[iaufwg] Appendix A, END keyword

[Mark Calabretta]

On Tue 2008/03/25 12:37:59 EDT, William Pence wrote
in a message to: IAU-FWG <iaufwg at nrao.edu>

Dear Bill,

>In my experience, most, if not all, of the common errors seen in FITS 
>files are clear violations of the original FITS definitions and are not 
>a result of changes to the FITS requirements that retroactively caused 
>existing FITS files to become invalid.

This is a curious way of phrasing it.  I would have said that no change
to the FITS standard can retrospectively cause any existing FITS file to
become invalid - the FITS standard itself would be at fault.  In fact,
this is potentially the situation we have now.

>The one case of a new 
>requirement that is commonly cited, i.e., the change made back in 1999 
>to *require* rather than only *strongly recommend* that the keyword 
>value and comment field be separated by a slash character, is largely a 

Why hasn't this known instance been documented in the appropriate
section of the standard, either directly in the text or as a footnote?
The technical committee has never provided a reason not to, it has only
stated that it doesn't think it worth the effort to find other
anomalies.

While on that subject, the technical committee is already well
acquainted with FITS 2.1 because that was where it started, and FITS 3.0
because that's what it produced.  Really all it needs to do is compare
FITS 2.1 with the original documents.

Supplementary questions: are copies of the earlier standards publically
accessible?  Are the latex sources for the earlier NOST-100 documents
available (i.e. for diff'ing)?

>non-issue, I think, because I can't remember ever seeing a FITS file 
>that doesn't follow this convention.  Even if there are some FITS files 

You can't speak for all of the FITS files ever written, which is what
would be needed to argue convincingly along these lines.

>created before 1999 that don't have the slash character separator, this 
>is the sort of minor infraction that most FITS reading software would 
>ignore anyway.

Likewise, you would need to be able to speak for all extant FITS
readers.

Regards, Mark