[fitsbits] FIT Keyword data dictionaries

[Doug Mink]

Peter Bunclark wrote in response to Doug Mink's citation of an image header
with DATE-OBS= '1611-04-15':

> The FITS standard tell us that DATE-OBS ``...value field shall contain a
> character string giving the day on which the observations represented by
> the array were made...''  so is it, in fact, an appropriate keyword to
> represent the epoch of an artificially-generated map given that the
> observations used were presumably made much later?

In this case, the visible planets in the image were actually from the epoch
of DATE-OBS.  If I had a catalog with proper motions deep enough to show
the Milky Way well, I would have used it, and the stars in the image would
have been of the appropriate epoch as well.  I might try it again with
the Tycho-2 to see how it looks--I unsuccessfully tried the PPM, which was
the biggest all-sky available when I did the project, just before the release
of the Tycho catalog (whose faint star proper motions wouldn't have made the
17th century star positions any more accurate than the HST GSC anyway).
Part of my project (without the FITS files) is online at

http://tdc-www.harvard.edu/software/skypic/rome97.html

-Doug Mink