Re: owl:subClassOf ? [was: Re: working on it]

>On Thu, 2003-06-26 at 06:28, pat hayes wrote:
>>  Guys, I now have connectivity.
>
>:)
>
>>
>>  I have been working on a version of the semantics doc along the lines
>>  several of us discussed after last Friday's telecon. It is still
>>  incomplete, but I will keep a version here:
>>
>>  http://www.ihmc.us/users/phayes/RDF_Semant_Edit_Weak.html
>>
>>  I hope it will be reasonably final by Friday, though I may not have
>>  the proofs written out until early next week. At the time of sending
>>  this email, only section 4 has been rewritten and that not fully; the
>>  rule section hasnt been done yet.
>>
>>  In this, *all* the RDF/S semantic conditions are 'if' not 'iff', so
>>  the correspondence to the rules will be easy to establish, and the
>>  relevant closure lemmas easy to prove. The only extra condition added
>>  is that all classes are subclasses of rdfs:Resource, which I think is
>>  needed to make sense of the 'empty' domain and range conditions, and
>>  in any case is handled fully by rdfs 7a
>>
>>  The complete RDFS rules are now rdfs 1, 2, 3,4a,4b,6, 7a, 9,10,11;
>>  the correspondence to the semantic conditions is very clear.
>>
>>  I plan to discuss the 'extensional' versions and mention the extra
>>  rules they sanction (5,7b,8,12) but without any claim of
>>  completeness. Thus, the overall content will be similar to the last
>>  version except that a simple basic RDFS rule set can be complete and
>>  included in the normative spec in the way that will make DanC's heart
>>  lighter.
>
>I'd like to be sure we understand the relationship of rdfs:subClassOf to
>Owl.  As I understand things, the intent is that Owl would retain
>extensional semantics.  Is that correct.

Yes: at elast, OWL Dl certainly will. I am not sure about OWL Full as 
Peter has defined it.

>
>If so, then Owl want a stronger notion of subClassOf than that proposed
>in this document for rdfs:subClassOf.

It will want to impose stronger conditions on the meaning of that term, yes.

>
>Would that meant that the 'correct' thing for WEBONT to do would be to
>define owl:subClassOf as a subproperty of rdfs:subClassOf?

I no longer have any firm idea what 'correct' means in the W3C WG 
land. We defined the notion of 'semantic extension' carefully so that 
things like OWL could impose stronger conditions on RDFS vocabulary 
meanings as long as they didn't contradict the RDFS meanings. OWL 
already does this for domain and range, for example, and nobody seems 
to find that odd. It can do similarly for subClass and subProperty, 
seems to me, on a similar basis.

Pat

-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Thursday, 26 June 2003 18:45:58 UTC