Re: URGENT: train wreck coming, action needed. (was: Fwd: URI-CG group chartered)

On Saturday, Apr 5, 2003, at 05:18 US/Eastern, Graham Klyne wrote:
> <aside>
> I think there has long been a tension that URIs serve (at least two) 
> different masters:  in the web architecture (wherein the concept 
> originated), as a framework for universal identification, but within 
> the IETF (who "own" the specification) I sense a broad view that URIs 
> are some kind of glorified address.  For many purposes, these are 
> reconcilable views, but when issues like this come up one sees the 
> fault lines emerge.
> </aside>

Yes, that is well put.  This has been a certain tension all along.  
While the URI spec's syntax and the HTT protocols allowable message 
sequences have been indeed envelope within the IETF, the web philosophy 
was not understood by everyone concerned.

To take one example, when the web architecture called for a universal 
identifier space, and wanted to grandfather in Gopher servers, some 
folks wanted to keep the ":" hierarchical separator in the gopher URI, 
even though the web architecture called for standardization on "/" to 
make relative URIs work. In essence, while the web architecture was a 
more encompassing one, but some Gopher users didn't see any need to 
have to fit in with it. And of course they didn't have to.

Now, we have the Semantic Web and the World Wide Web as it is. The 
Semantc Web is a user of the WWW in one way, in that it uses the web to 
put its document in.  However, when it comes to the languages, the 
Semantic Web languages are more encompassing, in that RDF models real 
things and abstract concepts, and includes documents messages as 
special cases.  So while URI users who work in a world of documents 
feel they have the right not to be encompassed by something they don't 
[think they]  need.

This is behind some of the discussions of "Resource" and so on.

In fact, web architecture and semantic web architecture can be 
integrated fairly seamlessly , with a bit of understanding on each 
side. I think it needs a formal description, so that we don't rely on 
english terms.

Tim

Received on Monday, 7 April 2003 15:05:34 UTC