Re: UPDATE: initial message concerning syntax

From: Dan Connolly <connolly@w3.org>
Subject: Re: UPDATE: initial message concerning syntax
Date: 14 Dec 2001 12:43:13 -0600

> On Fri, 2001-12-14 at 12:01, Peter F. Patel-Schneider wrote:
> > From: Dan Connolly <connolly@w3.org>
> > Subject: Re: UPDATE: initial message concerning syntax
> > Date: 14 Dec 2001 11:57:47 -0600
> > 
> > > On Fri, 2001-12-14 at 11:43, Peter F. Patel-Schneider wrote:
> > > [...]
> > > > Because that is how it should work.  Do you really want SWOL entailment to
> > > > be different from RDF entailment on RDF documents?
> > > 
> > > Yes; I expect the SWOL semantics to make more conclusions valid.
> > 
> > Name some.
> 
> 	C intersectionOf L
> 	L first A
> 	L rest L2
> 	L2 first B
> 	L2 rest nil
> 	x type C
> SWOL-entails
> 	x type A

I see now what you are getting at.  

Pat and I have been having a discussion with Jim over this point, which for
some reason Jim took off the wg mailing list.  

However, the problem here is that I was using a term ``RDF document''
without giving it a precise definition.  What I meant was:

If D1 and D2 are two RDF documents that do not use any non-RDF DAML+OIL (or
SWOL) constructs then D1 RDF-entails D2 precisely when D2 DAML+OIL- (or
SWOL-) entails D2.

Sorry for not being sufficiently precise.



> > > > > I'd say SWOL entailment reduces to FOL entailment (less
> > > > > the excluded middle).
> > > > 
> > > > SWOL entailment does exclude the middle.
> > > 
> > > Er... argument by assertion. I can do that too:
> > > No, it does not.
> > 
> > I'm not making an argument by assertion, just stating a simple fact that
> > can easily be determined from a quick perusal of the appropriate documents.
> 
> The word "middle" doesn't occur in
> http://www.daml.org/2001/03/model-theoretic-semantics.html
> 
> I assume that's the the document you were talking about.
> The text of that document is sufficiently dense that
> I don't get much out of a quick perusal. I'd appreciate
> if you'd point out where it says that DAML+OIL
> excludes the middle.

If you are going to argue about logic and semantics please try to
understand what you are arguing about.

The entire model theory uses constructs that assign exactly one of two
situations to ``facts''.  For example, a pair is either in the extension of
a property or it is not in the extension of the propery.  It is not
possible for a pair to be neither in nor not in the extension of the
property.  As another example, the extension of the complement of a class
is precisely the set of resources that are not in the extension of the
class.  This is the law of the excluded middle.

> -- 
> Dan Connolly, W3C http://www.w3.org/People/Connolly/


peter

Received on Friday, 14 December 2001 14:08:45 UTC