[fitsbits] Five FITS Proposals

[Preben Grosbol]

On 24.10.2013 08:54, Thierry Forveille wrote:
> The proposals would put a moderate but still significant upgrade burden
> on existing readers, so I think need a better argued justification before
> I would seriously consider them. In other words, I would like to see a
> detailed rationale for each proposal and a description of some real  
> life problems that it solves.
I share the concern of possible overheads in expanding the FITS header
decoding rules.  One may argue that the 3 first issues could better
be solved by introducing a new 'METADATA' extension which would contain
a XML formatted character string.  This would have the advantages of
 - providing a very flexible definition of meta-data, and
 - only applications interested in decoding such data would need to.

For the FITSVERS keywork, the old argument against it is that it, in the
future, could be misused i.e. making FITS decoding dependent on the
the version number.  Thus, it was always viewed to be too dangerous.

Best regards,
Preben