Notes for science tools meeting 11/30/05
 

• Sky model environment variable

  I would like to propose a standard environment variable for the location of the sky model files, with the assumption that all files are in the same folder, as now:

  checkout3_blazars.xml
  checkout3_diffuse.xml
  checkout3_dm.xml
  checkout3_isoblazars.xml
  checkout3_other3eg.xml
  checkout3_srcIds.dat
  checkout3_xml_files.dat
  EgretCoinc_library_trim_new.xml
  EgretPulsars_fierro_library.xml
  GP_gamma_v0.fits
  GRBobs_user_library.xml
  gtobssim.par
  source_names_full.dat
  Tosti_lc.fits
  Wang_mapcub1.fits

  This means that every filename specification should have
  $(SKYMODEL_PATH) (name to be decided) prepended to the file specification.
  Example from the file checkout_3_blazars.xml:
   

  <source name="J0737p1721">
      <spectrum escale="MeV">
          <SpectrumClass name="SpectralTransient"
  params="flux=2.153753e-02, tstart=900., tstop=4.752000e+06, templateFile=Tosti_lc.fits, emin=20, emax=2e5, lc=17, z=4.200000e-01, useLogParabola=0"/>
      <celestial_dir ra="114.47" dec="17.35"/>
      </spectrum>
  </source>

  
  I would like to replace the fragment "templateFile=Tosti_lc.fits" with "templateFile=$(SKYMODEL_PATH)/Tosti_lc.fits"
  This is supported by the code in SpectralTransient.

  Similarly, but not of interest in this forum, I would then have a line in the job options,
  
      FluxSvc.source_lib += {"$(SKYMODEL_PATH)/checkout3_blazars.xml"};
  
   

• The status of time.

  All times are expressed as Mission Elapsed Time (MET).  This is the number of seconds following the official Mission Start, 1/1/2001 UT. The code has a global function defined, astro::JulianDate::missionStart(). From the file astro/JulianDate.h;

  /// the GLAST official mission start: 1 Jan 2001 00:00, which should be 2451910.5
  static JulianDate missionStart(){ return JulianDate(2001,1,1,0); }

The current start (launch?) for DC2 is 1/1/2008, which was used to generate the pointing history file. I presume that a time, like the tstart=900. above are not MET, but should be referenced to the start of the simulation, like DC2. That is, how do we interpret tstart and tstop? I propose that the code in SpectralTransient simply add an offset corresponding to the difference between Mission Start and the Launch. In DC2, this is done by the line
    FluxSvc.StartDate = "2008-1-1 00:00:00"; // launch

The corresponding MET value is then available from the static function flux/Spectrum/startTime().