Attendees - Gaines, Canal, Berman, Kallenbach, Watts, Murat, Kennedy, Don Holmgren

Jeff presented slides on the status of the MATLAB evaluation, and Eileen presented a preliminary comparison of Octave to the PASREC and PASSUMA requirements.

Points from the ensuing discussion:

Both products seem to have some pretty severe deficiencies for our use. Both products are limited to whatever data can reside in memory. In both cases, the basic unit of data manipulation is the double-precision matrix. All data to be manipulated must be converted to this form. No integer data type exists. The programs cannot deal with 2-D histograms.

While MATLAB has good visualization capabilities, Octave graphics are very crude (GNUPlot).

The investigation into both of these products needs to be focussed on the things that PAW, ROOT, and Nirvana are built for - HEP-specific processing. The following questions need to be answered:

What upgrades will have to be made to allow the processing of 2-D histograms?

Are the programs capable of looping through HEP events?

In the case of Octave, can a good graphics package be connected to replace GNUPlot?

How can they be upgraded to accept integers?

Can they be upgraded to work around the size limitation?

Upcoming meetings: Aug 4 - detailed discussion of ROOT. Aug 11 - detailed discussion of Nirvana. Aug 14 - MATLAB and Octave.