[fitsbits] updates to the FITS standard document

[Tim Pearson]

Dear Lucio,

I agree with Calabretta and Forveille that it is far from clear that these "conventions" have any place in the standard. There are good reasons why they have not been made part of the standard in the past. Concurrent three-week public comment periods for six separate changes to the standard will not give us enough time to fully air all the issues involved, and I urge that this be delayed until after the IAU reorganization. I note that the reduction of the comment period to three weeks is in violation of the IAU FITS Working Group's own rules: "This public discussion period will normally last for at least 4 weeks, and the EC may extend it as necessary to allow time for the discussions to reach a conclusion." This is a case where extension of the period is needed, not curtailment.

Some of the changes require FITS readers to make changes to their code to conform to the new standard; e.g., readers that are unaware of the "Continued string keyword" convention can corrupt headers while processing a file. This is dangerous. At the very least the version of the standard that the reader needs to follow has to be recorded somewhere in the header, e.g., by changing the value of the SIMPLE keyword.

Tim
--
Timothy J. Pearson
Research Professor in Radio Astronomy
Mail Code 249-17, Caltech, 
Pasadena, California 91125, USA
E-mail:    tjp at astro.caltech.edu

On Jun 20, 2015, at 12:57 AM, THIERRY FORVEILLE <thierry.forveille at ujf-grenoble.fr> wrote:

>> 
>>> 2) a task force within the IAU FITS WG has been considering the inclusion
>>>    of a number of registered conventions as part of the standard.
>>> 
>>>    For 6-7 of them we have (or nearly have) a text to be included in
>>>    the FITS Standard Document. According to the rules, this will be
>>>    subject to a Public Review Period here on FITSBITS, with the goal
>>>    to have the IAUFWG vote (on each convention) during July.
>>> 
>>>    Considering that the conventions are in use since several years
>>>    (so their operablity and interoperability is surely proven), and
>>>    were discussed when registering the conventions themselves, to
>>>    speed up things the Public Review Period will be reduced to 3 weeks.
>> 
>> Noting that the process of registering conventions was concerned only
>> with ensuring the adequacy of their documentation, not the adequacy of
>> the conventions themselves, I just want to clarify whether we are being
>> asked to consider adopting these conventions, unaltered, as standards,
>> or whether they are subject to change.
>> 
>> In particular, I note that the INHERIT convention attracted more than
>> the usual amount of discussion on fitsbits and iaufwg, with many
>> respondents expressing strong reservations about it.  At the urging of
>> its proponents, INHERIT was ultimately registered as a convention on the
>> basis of its wide usage within a particular sector of astronomy, not
>> necessarily as the model for a new component of the standard.
>> 
>> With eight years worth of hindsight, the links accessible from
>> http://fits.gsfc.nasa.gov/registry/inherit.html make for interesting
>> reading.  It seems clear now that those members of the FWG who took an
>> interest should have been allowed to record a short summary (say 150
>> words) of their arguments for or against the conventions.
>> 
> I share Mark's concerns, and at least at this stage I intend to vote NO 
> on the less obvious of the conventions becoming part of the standard. 
> Every addition to the standard has a cost in FITS readers that must
> be balanced against what it gains us, and some of the conventions
> don't make my cut.
>