- From: Dan Connolly <connolly@w3.org>
- Date: Wed, 13 Jun 2001 09:07:47 -0500
- To: Brian McBride <bwm@hplb.hpl.hp.com>
- CC: pat hayes <phayes@ai.uwf.edu>, Jan Grant <Jan.Grant@bristol.ac.uk>, w3c-rdfcore-wg@w3.org
Brian McBride wrote:
>
> Dan Connolly wrote:
> >
> > This business of cycles in the subject/predicate/object graph
> > is an issue
>
> Dan, please could you clarify what the issue is. You have a test case.
> The output is quite clear. What's the problem?
Well, the issue is that there's no constraint that a
statement can't refer to itself; i.e. the liar's
paradox ("this sentence is false") is expressible
in RDF, and if anybody thinks that statements
form a tree or even an acyclic
graph, they should look out.
That design is not useful for the sorts of things
I want to do with rdf:Statements.
Maybe the stuff I want to do isn't typical; but
I can't imagine any use case for rdf:Statement
where cycles wouldn't be a problem.
Maybe I can handle this as an axiom local to my
applications. But it would leave me wondering,
along with lots of other folks, evidently,
just what reification is there for. i.e.
eventually we should have some design rationale
for rdf:Statement/subject/predicate/object.
If it doesn't seem worth adding to the issues list,
I think there are enough other issues on the issues
list that relate to this that I'll eventually
get to make my point.
--
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Wednesday, 13 June 2001 10:09:07 UTC