Lists Home |
Date Index |
- From: Jonathan Borden <firstname.lastname@example.org>
- To: Uche Ogbuji <email@example.com>
- Date: Sat, 16 Dec 2000 22:52:49 -0500
> Hmm. I'm not sure I buy this. I see RDF Schemas as no more than an arc
> constraint mechanism. I don;t think this is quite enough to create
> an "ontoloty". Of course I know that ontologoes can be built upon RDF,
> but I think it requires _much_ more specialized processing layered on top
> of RDF and RDF Schemas.
> Maybe this doen't contradict what you say...
I think we are in agreement with a caveat.
As DTDs and XML Schemata provide syntactic constraints, RDF Schemata provide
semantic constraints (which for the purposes of this discussion is defined
as placing constrainsts on arcs and nodes in a directed graph). So it is
true that RDF Schema provides no more than a constraint mechanism, but I
offer that this is precisely how ontologies are built. Perhaps I can offer
an analogy to the process of creating a piece of sculpture from a block of
granite being simply constraining the spacial location of the granite :-)
Ontologies can be defined using directed graphs and hence ontologies can be
using languages built upon the RDF Schema language (which is itself built
upon RDF). Notably both DAML-ONT and OIL are built upon RDF Schema and
these languages are indeed designed to define ontologies.
So the caveat is that I would say that what we do with these ontologies once
we have them is what requires the specialized processing layer. Of course
what we do with the ontologies once we have them is what is interesting.
The Open Healthcare Group