Re: astrometry

[<aah@nofs.navy.mil>]

Michael asked some questions:

    - what if 2 stars are detected one night, but only 1 (merged)
               image on another night?

    - what if 2 stars are detected one night, but only 1 (single)
               star another night (if one star fades, for exampl)?

    - what if 2 stars are detected one night, and 2 stars are detected
               another night, but the astrometric solutions on the 
               two nights are sufficiently different that the 
               detection of star A on night 1 matches up with the
               detection of star B on night 2?

Remember, the limitation here is *not* seeing, but optics.
If things are focussed properly, blends will stay blends and
vice-versa.  Also, astrometric solutions are always good to an
arcsec or so; again, you shouldn't have confusion.  Situation 2 is
more rare -- you can't guard against everything.
  I've already mentioned that I think software gurus should think
strongly about the blending/crowding problem.  Another area of
extreme interest is real-time analysis.  I intend to produce
starlists on-line, but the analysis of those starlists is much
more complex.  However, *noone* is currently doing real-time
analysis.  If you did, you could find the cataclysmic variables
that were in outburst, be the first to see a nova, etc.  What
I think is needed is a post-starlist program that can be spawned
after the starlist generation and can do the correlation between
existing catalog and new photometry to find those bursting objects.
For CVs, you have a catalog of ~1000 known or suspected variables,
which could just be checked against the incoming photometry to
find the outbursts; no discovery is necessary.
Arne