Re: WOWG: minutes Jul 10 telecon

Jim Hendler wrote:
>At 10:21 PM +0200 7/10/03, Jos De_Roo wrote:
>>>  > 22. RDFCore Comments on OWL Reference
>>>  > ACTION: Frank van Harmelen to respond to one open element
>>>  > vs. owl:Class)
>>>  WITHDRAWN; transferred to PatH
>>>  ACTION PatH: draft rationale for rdfs:Class vs. owl:Class situation
>>>  (less controversial than last time)
>>I was joining late and missed the disicussion; my bad
>>What I was also trying to say at the end of the telecon
>>(but had no chance due to sound prloblems) was that
>>I am stuck at the following:
>>When D is a consistent OWL Lite/DL document then
>>it is not necessarily a consistent OWL Full document
>>for example the document
>>owl:Thing owl:oneOf _:x.
>>_:x rdf:first eg:a.
>>_:x rdf:rest _:y.
>>_:y rdf:first eg:b.
>>_:y rdf:rest rdf:nil.
>>seems to be DL consistent but Full inconsistent.
>>We try to make our assumptions explicit in a "global" sense
>>(using URI's, triples, implications and some such)
>>and I can't see those for owl:Class and owl:Thing
>>(seems to me that everybody could mean it's own local thing)
>Jos et al -
>  I'm a bit confused on the issue of full inconsistent, in part
>because I haven't been tracking the RDF Core semantics -- supposing I
>did the following in RDF
>rdf:resource rdf:type eg:a.
>or - suppose I said:
>rdf:resource owl:equivalentTo owl:nothing.
>these would surely cause problems as well, wouldn't they?  They seem
>to be syntactically legal RDF - does the semantics rule them out in
>some way?  If not, wouldn't our documents where we change owl:Thing
>(in Full) be he equivalent of changing rdf:resource since they are
>equivalent?  Guess what I'm really asking what "Full inconsistent"
>means in this sort of case...

I meant to say "as an OWL Full document it is not consistent"
(see also in Test for various otest:InconsistencyTest cases)
and indeed your example

rdfs:Resource owl:sameAs owl:Nothing.

is also such a case (which we can prove).

Now, having an OWL document that is both consistent
and inconsistent is something that I must object to
(in fact I am trying to make sense of owl:Class and
owl:Thing as "global" W3C owned URI identifiers...)

Jos De Roo, AGFA

Received on Friday, 11 July 2003 19:19:07 UTC