Re: RDF Namespaces and Attributes (again)

Dan Brickley wrote:
> I'm not an rdf syntax expert, so my honest answer is "I don't know". My
> instinct would be to accept qualifications of about, resource, parseType
> since the intent is clear, even if the spec isn't.

That's my interpretation. Our RDF processor accepts rdf: qualification of
all RDF syntax "keywords", regardless of context. It generates unqualified
"about" and "ID" IFF they are attributes of an rdf:Description, otherwise
all keywords are rdf: qualified.  This unfortunately means I have to have
special case code for typedNodes (must use rdf:about or rdf:ID), production
[6.12] abbreviation, etc.

The formal grammar BNF is oversimplified: it doesn't take into account the
XML scope of about, aboutEach, aboutEachPrefix, ID, and bagID.  If you
forget about the fact that this represents RDF, and just look at the syntax
in the context of any conforming XML application, these attributes are
"global" in scope, since they can appear in "any" element (by virtue of
typedNodes and abbreviated properties), not just a "local" type like
rdf:Description. As such, they should always be qualified with rdf:.

I don't think there is much we can do about the spec, since we don't want to
make a change that invalidates existing RDF.  If there is a FAQ or Errata,
we might deprecate the omission of the rdf: qualifier, and recommend that
all keywords be qualified in all contexts.

Perry

P.S. RDFMS trivia for the day: is the following well-formed RDF? What
N-tuples should it produce?

<?xml version="1.0"?>
<Foo about="me" and="you" xmlns="schema:"/>

Received on Saturday, 30 September 2000 17:26:41 UTC