[fitswcs] Keywords for other FITS image representations

[Mark Calabretta]

On Mon 1998/08/17 15:11:00 -0400, William Pence wrote
in a message to: fitswcs at fits.cv.nrao.edu

>I've changed the notation used in the table, and fiddled with
>a couple of the keyword names themselves, so please review it 
>carefully to see if you have any objections.

Looks good to me.

>   Leading zeros for the projection parameter, kk, are only 
>   required for the pixel list representation.

(BINTABLE also, as in your table.)

>2. The version number (A-Z) is optional in all cases EXCEPT
>   for BINTABLE vector represention where it must be present
>   to separate the axis number from the column number. 

Do we need some rule about what version number is implied if "v" is omitted.
For example, for two versions in one primary header should you write

   LNGPOLE & LNGPOLEA

or

   LNGPOLE & LNGPOLEB? 

>5. Currently, the maximum number of allowed projection parameters
>   is 99.  Is this sufficient?  The root name could be
>   changed to support up to 999 parameters.

I believe some others (e.g. Doug) want up to 999 but hopefully only in the
primary header.

>k     = projection parmeter number,  1 - 9,  10 - 99
>kk    = projection parmeter number, 01 - 09, 10 - 99

G&C allows k = 0 for the zenithal polynomial projection (ZPN, only) since it
was a natural choice in that case.  k=0 is pretty-well ingrained into WCSLIB
at this stage, e.g. in the C library PROJP0 corresponds to p[0].  Indexing in
the FORTRAN library also starts from 0.


Cheers, Mark