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

> -----Original Message-----
> From: ext Weibel,Stu [mailto:weibel@oclc.org]
> Sent: 09 July, 2003 19:32
> To: uri@w3.org
> Subject: RE: Proposal: new top level domain '.urn' alleviates all need
> for urn: URIs
> 
> 
> 
> [thinking to myself... oh boy, why am I doing this...???]
> 
> My friend Michael Mealling [mailto:michael@neonym.net] asserts:
> 
> > If you have two URIs then you, by definition, have two distinctly
> different Resources.  
> 
> The following three URIs are well formed and distinct, but resolve to
> precisely the same resource:
> 
> http://purl.org/dc
> http://purl.oclc.org/dc
> http://dublincore.org
> 
> It can be argued that the first and second actually point to 
> a resolution
> table, which in turn supports redirection of the link to the 
> actual resource
> (the third), though this seems to me a distinction without a 
> difference, and
> unhelpful.
> 
> The distinction between the first and the second resides at 
> the DNS level
> (two domains pointed at the same machine... not all that 
> uncommon).  Clearly
> these two are different URLs that resolve to the same resource.

Quite.

Perhaps it would be more precise to say that all three
of the above URIs

1. have equivalent denotations (refer to the same resource)
2. resolve to the same set of representations of that resource
3. (potentially) have distinct intensions

I see the intersection of the traditional (RESTful) Web and
the Semantic Web as being the shared denotations of URIs.

Both the Web and Semantic Web share the same set of identifiers
and a given URI denotes the same resource for both a web agent
and a SW agent. 

The Web is concerned with representations of those resources
retrievable via their URIs. Navigating the Web constitutes
traversing from representation to representation via URIs
referring to related resources.

The Semantic Web is concerned with knowledge about resources
retrievable in some fashion (e.g. URIQA) via their URIs.
Navigating the Semantic Web constitutes traversing from 
description to description via URIs referring to related
resources.

Different URIs may denote the same resource, even if different
representations or descriptions are available via those URIs.

The particular URI chosen often (and usually unknowingly) 
reflects a distinct intension from other URIs that might
denote the same resource, for which we need languages such as 
RDF and OWL to model effectively.

While my assertion that http: and https: variants (typically)
have identical denotations might be questionable (at least
insofar as the letter of the specs), it is certainly the case 
that the assertion that two different URIs always denote two 
different resources is not only false, but would render both 
the Web and Semantic Web unworkably constrained and more
seriously, semantically disjoint.

Regards,

Patrick


--
Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com

Received on Thursday, 10 July 2003 09:09:51 UTC