Re: Semantics of the schema.org data model (for integration with RDFS, OWL, etc.)

On 29 August 2012 17:15, Christoph LANGE <c.lange@cs.bham.ac.uk> wrote:
> Dear schema.org community,
>
> [trying again; first mail didn't go through]
>
> I have some questions about the semantics of the schema.org data model
> (http://schema.org/docs/datamodel.html).
>
> The documentation of the data model speaks of
>
> * types (with multiple inheritance)
> * properties (with multiple domains and ranges)
> * and finally there is the "Thing" type.
>
> The formal semantics of these is not defined, but am I right to assume that
> it is the same as the one of
>
> * rdf:type
> * rdfs:Class (or owl:Class, maybe in an OWL Full semantics?)
> * rdfs:subClassOf
> * rdf:Property
> * rdfs:domain
> * rdfs:range
> * owl:Thing
>
> ?
>
> My background: I am one of the authors of the draft ISO standard on
> "ontology integration and interoperability" (http://ontoiop.org), which
> specifies the Distributed Ontology Language (DOL).  DOL allows for
> integrating ontology modules written in different languages, with an
> extensible supply of ontology languages.  Integrating the schema.org data
> model as one of these "ontology languages" would allow for connecting
> structured data all over the Web (i.e. instances of the schema.org data
> model) to more expressive logical formalizations of the domains involved.
> This goes beyond the OWL ontology at http://schema.org/docs/schemaorg.owl:
>
> 1. We would support ontological descriptions not just of the standard
> schema.org schema, but of any non-standard extension.
> 2. DOL allows for such ontological descriptions to be given not just in OWL
> but, e.g., in Common Logic, F-logic, etc. (i.e. possibly languages that are
> more expressive than OWL, and that do not necessarily natively support
> IRIs).  This is e.g. useful for validation purposes: "the lowPrice of an
> AggregateOffer must be <= the highPrice".
>
> But, of course, integrating the schema.org data model into the DOL framework
> requires a formal semantics, which is why I'm asking these questions.
>
> Cheers, and thanks in advance for any help,

Interesting. Well, more or less "what you see is what you get". We
don't have any formal axioms etc., beyond basic type hierarchy. The
notions of range and domain we use are pretty soft too. If you find
opportunities for formalization (such as the lowPrice example), it
would be interesting to see those. But for now, I think it's best to
treat schema.org's schema as a pretty simple RDF vocabulary. So
schema.org extensions would also just be treated as  RDF/S or perhaps
OWL. The schema.org search engines don't claim to do anything with 3rd
party vocabularies that extend schema.org, but others are always
welcome to make new uses of the vocabulary.

cheers,

Dan

Received on Wednesday, 29 August 2012 17:34:18 UTC