W3C home > Mailing lists > Public > www-webont-wg@w3.org > April 2002

RE: Problems with dark triples approach

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Tue, 23 Apr 2002 16:25:47 +0100
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
Cc: <www-webont-wg@w3.org>
Message-ID: <JAEBJCLMIFLKLOJGMELDMEMOCDAA.jjc@hplb.hpl.hp.com>
> 
> I (haven't yet) even touched the complexity issues.  I'm still trying to
> get at what the approach is supposed to do here, let alone how hard it is
> to do whatever that is.

In RDF and RDFS none of the triples are dark, in the following 
sense. When you write:

foo rdfs:subPropertyOf bar .

that you have said:

that <foo,bar> is a pair in the rdfs:subPropertyOf relationship
just as much as

foo eg bar .

says that they are in the eg relationship.

i.e. at some level the triple

foo rdfs:subPropertyOf bar .

is treated just like any other triple; and then the
extension of *rdfs:subPropertyOf* within the model is 
treated specially.

I was moving forwards assuming that was how DAML+OIL works
(incorrectly according to Peter?).

The dark triple approach is to assign the desired semantics
(e.g. that every pair in the property extension of foo is
in the property extension of bar) directly to the
dark triple in the graph, rather than indirectly as in
RDF MT.

Jeremy
Received on Tuesday, 23 April 2002 13:55:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:49 GMT