Re: more on the relationship between RDF and DAML+OIL

From: Dan Connolly <connolly@w3.org>
Subject: Re: more on the relationship between RDF and DAML+OIL
Date: 12 Dec 2001 11:21:21 -0600

> On Wed, 2001-12-12 at 06:36, Peter F. Patel-Schneider wrote:
> > Here is a more-formal discussion of the problems that would arise if
> > DAML+OIL syntax was treated the same way that the RDF model theory treats
> > RDF syntax.
> 
> I'm not sure I understand this argument. Could you show how
> it applies to the axiomatic semantics of DAML+OIL?
> 
> 
> -- 
> Dan Connolly, W3C http://www.w3.org/People/Connolly/
> 

It doesn't apply to the axiomatic semantics for DAML+OIL because the
axiomatic semantics does not work ``right'' with respect to RDF entailment.

For example, in the axiomatic semantics

	John is Person .

does not imply

	John is :_1 .
	:_1 complementOf :_2 .
	:_2 complementOf Person . 
	
The axiomatic semantics, instead, corresponds to the DAML+OIL model theory
way of doing things.


Even if you fixed the axiomatic semantics, assuming that this is possible,
you don't get the same situation.   In the kind of axiomatization that is
used as the DAML+OIL axiomatization the only ``error'' condition that you
get is an inconsistency in the first-order logic.  This ``error'' condition
is used for both semantic ill-formedness and inconsistency in the knowledge
base so there is no way of telling which is the case in this style of
axiomatization. 


peter

Received on Wednesday, 12 December 2001 13:22:42 UTC