FAIR vocabularies

[Markus Demleitner]

Dear Françoise,

On Wed, Dec 02, 2020 at 08:57:03AM +0100, Francoise Genova wrote:
> I have the feeling that our vocabularies are not so bad in terms of
> community-driven governance. Once again there is an issue with licence. I
> just asked the question and the answer is that something which is openly
> available on the web cannot be reused if it does not have a licence (I have
> to admit that I was expecting this answer because I got the same about data
> when I asked in in other contexts). A hot debate followed my question in the
> chat about the type of licence and as usual nothing is simple in the licence
> domain. When we discussed licences for data the decision of licences was
> clearly on the data provider side. For vocabularies maintained by the IVOA
> it should be in the hands of the Exec but there is a need to do a
> preliminary assessment before any decision.

With vocabularies governed by Vocabularies in the VO 2, one use case
is offline use, and hence we want the vocabularies to be
distributable with software.  To ensure that, we could use a software
licence, but then that only marginally makes sense with semantic
resources.  On the CC side, everything but CC-0 is incompatible with
software licences and would thus break the use case, and hence we
settled for CC-0 globally a while ago.

https://www.ivoa.net/rdf/ and the individual renderings (at the foot
of the HTML rendering, in dc:license triples for the RDF renderings;
desise doesn't include the declaration yet) of the vocabularies say
as much.

There is an exception for the UAT, which (IMHO regrettably) has
chosen CC-BY-SA.  That's not something we can fix, and so we can only
declare that and live with the consequences.


By the way, CC-0 isn't a licence, as it does not contain claims of
copyright[1] and hence there can be no licensor.  So, the right way
to state what we do is: "The IVOA distribute their own vocabularies
under CC-0".

> There are likely other issues than licences in our practices with respect to
> what emerges in the discussion on the definition of FAIRess for
> vocabularies.

Please do bring them up here as they emerge.  As we'd like to bring
the Vocabularies 2 specification to RFC soon: Is there a schedule for
these discussions?

         -- Markus

[1] Which, in collaborative resources that vocabularies often are, is
another big advantage, as there are no problems of copyright
transfer or its substitutes in jurisdictions that don't allow it.