[fitsbits] Fits Table Entries Referring to Keywords

[William Pence]

Randall Smith wrote:
> 
> What I'd like to do is change this to store the complete,
> journal-quality references in the header and then give a lookup value
> in the file.  Ie, on a particular line give the reference as "ZS87",
> and then in the header somehow list
> ZS87 = "Zhang,~H \& Sampson,~D.  1987, ApJ, 581, 183"
> 
> What I'm wondering is, is there any kind of standard for this?  I could
> make the ZS87 a keyword with a long string value equal to the full
> reference.  However, I'd like to have complete freedom in making up
> the lookup value, and that means there could be a conflict between the
> lookup value keyword and a real, well-defined FITS keyword.   Or I
> could use COMMENT statements which I could then parse, but this
> information is an important part of the file, and relegating it to
> "COMMENT" status doesn't seem right either.

A more straightforward and efficient way to do this would be to use a
variable-length string (TFORMn = '1PA') for the column containing the
reference.  If multiple rows of the table have the same reference, they
could all point to the same address in the FITS heap area.

Otherwise, if you use a single indexed keyword root name for all the
references (e.g. 'REFn' or 'BIBn') then you would not have to worry about
conflicts with other keywords:

REF1 = "Zhang,~H \& Sampson,~D.  1987, ApJ, 581, 183"
REF2 = "...

I agree that hiding important information inside COMMENT keywords is not a
good idea. 

-Bill Pence 
--
____________________________________________________________________
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)