Re: SemWeb terminology page

> On Fri, Dec 03, 2010 at 11:34:52AM -0500, Jeff Young wrote:
>> As an informal term, I don't think "controlled vocabulary" is that bad
>> from a semantic web perspective. We just have to be careful with the
>> definition.
>>
>> According to the OWL Web Ontology Language Guide:
>>
>> "In OWL the term ontology has been broadened to include instance data."
>> <http://www.w3.org/TR/2004/REC-owl-guide-20040210/#owl_Ontology>
>>
>> In other words, the semantic web world shouldn't balk at the informal
>> notion of "controlled vocabulary" as long as they are represented based
>> on OWL (e.g. SKOS).
>
> We're getting off on a tangent a bit here, but the definition
> at [1] says: "An OWL ontology may include descriptions
> of classes, properties and their instances."  It doesn't
> actually say "OWL classes" and "OWL properties" - and for
> that matter, it only says "may"!  I'm curious whether
> formal definitions of "ontology" explicitly require OWL -
> or explicitly _exclude_ sets of (non-OWL) RDF properties and
> classes.  No time to chase this one right now...
>
> Tom
>
> [1] http://www.w3.org/TR/2004/REC-owl-guide-20040210/#Owl_Ontology_definition
>


Don't forget that this is the definition of an "OWL ontology". I think it is in fact quite self-referential: an "OWL ontology" is a dataset defined with the constructs brought by OWL, and that's it :-)
But anyway, that's not important, I don't think the SW community would balk on "controlled vocabularies" as Jeff said.
On the other end, I'd be against using "ontology" for Group 2 since, as you put it, "ontologies" (which feature of course OWL ontologies) may include *any kind* of individuals (using owl:Individual), which brings us too far. Group 2 should focus only on these things which would give rise to (RDFS or OWL) classes and properties, i.e., "RDF vocabularies".

Antoine

Received on Saturday, 4 December 2010 10:40:47 UTC