[iaufwg] FITS Registry Proposal

[Lucio Chiappetti]

I was away in the last few weeks so I collect together below my comments 
on some of the messages circulated.

--------------------------------------------------------------------------
On Wed, 15 Mar 2006, Doug Tody wrote:

> Our main purpose with a registry of conventions should be to list and 
> describe these conventions so that people know about them.

  I agree

> I don't think we should try to control the quality of conventions or we
> are starting to duplicate the standards process.
  
  except in case of clear nonconformance (which I expect to be rare or
  nil)

> With these goals in mind I think the proposed process described is too
> strict. [...] My suggestion is to reduce the requirements for
> registration to the minimum, e.g.: the necessary information is
> provided, the convention is adequately documentated, and [...] Other
> issues, for example if two conventions provide similar functionality,
> should not be our concern

  I agree

--------------------------------------------------------------------------
On Wed, 22 Mar 2006, Preben Grosbol wrote:

>2) Some kind of indication of usage would be good e.g. historic, current,
>      
>3) Some of the 'approval process' is not relevant for conventions already
>   in use.  For new ones it may be a little heavy.  One could imaging
>   rather local ones which are nice to know about and widely used ones
>   going through the full thing.

  I agree also on those two statements

--------------------------------------------------------------------------
On Thu, 23 Mar 2006, Mark Calabretta wrote:

> Another example: the ATNF's internal format is called RPFITS which
> actually is not FITS [...] But RPFITS looks a lot like FITS except that
> it begins with SIMPLE = 'FALSE', has a block size of 2560 bytes, uses 
> VAX floating point representation, and a few other things, but otherwise
> looks just like FITS.

Hmm... well ... yes, but ...
I also defined for BeppoSAX my own "XAS" format which is not FITS (native 
binary representation, binary keywords) but mappable to it (concepts of 
header keywords, binary tables)

Or ESO has a format called "short FITS" which is not FITS.

I do not think it is appropriate to handle those or similar cases in the 
registry. Except perhaps for the cases which have "FITS" in their name but 
are not FITS. But they should just be LISTED in a SEPARATE page, not 
subject to any "approval" process, and documented ELSEWHERE.


--------------------------------------------------------------------------
Incidentally, this item reminds me of a small detail. Should we want to 
provide some "automatic" way to classify a FITS file according to the 
conventions supported ?  I'm thinking of the "file" command and its 
"magic" file. Should we encourage convention mantainers to supply a 
"magic" description or even help them to build one ?


--------------------------------------------------------------------------
On Wed, 22 Mar 2006, Andreas Wicenec wrote:
> On 21.03.2006, at 23:26, William Pence wrote:
> 
> > 1.  Bob and Doug both suggested opening up the registry so that proposed
> > conventions, or working drafts of conventions could be registered.  I

> I agree that there must be a clear distinction, but I still think there 
> should be a way to see all of them on a single page along with their 
> status ("proposed", "approved", ...). It looks like a fairly simple 
> web-application could actually accomplish both (see below).

I also share Andreas' preference to be able to see all from a single page 
(table with clear status and historical flag). However, also considering 
the subsequent discussion on "nominated" vs "proposed" vs "approved". we 
should consider the following points :

 - this work will require a burst of initial work to "clear the backlog"
   of existing, historical conventions. In this case most of the 
   "instructory" work could possibly be waived because the conventions
   will already be widely used, documented and conformant.

   The main point is if 

   - we wait for an existing convention to be "submitted" to us, 
   
   - we "solicit" authors of existing conventions we know to formally
     submit them to us

   - we proceed ex officio to handle conventions known to us using
     available documentation

   In all this cases the convention could stay in the "submitted" status
   temporarily for a short time, formally, but soon move to the "approved"
   status (what if it fails approval ? could it remain in a "notified"
   status for reference ?)

 - for NEW (future) conventions probably a slightly more thorough 
   "instructory"  work will be required, sort of what Bill proposes in his 
   page. But we will be facing two different cases :

   - a convention already well tested and in widespread use is brought
     to our attention ("submitted" status) and after the, say, 4+4 weeks
     moves to "approved" or "notified" as above

   - a draft convention is submitted to us for comments, and we could 
     suggest adjustments before it is "published". This is an altogether 
     different path, and I'm not sure we really want or need to do this.

> > 3.  Bob pointed out the strong similarities with the process used in 
> >    the IVOA to promote standards, which is encouraging.  [...] I think 
> >    the FITS community would be more comfortable with the traditional 
> >    practice [...] proposed/approved conventions are just listed on a 
> >    simple web site.  At least I would prefer to get things started 
> >    this way, and postpone merging into the VO registries to some 
> >    future time.

I share Bill's feeling, anyhow the final choice should mainly be with the 
mantainer of the registry.

> > 4. Finally, Steve raised a number of practical issues about where this
> > web site would be located and who could add new information to it.  For
> > now, I volunteer to maintain this under the FITS Support Office Web site
> > at the HEASARC, which I already maintain.   

I believe that in practice we should find a way to distribute the burden.

It can be true that requiring people to upload convention "data" to some 
web server may discourage somebody to do so (although we should be used to 
submit papers to journals and even drafts to internal project repositories 
in similar ways), and will require a bit of maintenance, more if some 
security is required to verify authorized users.

Perhaps it could be easier just to provide an e-mail address in some form 
of a moderated list. The moderator (Bill ?) will "filter" the submissions 
(from spam and other mail) and move them either to the web site, or 
dispatch them to a "referee".

I believe it is practical that we (members of IAUFWG) in turn act as a 
referee to check formal aspects before opening the "discussion".

The referee verifies the formalities are OK, and bounces the submission 
back to the moderated list with some tag. The moderator can then move the 
submission to the web site for the discussion.

The result of the discussion will just be a change of the "status flag"
from "submitted" to "approved" or "notified", easy for the website 
mantainer.

----------------------------------------------------------------------------
Lucio Chiappetti - INAF/IASF - via Bassini 15 - I-20133 Milano (Italy)
----------------------------------------------------------------------------
The reasonable   man tries to adapt himself to the world
The unreasonable man tries to adapt the world to himself
Therefore all progress depends on the unreasonable man      (G.B.Shaw)
----------------------------------------------------------------------------
For more info : http://www.iasf-milano.inaf.it/~lucio/personal.html
----------------------------------------------------------------------------