Re: Dark triples motivation

From: Dan Connolly <connolly@w3.org>
Subject: Re: Dark triples motivation
Date: 15 Apr 2002 13:38:58 -0500

> On Mon, 2002-04-15 at 12:33, Peter F. Patel-Schneider wrote:
> > The only problem that I see with your examples is that they concentrate on
> > daml lists (daml:collection) and containers.  I see the problem much more
> > as having to do with defined classes and restrictions, and lists and
> > containers only showing up because they are needed in some places in
> > defined classes and restrictions.
> 
> I'm still at a loss to see how 'dark triples' solves anything.
> So I'd love to see an even simpler example of how it can help...
> one that doesn't use lists/collections would be great.

Well, almost all the examples end up using daml lists, because so much
of daml syntax uses daml lists.

Here is one of the really bad examples

   :_2, rdf:type, owl:Restriction .
   :_2, owl:onProperty, rdf:type .
   :_2, owl:maxCardinalityQ, "0" .
   :_2, owl:hasClassQ, :_3 .
      :_3, owl:oneOf, :_4 . 
        :_4, owl:first, :_2 .
	:_4, owl:rest, owl:nil .

If you don't have dark triples, and you want to have reasonable
inferences, then this kind of restriction ends up being in all owl
interpretations.  As this restriction is self-contradictory, all owl
interpretations contain a contradiction.

Why does this restriction have to be in all owl interpretations?  Well,
because individuals can't belong to restrictions that don't exist.  Why is
existence a characteristic of restrictions in the first place?  Well,
because all triples are non-dark. 

So, dark triples will destroy this line of reasoning.  Is this a proof that
they make all the semantic problems go away?  Of course not.  That requires
an in-depth analysis of a proposal.  However, it appears to me that dark
triples are sufficient to make these semantic problems go away.  Is this a
proof that dark triples are needed to make all the semantic problems go
away?  Of course not.  There are already proposals that do not have
semantic problems and also do not need dark triples.  However, in my view,
all these proposals have their own problems.  For example, RDFS is not
expressive enough and DAML+OIL has far too weak entailment.

peter

Received on Tuesday, 16 April 2002 15:24:27 UTC