[fitswcs] Improving FITS header astrometric information

[Peter Bunclark]

EPOCH = 'J1991.25 '

I think the guy means equinox 1991.25, but fundamentally the misuse of the
word EPOCH has got to be stopped.  Perhaps if you find examples, we could
publicy ridcule the authors or excommunicate them from the IAU or
something.
	The real epoch is the mid-point of exposure for an image, and
should be expressed as JD (or MJD).  DATE-OBS, of course, is the same
information in human-friendly and much more complex form.
	If the FITS-writer doesn't understand the concepts, then you are
never going to be able to automatically determine what he might have
meant.   However, perhaps you could look for MJD and DATE-OBS, if both
then check they are self-consistent, if one then check if within some
precsion of EPOCH, if not then the EPOCH value might be the EQUINOX.
But would you publish a result based on this data?


Pete.