Vocabularies: next steps

[Frederic V. Hessman]

> The outstanding issues appear to be:
>
> The format of the concept labels (case and character set)
Careful!

Namespace labels (e.g. http:www.ivoa.net/thesauri/IVOAT#21CmLine):

	Case:  
MixedCapitalizationByOnlyRemovingNonAlphabeticAndNumericCharactersSoThat 
MercuryWillWorkTwiceButOtherwiseVerySimpleToImplement			


	Set: a-Z, 0-9

prefLabels (e.g. "21-cm line"):

	Whatever is used by whoever does it.  E.g. IAU93 and AOIM have  
spaces and things.

The namespace label format should be an IVOA recommendation/standard!


> The grammatical number of the concept names (singular or plural)
Singular, please! - it's a real pain to use the formal system of  
singular concepts and plural countables and I agree that singular  
should make the vocabulary simple to use

> The number of top concepts in the IAU thesaurus

Huh?  The IAU thesaurus is the IAU thesaurus.  If "top concepts" are  
defined either as 1) not having a BT or 2) having a NT, then the  
number is already fixed.  Basta.  So is this really a question about  
definition of "top concept"?  Solution #2 gives you fewer top  
concepts but #1 feels better: it gives you the maximum number of  
potentially top concepts.

If we're talking about what I've called IVOAT, then the more work is  
done, the fewer the top concepts.  As Rob has pointed out, a lot of  
us are only in this for the vocabulary and I'm doing the BT/NT/RT  
stuff only to make all the potential critics feel good about things  
in general.


> The number of vocabularies we intend to produce (in particular  
> whether we produce a pair of `IAU' thesauri, including a corrected  
> and updated one, and which UCD vocabulary we use), and which  
> interrelationships we plan to publish

Vocabularies:

	IAU93, AOIM, UCD1, A&A certainly (since they're easy)

	IVOAT hopefully (since this is hopefully the most useful)

Interrelationships:

	Tricky question:  we don't want to refer too much to IAU93, because  
the suggestion will be that it's useful (which it really isn't) and  
UCD1 really doesn't cover very many concepts contained in the above  
vocabularies.  Stationary targets like the first list are admittedly  
much easier to do, but I've already started to connect IVOAT and  
UCD1, which is a good exercise since they are only partially  
matchable.  IAU93 and IVOAT are so closely related - even with the  
syntactic and content cleanups - that one could automate that  
connection without too much trouble.


> Which namespace we use
I vote for a long-term short namespace directory (e.g. www.ivoa.net/ 
Thesauri/) rather than a short-term long one.  There are going to be  
a number of vocabularies added with time, I'm sure.

> The WD which documents this
I agree we need a very short document saying SKOS + a few standards  
(see above).

> How we manage the development and release of the vocabularies
>

If SKOS is the standard, then we put the SKOS documents in  
www.ivoa.net/Thesauri/ with good names (e.g. IVOAT-2007-11-26.rdf)  
but manage the development using text or N3 via a wiki or something  
(ya'll are the experts here).  I've tried out an automatic  
documentation script used at http:/www.astro.physik.uni-goettingen.de/ 
~hessman/rdf that I'd be glad to contribute and Douglas has produced  
a totally different set of docs using standard tools (probably a  
smart thing to do) but without the blunt bare list of topics/tokens  
(the best thing to have).  No reason why there can't be several  
different means of perusing the vocabularies, depending upon your  
needs and tastes, so the IVOA would only have to have links to docs  
of IVOA-supported or -favored vocabularies and not have to support an  
official set of vocabulary docs.    After all, we ultimately want  
each group to support their own vocabularies.   The IVOA could also  
strongly suggest a minimum amount and type of documentation (e.g.  
some fast bare-bones listing and some view showing the links to BT/NT/ 
RT's).

Rick