W3C home > Mailing lists > Public > www-rdf-interest@w3.org > September 2000

RE: abstract model and reification

From: McBride, Brian <bwm@hplb.hpl.hp.com>
Date: Wed, 20 Sep 2000 08:57:36 +0100
Message-ID: <5E13A1874524D411A876006008CD059FE7DC44@0-mail-1.hpl.hp.com>
To: Jan Grant <Jan.Grant@bristol.ac.uk>, Seth Russell <seth@robustai.net>
Cc: "RDF Interest (E-mail)" <www-rdf-interest@w3.org>
> > 2) If we want to say something about a statement that is 
> asserted in some
> > context, why cant we just designate it's id as an object (see [1])?
> 
> This is Brian's approach. My personal belief is that there may exist
> multiple resoruces which are reifications of any statement, and which
> have other differing properties that prevent their being equivalent.

Actually, I'm not sure I've got an approach at the moment.  Jan
has raised a good question and I've been trying to get an analytical
handle on it.

One of the problems here is that m&s provides an intuitive notion 
of the semantics of an RDF model, but it doesn't provide a formal
notion of a semantics the way say a logic text does.  And here I seem
to need to build arguments in terms of the semantics of an RDF model.

But lets see how far this runs ...

So if we imagine that we have two resources s1 and s2 both of which
represent some statement S.  Then any RDF statement
that was true of s1, would in fact be a statement about S and so
would also be true of s2.  And vice versa, any true statement about
s2 would be true of S and would thus also be true of s1.  A similar
argument applies to false statements about s1 and s2.

If any statement was true of s1, but not true of s2, or vice versa,
then s1 and s2 could not both be representing S.  At least one of s1
or s2 must be representing something else, possibly very closely
related to S, but not S itself, as no statement can be both true of S
and false of S.  (Dodgy ground that one - have to think about time,
contexts etc - but with a simple notion of the semantics of RDF, it
seems sound.}

That suggests that s1 and s2 are indistinguishable, and should
be considered to be identical.  If you are not sure about that
step consider what happens if s1 has a property foo:hasURI with
a particular value.

The trouble I'm having building an argument the other way is that I
don't have any examples where I might want multiple different
resources to model a statement.  I've got nothing to feed my
intuition of what that might be trying to accomplish.  Do you have
any in mind Jan?

Brian McBride
HPLabs






 
Received on Wednesday, 20 September 2000 03:57:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:44 GMT