Testing log - 2007.06.28
Steven T. Myers

-----------

Test using Jupiter 6cm Synth2000 tutorial data:
/home/sandrock2/smyers/NAUG2/Data/VLA_CONT/FLUX99-6CM.CBAND
or
http://www.aoc.nrao.edu/~smyers/naug/Data/tut2000/VLA_CONT/FLUX99-6CM.CBAND

Machine: sandrock
/home/sandrock2/smyers/Testing2/Jun07/

-----------

Note: this is RPM daily (CASA Version 2.0 Rev 2666 Built on: Wed 2007/06/27 01:15:06 UTC)


*** = serious problem
**  = serious problem
*   = important issue

plotxy :

   selectsp is gone (yea!)
   correlation now working, e.g. correlation='RR LL'

   killing the plotw window has no adverse effects

   if you do plotxy then plotcal keeping the same window
   you will get two sets of buttons - killing window in between
   fixes this

   if you do plotcal then plotxy, plotxy crashes or hangs

   if you run a plotxy, keep the window up, delete the ms, then
   try to flag, you will get a coredump

plotcal :

   I could not figure out how to unzoom without just iterating
   to the next panel

   I you want to iterate (antenna) when you do 'n' it keeps
   overplotting.  If you kill the window before doing 'n' you
   will get the clean plot w/o overplotted previous antennas.

split : 

   today it works ok (with field selection), though field is still
   specified as an ID

clean :

   * get "*** Error ***  index out of bounds" if field=''
   used field='*'

   used cleanbox='interactive' mode - sweet!  But should have
   a way to reset the npercycle on-the-fly in the interactive window

   should there be an alternate way to specify the cycle cutoff?
   some sort of threshold?  I dont know a robust way to specify this,
   maybe a cleaned flux limit (fluxpercycle)?  or a threshold factor
   (threshfac, in units of threshold)?

   looks like it does a 1-iteration clean to get the "dirty" starting
   image.  This is probably ok unless you have high loopgain (or does
   it use an infinitesimal loop gain for this one iter?).

   clean really needs to report more useful stuff like:
      total number of iterations
      total cleaned flux
      maximum residual in final image (not just final mask)
      maximum in final restored image
   (many of these can be computed later but its nice for the user
    to not have to do extra work)

viewer :

   when reporting statistics in a region should report the vertices of
   the region drawn also

imhead :

   need to put the mode options in the "inp" text
   mode                =     'list'        #   get or put header values
   should be:
   mode                =     'list'        # 'get','put','summary','list','stats'

   mode='summary' seems to be same as 'list' except no writing to
   terminal (same stuff in logger) - is this needed?

selfcal :

   currently have to string together ft/gaincal/(accum)/correct
   should we have a "selfcal" task to do this?

   (gaincal)

   get this warning with calmode='p':
   Thu Jun 28 20:16:59 2007      WARN calibrater::setsolve:
   Please note that the phaseonly parameter will soon be eliminated.
   You should begin using apmode='P' for phase-only.