RE: Are duplicate property/value pairs permitted for a resource?

(To those interested in following this thread, I will post my future
responses only to the 'www-rdf-comments@w3.org' mailing list.)

John,

I disagree.  RDF is a way to specify just "facts".  If "app2" wants to state
that it believes that (#A, #B, "2") is true, it needs to state that as a
reified Statement, as specified in section 4 of the RDF Model & Syntax Spec.

For example, the set of triples for such a reified Statement "S" might look
like the following (assuming that the "x:assertedBy" Property is defined
somewhere):

(rdf:type, #S, rdf:Statement)
(rdf:predicate, #S, #A)
(rdf:subject, #S, #B)
(rdf:object, #S, "2")
(x:assertedBy, #S, "app2")

Note that (#A, #B, "2") does not appear anywhere.

-----Original Message-----
From: John Cowan [mailto:cowan@locke.ccil.org]
Sent: Friday, April 09, 1999 12:20 PM

> Eric Prud'hommeaux wrote:

> > app1 adds (#A, #B, "2")
> > app2 adds (#A, #B, "2")
> > app1 dels (#A, #B, "2")
> > app2 queries for (#A, #B) values and should get "2".

> That depends on what the semantics of deleting statements might be.
> If deleting a statement merely means that *you* are no longer
> asserting it, then the above is correct, but has nothing to do
> with literals: all would be the same if "2" were replaced by #C.

> But if deletion means that the statement is not *true*, then app2 should
> fail.

> In general RDF has no mechanism for dealing with agents who think
> they know contradictory things.