[Semantics / Vocabularies] Proposition for a IVOA vocabulary main page

Markus Demleitner msdemlei at ari.uni-heidelberg.de
Thu May 24 14:43:25 CEST 2018


Hi Mireille,

On Wed, May 23, 2018 at 11:42:50PM +0200, Mireille Louys wrote:
> http://wiki.ivoa.net/twiki/bin/view/IVOA/VocabularyTestPage
> 
> Dear Semantics ,
> Here above is the link to a prototype page  that has been proposed to the
> exec commitee recently.
> The goal is to gather all vocabularies used in the IVOA framework and manage
> the various updates of these collections of terms
> uniformely.
> 
> The technical aspects for providing a central page at the IVOA main pages
> are discussed  with Exec Committee and the Media Group .
> 
> I would very much appreciate your feedback as developers or users on this
> proposal  in terms of :
> 
>  * lisibility: is it understandable ?
>  * usage :
>      o how you would consider using it  or refer to it
>      o tools you may like to plug on these resources
>      o etc ..

I think technically that should be some bit of XML with a style sheet
on the doc repo, curated by them, but I guess this proposal isn't
about the curation tech, really.

More importantly, I think this needs to be on the actual vocabulary
level, so the table structure would be

<name> | <vocabulary URI> | <description> | <working group>

which, for the vocabularies used by VOResource at the moment, would
work out to

resources content levels | http://ivoa.net/rdf/vocab/content_level | What is the intended audience...      | Registry
roles of evolution dates  http://ivoa.net/rdf/vocab/date_role      | What sort of action has been taken... | Registry


Further, I don't think the notion of version is useful for
vocabulary, as it's expected to "live" (i.e., is updated fairly
regularly), and there can't really be major versions. As suggested in
the W3C best practices, the evolution is tracked through dates, and
redirects are issued so clients can figure out which exact vocabulary
they're using (see datalink and the VOResource vocabulary as
examples).

I'm not quite sure at this point where UCDs fit into this -- I seem
to remember there once was a drive to make an RDF version of them.
Perhaps that's what should be done?  I'd say, though, that we should
first see how things work out for the new-style vocabularies.

           -- Markus


More information about the semantics mailing list