abdellah A. (phone), David L., Itaru, N.


Some slides shown:


Dave presented an overview of the new PEvent_Loop class recently committed into CVS. This incorporates existing code that was common to most of our analysis programs (bankdump, prim_ana, ...). The prim_ana program was reduced by about 100 lines (from 182 to 89). A short discussion about optional methods of implementing this functionality ensued.

Dave presented an overview of the new PEvent_Output class. This allows output of PrimEx banks into a CODA formatted file. The online data is written in CODA format and there is a desire to have the Monte Carlo and filtering programs write in that format as well. The current implementation of the output does NOT produce files with exactly the same structure as the online data. This led to a discussion over how much work we should invest to make them identical. The discussion stalled when an accurate description of the current online format could not be recalled. (After the meeting, I looked it up and it can be seen here: http://www.jlab.org/primex/documents/meetings/mar_4_2003/bank_format). The output format also has the option to compress and decompress files on-the-fly in gzip format. Preliminary tests on one file showed a 75% reduction in file size for a 40% reduction in event processing rate. Cross-platform testing is stll needed. There were no volunteers.

Dave presented a list of PrimEx Software Subdivisions. The idea was that these are semi-distinct areas where one person could be assigned as responsible. The list is posted on the web at the following URL. It has been updated to include names assigned at the meeting.

http://www.jlab.org/primex/documents/meetings/mar_4_2003/divisions

A preliminary agenda for the upcoming PrimEx Software Workshop was presented and discussed. This included assignment of names and a discussion of the optimal timing and format of the workshop. The workshop is planned to be near the collaboration meeting which has not yet been scheduled. It is expected to occur in late March, 2003. There was strong sentiment that the best format would be to have a 1/2 day workshop on the day before the collaboration meeting so a report summarizing the meeting could be given to the entire collaboration. The preliminary agenda can be found at the following URL. It has been updated to include names assigned at the meeting.

http://www.jlab.org/primex/documents/meetings/mar_4_2003/Agenda