RE: Proposal: new top level domain '.urn' alleviates all need forurn: URIs

On Thu, 2003-07-10 at 09:26, Patrick.Stickler@nokia.com wrote:
> > -----Original Message-----
> > From: ext Michael Mealling [mailto:michael@neonym.net]
> > On Wed, 2003-07-09 at 14:35, Sandro Hawke wrote:
> > > > > I'm using the word "same" in the normal dictionary 
> > sense of being
> > > > > truly and completely indistinguishable.  If I tell you <Jim>
> > > > > owl:sameAs <James>, then you know I'm using the terms 
> > "<Jim>" and
> > > > > "<James>" as synonyms, as two names for the same thing.
> > > > 
> > > > Absolutely correct! _You_ have told me that "Jim" and "James" are
> > > > synonymous. But RFC 2396 provides no method for making 
> > such statements.
> > > 
> > > I can put aside the OWL terminology entirely.
> > > 
> > > You seem to be saying that 
> > >     http://www.w3.org/
> > > and
> > >     http://WWW.w3.org/
> > > could not possibly ever both identify the same resource.  
> > You seem to
> > > be saying that if two URIs are textually different, they cannot
> > > possibly identify the same thing.
> > 
> > It depends on where in Tim's taxonomy of normalization you are. Since
> > you deal with the 'http' scheme a lot I assume you can normalize that
> > down to something that allows you to compare them via string
> > equivalence. If you were in the LDAP world and had no need to 
> > know about the 'http' scheme then you wouldn't be able to do that 
> > scheme specific normalization step that would allow you to do 
> > that comparison.
> > 
> > I.e. if I knew nothing about the http scheme then yes, they are
> > different URIs and thus different Resources. Remember, I live 
> > in a world that doesn't understand the 'http' scheme. No 
> > 'representations', 'redirects', or any other 'http' 
> > specific semantics. 
> 
> I could accept the assertion that they are different names;
> and thereby, have the potential, if not even presumption, 
> that they denote different resources.
> 
> But I don't accept the assertion that they *cannot* denote
> different resources.
> 
> And that has nothing whatsoever to do with concepts of representation,
> redirect, description, etc.

Its all about _layers_ Patrick. At the Layer that RFC 2396 is at there
is no language for talking about comparison of Resources because you
don't do it at that layer of the architecture. If you need to say that
two Resources are the same or different by some degree then you have to
go up a layer into something like OWL or RDF or some other application
(try LDAP where equivalence can depend on some funky DIT rules). Now if
I were OWL or RDF I would not use the term 'resource' since its to
easily confused with that term as its used one layer down. Call it an
'object' or something like that.

> > > If you are saying that, you are making the "unique names assumption"
> > > (UNA) for URIs, and living in a world in which RDF and OWL 
> > cannot work
> > > as currently specified.
> > 
> > I never said my very simplistic definition of 'sameness' is 
> > the only one
> > that's allowed. Its simply the only one that's universal. OWL and RDF
> > can make statements about degrees of sameness between two URIs all day
> > long. They just have to be explicit about doing so....
> 
> But you seem to be asserting that co-denotation is prohibited,
> not simply possible. There is nothing left for OWL or other
> layers to contribute, without coming into direct conflict with
> what you appear to be asserting, that lexically distinct URIs
> always denote distinct resources.

Its prohibited _at that layer_ by the simple fact that you have no
language in which to express any of those concepts. OWL and RDF can
contribute all manner of equivalences if they so desire but those
concepts are not universal to all URIs or applications, they exist
within their own application scopes at their own layer. 

-MM

Received on Thursday, 10 July 2003 10:22:55 UTC