Science Tools Preparation for DC2

S. Digel, 17 January 2006

Assume that we'll have a good sky model and that it and the interleaved background will be run in a timely fashion.  What's left in terms of getting ready for DC2?

Serving data at GSSC & SLAC

What to serve? FT1 & FT2 and (at GSSC) FT1++.  GBM data, phony trigdat messages, precomputed livetime cubes, ephemeris file, preliminary source list.

The Data Handling group at SLAC was last week considering two alternatives for having an event display on the Web for DC2.  (This assumes that we'll have an event number in the FT1 files that relates to the run and event id when the event was generated, and that the interleaving of backgrounds is transparent to the user.)

Defining instrument response functions

For DC2, Julie has organized a group (Jean, Jim, Toby, and Riccardo) to look at event classes and parameterizations of instrument response functions, and to verify the parameterizations.  The goals are accuracy and computational efficiency.  One not-so-subtle issue that I did not really appreciate is that a given event may belong to more than one event class (like Bill's base classes).

Specifically in terms of science tools, what else is left to do?

Tools for data formatting

makeFT1 (makeFT1++) updates.  makeFT2(a) needs updating, too; in terms of the data generation for DC2, the livetime accumulation should be made consistent with the position and attitude-dependent trigger rate and SAA passages.

Distribution and use of science tools

For Windows and flavors of Linux other than RHEL 3.

• Windows is still a mixed bag - builds are routinely made but user interface is not really workable
• Python GUIs and interactive likelihood - will the installer be able to provide them?

User interface and plotting

• Python GUIs - for observation simulation and likelihood
• ROOT GUI for parameter files in general
• Data visualization, exposure maps, response function visualizer (Jim's?)

Update and expand the documentation

Reference guides are in general terse and some are incomplete or out of date

Prepare tutorials for the workshop

Like for checkout 3, but I guess more so.  We will need to go into more depth on several topics, including:

• 'data exploration' using ftools and the science tools utilities (counts maps, counts spectra, exposure and intensity maps)
• the instrument response functions - what they are
• living with FITS files
• strategies for using likelihood analysis (binned vs. unbinned, speed/accuracy tradeoffs in general, interpretation of TS, maybe modeling residual background)

Either we or the collaboration science working groups will have to set up an effective system for sharing results, posting questions, and reporting bugs.  The latter is certainly JIRA.  The first 2 are probably Confluence, but Confluence gives you enough rope to hang yourself in terms of organization.