Re: URIs for terms: motivation [was: Requirements Document]

From: Jim Hendler <hendler@cs.umd.edu>
Subject: Re: URIs for terms: motivation [was: Requirements Document]
Date: Fri, 15 Feb 2002 10:36:39 -0500

> >This really needs to be discussed in more detail and deconstructed. 
> >Its full of nonsequiteurs and misunderstandings. The fact is that 
> >the entire concept of URI is completely confused right now, and 
> >nothing in any documentation put out by the W3C is enough to clarify 
> >it. We can't just take it as a given, we have to have some 
> >discussion of what it is supposed to mean. In particular, what is 
> >"on the Web" ?? If a document has a URL and uses a referring 
> >expression, is that expression on the web? Is the thing named by the 
> >expression on the web? What does 'identified distinctly' mean, 
> >exactly? None of this is clear, and getting it clear is one of the 
> >most important jobs we could do.
> 
> 
> Pat - thanks for the many very valuable contributions you made today 
> (glad you found the missing mail :->)   On this one, however, I want 
> to make a "Chair's statement" which is that while I personally agree 
> with the fact that getting this right is crucial, and hope fervently 
> that the W3C Technical Architecture Group is working the issue, I 
> think we can RULE THIS OUT OF SCOPE for our working group - we can 
> certainly state what version or document, or what assumptions about 
> URIs we are making, but trying to solve this is well-beyond our 
> charter and I want to make sure we don't spend too much time going in 
> this direction...
>   -JH

Let me come down firmly in the middle here.

I think that there are problems with URIs that need to be addressed before
WebOnt can finish.  However, I think that WebOnt doesn't need a complete
solution (even assuming that a complete solution is possible).

So, in particular, I suggest that the WG mandate the Chair to send a strong
message to the *POWERS*THAT*BE* saying that we desperately need a way of
accessing XML Schema document definitions.

Other issues that we probably need to have addressed have to do with
identity of URIs.  Right now URIs are the same if and only if they are
bit-for-bit identical.  However, we should all know many groups of URIs
that are not the same under this definition but that necessarily refer to
the same thing.

peter

Received on Friday, 15 February 2002 10:51:56 UTC