FITS header values

[Stupendous Man <richmond@a188-l009.rit.edu>]

  I've just read the paper by Griesen and Calabretta, 
A&A 1996, which describes the World Coordinate System for
FITS files in excruciating detail.  I note from the
WWW page on WCS systems at NRAO

     http://www.cv.nrao.edu/fits/documents/wcs/wcs.html

that the system described in that 1996 paper, with
keywords like CRVAL1 and CDELT1, is now obsolete,
replaced by a new manner of denoting coordinates.  Well,
fooey to the new system, I say.  Let's continue with
the old.

  After looking carefully at some of Tom's Mark IV files
from the latest Disk Set 17, and the WCS document, I think
the appropriate values are:

        CTYPE1 = 'DEC--TAN'          / first axis (column) is Declination
                                     / note that there are two hyphens
        CRPIX1 =      1016           / middle pixel of array in col direction
        CRVAL1 =         0.0         / Dec of this middle pixel, in
                                     / decimal degrees (not sexigesimal)
        CDELT1 =         0.002126    / moving forward one column increases the
                                     / Dec by 7.65 arcsec = 0.002126 degrees

        CTYPE2 = 'RA---TAN'          / first axis (column) is Right Ascension
                                     / note that there are three hyphens
        CRPIX2 =      1015           / middle pixel of array in row direction
        CRVAL2 =       129.0         / RA of this middle pixel, in
                                     / decimal degrees (not sexigesimal)
        CDELT2 =        -0.002126    / moving forward one column decreases the
                                     / RA by 7.65 arcsec = 0.002126 degrees

  I would recommend that you _not_ use the CROTA1 and CROTA2 keywords.
They are meant to indicate a small angular rotation of the image
coordinates relative to (RA, Dec), not a large 90-degree or 180-degree
rotation.

  Note that the way the Mark IV is currently read out, the first
rows are Eastern, the last rows are Western: thus, increasing row
number DECREASES the RA.  On the other hand, the first columns are
Southern, the last columns are Northern, so increasing column
number INCREASES the Dec.  That's why I have different signs
for the CDELT1 and CDELT2 values.

  I placed values similar to those above into one of the Disk Set 17
image headers, and then ran the FITS image viewer "DS9".  The image
display had (RA, Dec) values which were approximately correct, so 
I'm confident that I managed to get the signs right this time.

  It doesn't matter to me what the raw image orientation is.  I would
say you should write the pixels in the most convenient way.  It's
easy to flip or rotate the image in any image display program to 
match your favorite chart.

                                       Michael Richmond