[fitswcs] WCS documents

[Clive Page]

On Mon, 15 Oct 2001, Mark Calabretta wrote:

> It's not clear whether the "[]" units proposal is supposed to be machine
> readable or just an assist to human readers.  If machine readable then it
> would be contentious because it's questionable grammar to put semantic
> content inside comments.

I think there's a more general problem here, in that it's quite hard to
work out _how_ to extend the FITS Standard in a way which is compatible
with existing files and existing software (not to mention the existing
Standard).

Just as an example, several people have realised the need for FITS tables
to specify that columns of positional coordinates should be displayed in a
sexagesimal format, such as hh:mm:ss or +dd:mm:ss.  With a few others
(including Steve Allen and Clive Davenhall) I have been trying to work out
the best way of doing this but we have not yet come up with a really
satisfactory solution.

The ideal solution would be to add a new descriptor or two to the list
allowed in TDISPn, but files using this could easily break existing
software.  An alternative would be to append some extra characters to an
existing TDISPn value such as "Dw.d", but the Standard does not allow
extra characters here. Interestingly, it does allow something (undefined)
to be appended to a TFORMn specifier, which would provide a possible
mechanism, although it seems a less appropriate keyword to use, and might
still risk breaking some existing software which wasn't quite
standard-conforming.  Yet another possibility would be to specify
something in the comment field of TDISPn, <perhaps in angle brackets?>
which almost certainly would be compatible with existing software and
existing files.  But as Mark points out, it's not a nice way of providing
machine-readable as opposed to human-readable information.

-- 
Clive Page,
Dept of Physics & Astronomy,
University of Leicester,    Tel +44 116 252 3551
Leicester, LE1 7RH,  U.K.   Fax +44 116 252 3311