RE: QName URI Scheme Re-Visited, Revised, and Revealing

> OK, now I think I grasp what you're arguing. All you have to do 
> is understand that:
> RDF Does Not Deal With QNames
> and everything's OK.

The RDF model does not deal with QNames. I never said it did
nor should.

The RDF serialization *does* deal with QNames. It is the
mapping from the serialization to the graph where the problems
may arise.

I honestly don't know why folks keep thinking I'm talking about 
the model...  Suggesting the use of URIs which reflect QNames is
hardly suggesting that QNames themselves be first class objects 
in the model...

> > And those two sources are syndicated at run-time, where 
> both creators
> > of the knowledge were unaware of the other's use of QNames, and 
> > therefore
> > could not forsee any potential problem of collision, we end 
> up with the
> > following ambiguous RDF triples:
> The creators should have been thinking of URIs for their terms, 
> not QNames.  

Fair enough. But they are still then limited to URIs which allow
direct partitioning into QNames. I.e. they can't use URIs with
bracketing syntax.

It would seem to me that simply allowing predicates to be defined
in terms of URIs and not elements using QNames would solve this latter
discrimination. E.g. something like

   <rdf:Description rdf:about="...">
      <rdf:Property rdf:about="urn:partax:(foo(bar(bas)))"

Then the whole QName to URI mapping issue becomes a non-issue for those
who choose to live solely in URI space for resource identity.

I know that NTriples and other alternate serializations, XML or otherwise,
permit such things, but those are not the standard, etc. yada yada yada...

> Here's another maxim:
> RDF Has No Concept Of Namespace

Or rather, the RDF *model* has no concept of Namespace. And to that end, I
never said it did, never believed it did, and am surprised if my comments
gave such any such impression.

The RDF serialization model, however, most certainly has a notion of
namespaces, and one which impacts the presumably free use of arbitrary URIs 
for resource identity within the model space.

But if that doesn't impact what you're working on just now, I guess
it really doesn't matter.

> Now I will delete the rest of this thread from my mailbox,

That's certainly your perogative.

Patrick Stickler                      Phone:  +358 3 356 0209
Senior Research Scientist             Mobile: +358 50 483 9453
Software Technology Laboratory        Fax:    +358 7180 35409
Nokia Research Center                 Video:  +358 3 356 0209 / 4227
Visiokatu 1, 33720 Tampere, Finland   Email:

Received on Monday, 27 August 2001 04:45:21 UTC