W3C home > Mailing lists > Public > www-rdf-logic@w3.org > June 2001

Re: rdf as a base for other languages

From: <jos.deroo.jd@belgium.agfa.com>
Date: Sat, 2 Jun 2001 12:57:19 +0100
To: pfps@research.bell-labs.com
Cc: www-rdf-logic@w3.org, connolly@w3.org
Message-Id: <OF4F9301FC.92CED03F-ON41256A5F.003A2531@bayer-ag.com>


> > [...]
> > >If RDF wanted to do one or the other, i.e.,
> > >    store only ground facts (triples)
> > >or
> > >    store encodings of more-complex information interpreted using
> > >    standardized extensions,
> > >then I would not have a problem.  However, the two are, in my mind,
> > >completely incompatible.
> >
> > I must say (but I better should go to my bed) that I think otherwise.
> > If we write (in N3)
> >   [ :a :b].
> > we actually understand that as the statement
> >   [ :a :b] null null.
> > (null in the sense of empty (implemented as a Java null))
> > so that means that we actually DON'T assert statement
> >   _:anonid :a :b.
> > (that _ stands for an anonymous namespace prefix)
> > but are still able to compute unifiers for resolution.
> > Also writing
> >   :s :p [ :q :o].
> > is asserting statement
> >  :s :p _:anonid.
> > but is NOT asserting statement
> >   _:anonid :q :o.
>
> None of the above looks at all like RDF.  If this is supposed
> to be some other language, then it is not about RDF.  If it is
> some language that maps into RDF, then I think that the mapping
> needs to be provided, so that we can see what extra assumptions
> have to be made.  In particular I don't see any null in the
> definition of RDF, nor quadruples, nor unifiers, nor resolution.
>
> peter

Thanks Peter, good answer!
quadruples I don't see
We could also have written
  :u :is [ :lh [ :lh [ :lh :b; :op :times; :rh :f]; :op :minus;
                  rh [ :lh :e; :op :times; :rh :c]]; :op :slash;
           :rh [ :lh [ :lh :b; :op :times; :rh :d]; :op :minus;
                 :rh [ :lh :e; :op :times; :rh :a]]
as part of a formula to describe the intersection point
of 2 straight lines and that are just the triples
  :u :is _:a1.
  _:a1 :lh _:a2.
  _:a2 :lh _:a3.
  _:a3 :lh :b.
  etc...
The point is that statements about anonymous things
are NOT asserted as ground facts.
Now that I think about it, the first case
  [:a :b].
should be understood as something like
  this _:a2 _:a3.
  _:a3 :a :b.
(but I have to think/hack that out)

null, unifiers and resolution are implementation stuff
for some kind of inference engine or proof mechanism
so sorry about that confusion.

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Saturday, 2 June 2001 06:57:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:40 GMT