[fitswcs] WCS documents

[Steve Allen]

On Mon 2001-09-17T15:50:07 -0400, Bob Hanisch hath writ:
> We should not, however, endorse a solution in which the attributes of the
> WCS are implicit rather than explicit, and where the existence of WCS
> keywords with indices higher than the value of NAXIS might, or might not, be
> present.

This issue of standard practice in the use of NAXIS extends to the
situation of multi-HDU FITS files for mosaic images.

In the case of the NOAO mosaic the PHDU has NAXIS = 0, and no NAXISn
cards.  In the case of the DEIMOS mosaic the PHDU has NAXIS = 2,
NAXIS1 = 0, and NAXIS2 = 0.  Both cases are legal use of FITS, but
neither is well known outside its own user community.

Legacy FITS reading applications for optical astronomy often reject
the file if NAXIS != 2 in the PHDU.  The NOAO practice makes it
immediately clear that the FITS file does not conform to legacy
notions.  The DEIMOS practice permits a legacy application to see
header keywords even though it will see no image.

When NOAO software first saw a DEIMOS image it could not grok the
PHDU and did not recognize the file as a multi-HDU mosaic.
So, even within the optical community we already face a situation
where many applications will fail.  This should be considered a
natural consequence of evolution.

Rather than strive in futility for complete compatibility with
existing applications I think we should endeavor to produce new
standard practices for the use of NAXIS and NAXISn.

--
Steve Allen          UCO/Lick Observatory       Santa Cruz, CA 95064
sla at ucolick.org      Voice: +1 831 459 3046     http://www.ucolick.org/~sla
PGP: 1024/E46978C5   F6 78 D1 10 62 94 8F 2E    49 89 0E FE 26 B4 14 93