Re: Agenda for telecon -- May 29 2003

>>Jos De_Roo wrote:
>>
>>>
>>...
>>>but it would if owl:Class was replaced with rdfs:Class in the semantics.
>>>]]
>>>
>>>The entailment is perfectly OK in OWL Full
>>>so the trouble is that we have no test case
>>>to show what would break if we drop owl:Class
>>>
>>
>>
>>Moreover, if we did, it would be a bug in our design. Our design is
>>intended to prevent there being such a test case.
>>
>>The only part of our document where you can explore these issues is
>>the section entitled OWL DL in the rdfs compatible semantics. That
>>section does not relate to any of the conformance statements and is
>>hence essentially informative, since it has no impact on any
>>envisaged software.
>>
>>Jeremy
>
>There is something wrong here.  Why would dropping this be a bug in
>the design? The design depends on there being a clear distinction
>between owl:Class in DL and rdfs:Class, so it seems odd that we
>cannot articulate that difference.
>
>Maybe Ive been out of the semantic loop for too long, but it seems to
>me that the difference should be articulable in OWL Full precisely by
>owl:Class meaning the same there as it does in OWL-DL, and it being a
>tautology that owl:Class is a proper subset of rdfs:Class. Any other
>design will guarantee that OWL DL and OWL Full are not interoperable,
>since the OWL vocabulary will have different meanings.

My parser maybe failed, but I think that
owl:Class a-proper-subset-of rdfs:Class
is inconsistent with
owl:Class owl:equivalentClass rdfs:Class
There must be one class concept in OWL
Full, so I believe the latter is needed.

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

Received on Thursday, 29 May 2003 20:49:31 UTC