Re: Existential Quantification [Re: New RDF model theory]

Hi Pat,

> > > In danger of making another mistake ....
> > >
> > > I didn't quite see how the interpolation lemma worked when the LHS has
> > > anonymous nodes.
> > >
> > > How does it get the following entailment
> > >
> > >
> > > _:x <b> <c> .
> > >
> > >
> > > entails
> > >
> > > _:y <b> <c> .
> >
> >I just see that we have a bug in our stuff...
> >for
> >  _:x <b> <c>. |=> _:y <b> <c>.
> >we assert LHS and query it with RHS
> >and i think we should query with
> >  {_:y <b> <c>} log:forAll _:y.
> >and then there is a match found with
> >  {_:x <b> <c>} log:forSome _:x.
> >(i have to think about that, and try...)
>
> That seems right to me. The log:forAll on the RHS tells the reasoner
> it is free to substitute anything for the variable, right?

Well I think I was wrong with that forAll
that entailment doesn't seem to make sense
We have it now like

==== jeremy3.nt
_:x <b> <c>.
====

==== jeremy4.nt
_:y <b> <c>.
====

====
C:\n3>java Euler -core -trace jeremy3.nt jeremy4.nt
@@ assert [] <b> <c>.
[1]CALL: [] <b> <c>.
[1]EXIT: [] <b> <c>.
@@ continue
# Generated with http://www.agfa.com/w3c/euler/#27.055 on Tue Aug 28 21:16:35 CEST 2001
# for query file:/n3/jeremy4.nt
# given [file:/n3/jeremy3.nt]

@prefix log: <http://www.w3.org/2000/10/swap/log#>.

[] <b> <c>.

# Proof found for file:/n3/jeremy4.nt in 1 step (573 steps/sec)
====
so like 'blank' node tokens

But if we have
e.g.
  {:s [ :subPropertyOf :p] :o} log:implies {:s :p :o}.
we treat it like
  {:z :subPropertyOf :p. :s :z :o} log:implies {:s :p :o}.
where :z is universally quantified
and that seems to work quite well

and BTW for so called 'variable' predicates we find
your MT thing/extension distinction really wonderful!
(I mean the :z once in subject and once in predicate
position and treating them without and with extension)

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

Received on Tuesday, 28 August 2001 16:48:50 UTC