- From: Bijan Parsia <bparsia@cs.man.ac.uk>
- Date: Tue, 11 Mar 2008 09:12:00 +0000
- To: RIF WG <public-rif-wg@w3.org>
On 11 Mar 2008, at 08:51, Jos de Bruijn wrote:
>>> Well, the RDF model theory gives a precise definition.
>>> In short, like any other symbol, a blank node is mapped to an
>>> element in the domain. Then, satisfaction of triples is
>>> determined by the extension function IEXT: IP -> 2^{IR x IR}.
>> Interesting...this suggests that blank nodes are not treated as
>> variables (but as constants)?
>
> No. Both blank nodes and names are mapped to objects in the
> domain. The former are mapped using a blank node mapping; the
> second are mapped using an interpretation.
>
>> (I ask because this is in heavy debate in the OWL WG...if RIF is
>> going to be playing with RDF (e.g., generalizing it) it's probably
>> important, or at least helpful, to coordinate.
>
> Indeed. Although, RIF does not propose to generalize RDF as such.
> If we merely define interoperability with respect to a superset of
> RDF.
[snip]
I think I understand your point, but I think this may be a
distinction without a difference. I guess the minimal constraint
would be that the generalization isn't incompatible with any future
(or likely) sanctioned generalization. That seems to be met (though,
I suppose some future RDF could change the meaning of BNodes for
predicate positions...but that doesn't seem likely). So, technically
I'd personally say that you are in the clear. OTOH, the mere
existence of such an account might be taken as evidence that the W3C
endorses it or that it presents a likely roadmap.
Meh. I personally don't care, but given how surprised I've been by
some reactions to simple proposals, I guess I'm a little gunshy.
>> 3 is akin to what sparql allows, yes? I guess that provides *some*
>> reason.
>
> I guess so.
At best it gives a political reason, and perhaps not a good one. Dunno.
Cheers,
Bijan.
Received on Tuesday, 11 March 2008 09:10:06 UTC