Re: problem with blank node identifiers and rdf:nodeID

On Thu, 28 Nov 2002, Peter F. Patel-Schneider wrote:

>
> From: Brian McBride <bwm@hplb.hpl.hp.com>
> Subject: Re: problem with blank node identifiers and rdf:nodeID
> Date: Thu, 28 Nov 2002 17:01:01 +0000
>
> > At 10:55 28/11/2002 -0500, Peter F. Patel-Schneider wrote:
> >
> > >It appears to me that the use of rdf:nodeID makes it impossible to
> > >correctly transform an RDF/XML document into n-triples without first
> > >checking for all the rdf:nodeID uses in the document.  Otherwise a
> > >generated blank node id could accidentally be the same as one used (later)
> > >in an rdf:nodeID in the document.
> >
> > Is there anywhere in the spec where it says that the node id used in the
> > n-triples MUST be the same as the nodeID used in the rdf/xml?
> >
> > Brian
>
> As indicated in my response to Jan, I believe that the spec explicitly
> requires that the node id in the triple is the same as in the RDF/XML.

There are words that might give that impression here:
	http://www.w3.org/TR/2002/WD-rdf-syntax-grammar-20021108/#nodeElement

However, that section also links to
	http://www.w3.org/TR/rdf-concepts/#dfn-blank-node-id

which gives a good definition for blank nodes. The section also uses the
text, "generate a blank node identifier" without making explicit the
need for a disambiguation step so that generated ids are distinct from
ids used in the document.

Reading section 2.3.2 of the concepts document, I think it is clear
that a simple "name mangling" function such as I suggested is sufficient
to generate a correct transformation to N-Triples. Nevertheless, the
syntax spec needs some words to catch this:

[[
If e.subject is empty, generate a Blank Node Identifier i and ...
]]

so that the space of generated identifiers and ids supplied by
rdf:nodeID are clearly distinct.

-- 
jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287088 Fax +44 (0)117 9287112 http://ioctl.org/jan/
New Freedom of Information Act: theirs, to yours. Happy now?

Received on Thursday, 28 November 2002 12:26:55 UTC