Re: What is an RDF Query?

Well, I may have been confused with locally scoped names.
If I assert
  (1) "Document D has an author with name N"
then the author object would be an anonymous node.

If I assert
  (2) "Document D has author A who has name N"
then the author is a named resource, with a URI of its own.

What I would like to assert is something like
  (3) "Document D has an author, which I will (locally) call X,
       who has name N"

It might be necessary to defined the name X, in order to assert
something else about that resource (like document D' also has author X),
but actually it is much more like an anonymous resource than a named
one.

If I slightly change (1) (2) and (3) to be queries instead of assertions
("Is there a document..."), X would have to be considered a variable
rather than a constant.

Note that this is a syntactical problem: in the RDF model, there is no
reason why I should give a local name to the author. I would just draw
an empty node, and add arcs to it. Which really makes me think that
recomanded syntax is a pain ;-)

  Pierre-Antoine


On Tue, 2001-09-11 at 20:41, pat hayes wrote:
> >A question has been bothering me for a while.
> >In RDF we only have named nodes (with URIs) and anonymous nodes.
> >So apparently the models does not enable us to create locally scoped
> >names.
> >
> >On the other hand, the syntax has two different constructs for named
> >nodes : rdf:about and rdf:ID.
> >From an XML point of view, IDs are locally scoped. However, RDF makes
> >them global by prepending the document URI to them, with a '#'.
> 
> Right, RDF treats anonymous nodes like locally bound existential 
> variables, and all other names - URIs - as having global (WWW) scope.
> 
> >Do we agree that the same resource could have several URIS ?
> 
> Yes, but that is a different issue. Two globally scoped names may 
> have the same denotation.
> 
> >I do think
> >so. For example, if I write
> >
> ><rdf:Description ID="someBook">
> >  <pac:readIt>9/11/2001</pac:readIt>
> >  <pac:rating>Very good</pac:rating>
> ></rdf:Description>
> >
> >I do not deny myPieceOfRDF#someBook to identify the same resource as to
> >urn:isbn:12345. I may just have given it an rdf:ID for syntactical
> >reasons...
> >
> >In short : am I wrong to use rdf:ID as a kind of locally scoped IDs?
> 
> I would say so, yes.
> 
> >If
> >yes, how could I do it else? If no, how do we manage (and first of all
> >recognize) those "not global URIs" in queries ?
> 
> But when prefixed with yourWebSiteAddress#, they *are* global. Why 
> would they not be?
> 
> Pat Hayes
> 
> ---------------------------------------------------------------------
> (650)859 6569 w
> (650)494 3973 h (until September)
> phayes@ai.uwf.edu 
> http://www.coginst.uwf.edu/~phayes
>

Received on Wednesday, 12 September 2001 10:41:33 UTC