W3C home > Mailing lists > Public > public-owl-dev@w3.org > October to December 2007

Re: [OWLWG-COMMENT] ISSUE-55 (owl:class)

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Thu, 20 Dec 2007 03:57:00 -0500 (EST)
Message-Id: <20071220.035700.186614277.pfps@research.bell-labs.com>
To: schneid@fzi.de
Cc: alanruttenberg@gmail.com, public-owl-dev@w3.org

From: "Michael Schneider" <schneid@fzi.de>
Subject: RE: [OWLWG-COMMENT] ISSUE-55 (owl:class)
Date: Thu, 20 Dec 2007 09:40:41 +0100

> Hi again, Alan!
> 
> Alan Ruttenberg wrote on Monday, December 10:

[...]

> Looks like things have changed, because Peter has found a way to make the
> difference "perceivable":
> 
>   <http://lists.w3.org/Archives/Public/public-owl-wg/2007Dec/0227.html>
> 
> The trick is to force owl:Thing to be a /finite/ class (in Peter's ontology
> owl:Thing happens to equal the union of the two defined classes, which are
> both finite). In OWL-Full owl:Thing is always infinite, while it may be
> finite in OWL-DL. 

The only thing new in my example (which isn't really new), is that it
doesn't mention owl:Thing at all.  I put the example together this way
to counter a claim that it would be easy to recognize those ontologies
that did "weird" things like fiddle with owl:Thing and then refuse to
"repair" these ontologies. 

> @Peter: However, please consider Jeremy's ISSUE-73 in this context:
> 
>   <http://www.w3.org/2007/OWL/tracker/issues/73>
> 
> What I cannot tell is whether this is a show stopper for the "rdfs:Class
> repair" or not. I wouldn't believe so, but I have to ponder about this
> further.

This may be related in that if the semantics of OWL DL are changed in
this way the correspondence between OWL DL and OWL Full is closer.  (I
seem to remember that there are other observable differences, but I
can't bring any to the fore just now.)  

> Btw, I think the task of forcing owl:Thing to be a finite set can be even
> easier achieved by simply making owl:Thing equivalent to some enumeration
> class.

Sheesh, why aren't we (the DL community) using this example.  Silly us. :-)

The reason is probably historical - spy points are *hard* to reason with
and thus stick in the minds of theoreticians.

So another example of forcing the domain to be finite without mentioning
owl:Thing would be

	C = { a }
	C- = { b }

> Cheers,
> Michael

peter
Received on Thursday, 20 December 2007 09:18:32 GMT

This archive was generated by hypermail 2.3.1 : Wednesday, 27 March 2013 09:32:55 GMT