RE: Triples from DAML

From: "Charlie Abela" <abcharl@maltanet.net>
Subject: RE: Triples from DAML
Date: Tue, 27 Nov 2001 20:10:20 -0000

> First of all thanks for your reply.
> 
> This first statement
> 
> >Actually DAML+OIL doesn't really care about either the XML/RDF encoding or
> >the RDF triples.
> 
> highlighted some of the unsolved problems I still have.
> Is there a difference between XML/RDF and RDF? Some ppl refer to the prior
> rather than the latter.

Well, that brings up the question of just what RDF is.  To me RDF is (or
should be) a logical formalism.  As such it has both a syntax and a
semantics.  Well, actually RDF has even more.  It has a data model (the RDF
graph), which, to me, is sort of an abstract syntax for RDF.  It has at
least two surface syntaxes, the XML syntax (often called RDF/XML, which I
got the wrong way around in my original message) and an n-triples syntax.
It also now has a model theory.

So RDF/XML is a particular surface syntax (usually called an encoding) for
RDF.  RDF by itsel should refer to all of the above, but often is used to
refer to the data model.

> And I am not sure I understood the second part
> 
> >What counts as far as DAML+OIL is concerned is the models of the above.
> 
> What is implied by " the models of the above": is it referring to how they
> are structured?

DAML+OIL is a logical formalism, with a model theory.  In a logical
formalism what is important is the model theory.  Any syntax is just a
means for communicating with that model theory.  

In a model theory what really counts is what models there are for a formula
(or, if you prefer, knowledge base, or if your formalism has several
layers, the data model or any other syntactic layer).  

peter

Received on Tuesday, 27 November 2001 14:19:15 UTC