Re: Layering LX (or FOL) on RDF

> If you forbid self-reference, then it is possible.  If you don't correctly
> capture FOL entailment, then it is possible.  If you extend the syntax of
> RDF, then it is possible.  If you don't use RDF semantics for RDF triples,
> then it is possible.
> 
> What it depends on is whether you think any of these violates what a
> same-syntax semantic extension of RDF is.

So you agree that forbidding self-reference (in the sense discussed
earlier in this thread) is enough, but you think that doing so
violates some essential part of RDF?  I've had the impression that the
RDF Core WG is attached to reification for issues like provenance
(using RDF to record the sources of RDF information), but I'd be
rather surprised if they were attached to self-referencing sentences.
I imagine the only reason they allow them is because (1) it's not a
problem in their negation-free world, and (2) it's more complicated to
disallow them.

To bring this full circle (this thread started with Sean's document),
I'm finishing up a modifation to cwm which will let it translate N3 to
RDF triples using the LX vocabulary.  That will demonstrate that N3,
while not 1-1 equivalent to RDF, is equivalent to RDF + some
vocabulary.

    -- sandro

Received on Tuesday, 27 August 2002 11:45:09 UTC