SUO: Re: W3C approves RDF and OWL as recommendations
Danny, Adam,
Danny, thank you for the rationales for the use of XML. Although I still
think that the use of a high-level notation supporting URLs is better and
easier from any viewpoint I can think of (knowledge entering, knowledge
correctness/exploitation/sharing, inference engine development, ...),
your list is important to me.
> rdf:Property rdf:type rdfs:Class
I agree with that one, and I can't retrieve the part of the documentation
that led me to think that an instance of rdf:Property cannot be a subclass
of an instance of rdfs:Class even in OWL Full, so I may have been fooled
by the documentation about OWL DL. Sorry for my over-confident assertion.
> Hmm, I'm not sure about this - if you're wanting the construct to behave
> as a relation (and follow appropriate rules) under some circumstances,
> and behave as a class (or instance of a class) under others,
No, roleTypeDomain is only a relation type.
> you intend to define some general semantic transformation for every
> class to spawn a new relation?
Every class in parameter of roleTypeDomain, yes.
> If so, what's the mechanism? Is it some sort of procedural macro?
I'll try to write the KIF definition of roleTypeDomain tomorrow (it
is far too late for me today).
> Does the user see the generated axioms when browsing a knowledge base?
If they are explicitely generated, yes. If the definition of
roleTypeDomain is only used by the inference engine during inferencing
(and the intermediary results are not kept), no.
Philippe