[iaufwg] Revised FITS Registry Proposal

[William Pence]

Dear IAU-FWG members,

I've delayed replying to the few comments about the revised FITS registry 
proposal, because many of you were attending the Virtual Observatory Interop 
meeting last week (at least 6 IAU-FWG members were on the List of 
Participants).  As an aside, I think it is a very good sign that so many of 
us  are actively involved in VO activities and are keeping informed about 
developments that may impact the use of FITS in the future.

Getting back to the Registry proposal, I hope more of you will indicate here 
whether you think the current proposal is ready to be implemented (or not). 
  I don't want to proceed unless at least a reasonable fraction of you think 
it is a good idea.

In the meantime, here's my response to the previous comments.
Lucio Chiappetti wrote:
>
> I've read the new proposal, and it's mostly fine with me, particularly the 
> procedural parts (section 4 and 5).
> 
> I have a few comments on the documents described in section 3.
> 
> - for the "main" document in item 1, there is of course the "picky"
>   question of supporting or not a proprietary format (MS Word) vs
>   other. I suggest that documents can be accepted IN INPUT in those
>   format (one could also add OpenOffice sxw, but it's so easy to
>   generate the PDF straight from OO that it's not necessary), but
>   they will be made available in the archive ONLY as text or PDF
>   ...

It would be valuable to have the source document available, regardless of 
it's format, in case it is necessary to make revisions to it in the future. 
  Whenever possible we should try to provide both .ps and PDF format 
versions on the web site for download.

> - for the sample FITS files (item 3) one could also consider an URL
>   pointing to the site of the proponent (specially for intrinsically
>   bulky files)

URLs to larger sets of files would be useful, but since there is no 
guarantee that the URLs will be maintained indefinitely in the future, it is 
important to always have at least one example file available directly from 
the Registy web site.

>> While I realize that this version does not completely address all the 
>> issues that have been raised (especially regarding a XML definition of 
>> the convention),
> 
> - this is covered by the interim item 4 ...
>   ... but I believe also its current simpler replacement (item 2)
>   needs some standardization. I mean, we should at least provide a
>   recommended template for the text file in item 2 ... specially if
>   we want it also machine readable in some automated way. 

Once a couple conventions are entered into the Registry, they could serve as 
examples for future submissions.  I'm not too concerned about making the 
files machine readable initially, but this is something we could evolve 
towards over time.

> Another topic is : how do we start ? We just wait for inputs to come in, 
> or do we send "individual solicitations" to the authors of main known 
> existing conventions ?

I'm sure the IAU-FWG members can produce enough submissions to get started 
with (I could submit the CHECKSUM and the TLMINn/TLMAXn keyword conventions 
for instance).

-------------------------------------------------------------------
Walter Jaffe wrote:
> Does the requirement to list "a short definition of every keyword that is used in the 
> convention that would be suitable for inclusion in a FITS keyword dictionary." 
> also imply the requirement to define the columns in a FITS BINARY TABLE?  In many applications
> the description of the columns (names/formats/physical meaning) is the most important
> part of the format specification.

If the information about the columns is readily available in the main 
document, then there may not be much need to ask for the same information 
again in a separate text file.   I can see an immediate need for a concise 
definition of the convention and of all the keywords, however, which is why 
I think that information should be required.

----------------------------------------------------------------
Finally, as a reminder, here is my original message:
> Dear IAU-FIT Working Group members,
> 
> I would like to return to the issue of establishing a registry of FITS 
> conventions, and try to push ahead to get something implemented in the 
> near future.  I've revised the proposed rules governing the registry at 
> http://fits.gsfc.nasa.gov/iaufwg/fits_registry.html.   Compared to the 
> previous drafts, the rules in this version are somewhat less formal, and 
> hopefully will make it easier for people to prepare and submit 
> conventions to the registry.  For reference, you can view the previous 
> versions in the files fits_registry_v1.html and fits_registry_v2.html.
> 
> Please review this in the next week or 2 and reply with any comments or 
> suggestions.   While I realize that this version does not completely 
> address all the issues that have been raised (especially regarding a XML 
> definition of the convention), I hope there is enough of a consensus to 
> proceed with something close to what is proposed here.  I suggest that 
> we try implementing this on an experimental basis, and then reassess how 
> well the process is working after a few conventions have gone through 
> the whole process.  If anyone seriously objects to implementing this now 
> for some reason, please let me know!
-- 
____________________________________________________________________
Dr. William Pence                       pence at milkyway.gsfc.nasa.gov
NASA/GSFC Code 662       HEASARC        +1-301-286-4599 (voice)
Greenbelt MD 20771                      +1-301-286-1684 (fax)