Re: classes as instances

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: Re: classes as instances (was Re: comments on RDF mapping)
Date: Fri, 02 Nov 2007 16:20:13 +0000

> touché
> 
> Peter dug about in the archives and found me saying:
> > One way of moving forward would then be to allow classes-as-instances in
> > all OWLs but to make (large owl) entailments that relate the two (like
> > this test case) not a Fast OWL requirement (and hence not an OWL Lite
> > requirement).
> > 
> > Jeremy
> > 
> 
> Personally I would be happy with this sort of solution - notice the 
> entailment still holds in Fast OWL (i.e. OWL DL), it is just not a 
> requirement for reasoners to find it (i.e. OWL DL reasoners are no 
> longer expected to provide the same completeneess guarantees). OWL DL 
> would simply be one of several suggested profiles.

My take on this would be slightly different, as you might expect.  The
entailment holds in OWL Full, but not in OWL DL.  OWL DL reasoners would
be expected to be complete for OWL DL.  OWL Full reasoners would be
expected to be as complete as possible - for some definition of posssible.

[...]

> Jeremy

peter

> PS As you may have noticed I have changed in many ways over the last 
> five years.

Oh?  I hadn't noticed any changes at all.  :-)

> PPS This presents me with a challenge to dig about in the archives for 
> an appropriate riposte citing a pfps message ... maybe another day

That should be PFPS.  :-)

> PPPS There probably is an archived message from me in which I challenge 
> the myth of the continuity of the self: that seems to have been a 
> repeating theme in my fragmented existence.

Received on Friday, 2 November 2007 16:42:05 UTC