W3C home > Mailing lists > Public > www-rdf-interest@w3.org > June 2002

Re: Implementing statement grouping, contexts, quads and scopes

From: Jonathan Borden <jonathan@openhealth.org>
Date: Mon, 24 Jun 2002 13:58:09 -0400
Message-ID: <044f01c21ba8$b401fcc0$0a2e249b@nemc.org>
To: <areggiori@webweaving.org>
Cc: "RDF Interest" <www-rdf-interest@w3.org>, "RDF Logic" <www-rdf-logic@w3.org>

Alberto Reggiori wrote:
I think I can see what I am trying to do here, which is proposing a
practical solution (like Patrick I guess) to layer new semantics on top of
RDF using reification and the current XML/RDF syntax which most parsers
understand already :-)
Isn't the WebOnt supposed to layer OWL on top of RDF  using the XML/RDF _as
it is today_  [1] or am I missing something here ? :-)

Well yes, fair question. But what is RDF today? The 1998 REC or the
specifications being produced by RDFCore at the present time? Part of the
answer is that if the current RDF recommendation were sufficiently precise,
there would be no need for clarification, and the current RDFCore work might
be toward RDF 2, rather than a 'clarification' of RDF 1.

Can somebody better explain to me what's wrong about using reification for
layering ?

I am not sure _what_ the status of RDF reification is in the current RDF
WDs, i.e. the model theory. There does not appear to be much discussion. I
believe part of the 'problem' with RDF reification is that its definition
was not sufficiently precise to be useful.

The past version of the RDF MT WD discussed 'asserted triples' with the
implication that 'unasserted triples' might also exist. This language was
removed in the current WD, so we must conclude that the current intention of
the current RDF MT WD is that RDF reification in fact is not an acceptable
means of representing an 'unasserted triple'. If I am incorrect, the RDF MT
should state this relationship -- that is the purpose of a model theory,
isn't it?

picking up bit and pieces from my previous example, I can define N3 formula
X as follow:
<rdf:Statement rdf:bagID="X">
   <rdf:subject rdf:resource="&ex;s" />
   <rdf:predicate rdf:resource="&ex;p" />
   <rdf:object rdf:resource="&ex;o" />
<owl:OWLPredicate rdf:about="&ex;p" rdf:bagID="#X"/>

This syntax is not acceptable (IMHO of course) for the development of a
language (such as OWL) 'layered' on RDF -- well of course this could be
fixed with a new rdf:parseType :-)) of course that just moves the problem to
those that have to deal with triple stores (good thing memory and disc space
is getting cheaper :-))

Received on Monday, 24 June 2002 14:03:26 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:41 UTC