Pysynphot Commissioning Effort
Status Report 16 June 2008

- A test procedure has been defined.

   SSB and INS have agreed that synphot will be used as a "brass
   standard", and initial commissioning tests will be run to compare
   synphot and pysynphot results for the same cases. Significant
   discrepancies will be investigated to determine which software
   package gives the more correct answer.

   Additionally, tests will be defined for cases in which synphot is
   known to give incorrect or questionable results, which will be
   scrutinized more carefully.

- A set of ETC-related cases has been defined and run for all
  instruments that use the ETC.
     - these are still missing the thermback cases.
     - there are over 23000 tests in this set
     - several new pysynphot errors have been identified
     - an ETC obsmode specification error has been identified
     - a sporadic IRAF problem occurs during large test runs but not
  when tests are run individually; investigatino ongoing.
     - synphot.calcspec sometimes generates errors pertaining to the
  renormalization wavetab when countrate does not
     - for more detail see errors.txt in this repository

- A set of science-related cases has been started.
     - these include the narrow emission lines used in the ETC
     background, which demonstrate some interesting differences

- Scripts have been written to automatically run the above test sets,
  perform statistics, and generate plots.

- A multi-platform set has been run.
     Two test cases were reported to me that demonstrated discrepant
     results on Solaris, Mac, and Linux platforms. I have run these
     cases on various platforms and determined that pysynphot
     performs identically platforms on which synphot differs.
     results and list of platforms tested are
     located in /eng/ssb/syn_pysyn/platform. )
