RE: ACTION-129 and ACTION-132

[related to ISSUE-67 and ISSUE-81]

Bijan Parsia answered to me:

>> I remember having heard several times that many people would like
>> to change
>> the semantics of RDF reification, or even drop it completely from
>> the RDF
>> spec.

[...]

>I'm not sure how it can be "dropped" per se. Surely legacy
>considerations will force it to be maintained.

That's a point.

And regarding the possibility of a semantic change: As I have stated in an
earlier discussion, the actual RDFS semantics of RDF reification is very
weak and straightforward. It's effectively: rdf:subject, rdf:predicate, and
rdf:object are all rdf:PropertyS, and rdf:Statement is an rdfs:Class. This
is harmless, and I wouldn't see an immediate reason to make stronger or
different assumptions about a reified statement in general. Extending the
semantics for special purposes, like in the case of negative property
assertions in OWL, is then always possible without big problems.

What will have to be changed or dropped is the strange "MAY" semantics in
[1]. I have heard people believing that this /is/ the actual semantics, but
this is /not/ the case. Changing or dropping this MAY semantics will,
however, not have any effect on OWL 2 Full, since the semantics of OWL 2
Full will (most likely :)) not depend on it.  

(But this all doesn't mean that I would have a good feeling about using RDF
reification in OWL...)

Cheers,
Michael

[1] <http://www.w3.org/TR/rdf-mt/#Reif>

Received on Wednesday, 7 May 2008 11:31:16 UTC