Re: Changes to make S&AS consistent with RDF Semantics document / (fwd)

• From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
• Date: Wed, 03 Dec 2003 09:56:34 -0500 (EST)
• To: horrocks@cs.man.ac.uk
• Cc: herman.ter.horst@philips.com, www-webont-wg@w3.org
• Message-Id: <20031203.095634.93847123.pfps@research.bell-labs.com>

From: Ian Horrocks <horrocks@cs.man.ac.uk>
Subject: Re: Changes to make S&AS consistent with RDF Semantics document / (fwd)
Date: Mon, 1 Dec 2003 18:43:14 +0000

> herman.ter.horst@philips.com

> My proposal would be to reformulate the second bullet
> in the definition of abstract OWL interpretations
> in something like the following:
> 
> "LV, the literal values of I, is a subset of R that contains
> the values of plain literals in V, and each value in V 
> for each datatype in D."

I view this as a terrible change.  It causes changes to entailment in OWL Full.

Under the CR semantics, the following is inconsistent in OWL Full.  (I'm
using the abstract syntax, slightly extended, to make the example easier to
show.) 

Ontology(
  ObjectProperty(SSNi)
  DatatypeProperty(SSN Functional InverseFunctional inverseOf(SSNi)
	domain(Person) range(xsd:integer))
  Class(xsd:integer partial restriction(SSNi atleast(1 Person)))
  Class(Person partial restriction(SSN atleast(1 xsd:integer)))
  EnumeratedClass(Person John Susan Sally)
)

This ontology is satisfiable in interpretations that have between one and
three integers, which is allowable in Herman's proposed change to OWL.



peter

Received on Wednesday, 3 December 2003 09:59:49 UTC