[fitswcs] NAXIS vs. WCSDIM

[William Pence]

After rereading the recent postings in this thread, I have come to agree
with the NOAO contingent that it makes sense to decouple the NAXIS value
from the number of coordinate axes in the image.  As I see it, the original
primary purpose of the NAXIS keyword was simply to tell primitive FITS
readers how many NAXISn keyword immediately follow it in the header.  The
product of these NAXISn values is used to compute the size of the data array
that follows the header.  This certainly allows degenerate axes with NAXISn
= 1, but I don't think it should be required that NAXIS be equal to the
number of coordinate axes.  I would argue instead that software should 
assume a default value of NAXISn = 1  if the NAXISn keyword is not
physically present in the header (i.e., for  NAXIS < n < 999) which is at
least the mathematically correct assumption. (Except in the degenerate case
where the NAXIS keyword itself = 0).

I don't have a strong opinion about requiring a WCSAXES keyword to specify
the number of coordinate axes, but it is probably safer to include this
keyword as a consistency check of the other WCS keywords.

-Bill
-- 
____________________________________________________________________
Dr. William Pence                          pence at tetra.gsfc.nasa.gov
NASA/GSFC Code 662         HEASARC         +1-301-286-4599 (voice)     
Greenbelt MD 20771                         +1-301-286-1684 (fax)