[fitsbits] Primary & Alternate WCS Keyword Order

[William Pence]

My preference would be to remove entirely any requirement or 
recommendation about the position of the WCSAXESa keywords.  As Mark's 
previous email explained, the present positional requirement serves no 
useful purpose for software that has to parse and interpret the WCS 
information in the header.  Also, the WCSAXES keyword is the only 
non-mandatory FITS keyword defined in the Standard that does not follow 
the general requirement (in section 4.1.1) that "keywords may appear in 
any order...".  I don't think there is sufficient justification to make 
the WCSAXES keyword the sole exception to this rule.

-Bill

On 6/29/2012 10:54 AM, Lucio Chiappetti wrote:
> On Sat, 30 Jun 2012, Mark Calabretta wrote:
>> On Fri 2012/06/29 10:55:06 +0200, Lucio Chiappetti wrote
>
>>> What Bill did for the verifier (excerpt below) appears consistent with the
>>> LETTER of the standard, which talks only of the position of WCSAXES and
>>> not of WCSAXESa
>>
>> That's only because WCSAXES is introduced in Paper I before alternate
>> representations are introduced, so it should follow that what applies
>> to WCSAXES also applies to WCSAXESa.
>
> Well ... that the distinction between the LETTER and the SPIRIT of the
> standard :-)
>
> If we'd go too deeply in a literal interpretation we'd have WCSAXES shall
> precede all WCS kwds (including WCSAXESA or WCSAXESK), but also WCSAXESK
> shall precede all WCS kwds (including WCSAXES itself and WCSAXESA !)
> and WCSAXESA shall precede all etc. etc. !!!
>
> If instead we say that the spirit is that each WCSAXESa precedes its own
> WCS kwds (as recommendation) and only WCSAXES precedes all, we could claim
> the standard is complete, and the main s/w (fverify, WCSlib etc.) is
> compliant.
>
> All your suggestions are extremely sensible ...
>
> ... I was just trying to understand whether we should make an amendment to
> the standard at this stage (which will be a couple of lines on a 52 page
> document, and will introduce a deviation from the A&A published stuff,
> both Pence et al 2010 and WCS Paper I ... which is "incorporated by
> reference")
>
> ... or we could pretend that the standard is clear enough, and issue only
> some sort of explanatory note
>
> ... reserving the amendment for the next round (3.1 or 4.0) whenever it
> will be (not too soon hopefully).
>
> Since I see it's already Saturday down there ...
> ... nice weekend to everybody.
>
> _______________________________________________
> fitsbits mailing list
> fitsbits at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/fitsbits