RE: URI denumerability

+1.  It is the "angels on a pin" debate. 
Infinities of infinities are big enough 
to support any answer.

To me, the problem of URIs is that the 
property they assert, "identity" only 
emerges out of the properties of name 
and location; otherwise, they remain 
strings and not much more.   The point 
you make about transfer of meaning is 
well taken.   Presupposing a URI 
(all resources have URIs) is an indelicate 
way of saying "for a resource to become 
a member of the linguistic community, 
a URI must be assigned" and that is about 
all one has to say.  A URI is assigned 
and maintained. Full stop.

Cantor rants are fun but precisely 
what is one trying to prove/clarify 
with that?

len

From: Larry Masinter [mailto:LMM@acm.org]

So insofar as W3C documents disagree with IETF
documents about what a 'URI' is, we should certainly
work at resolving the difference. I'm not sure
it's necessary to actually decide how many resources
there are before agreeing on what it is a Uniform
Resource Identifier identifies, at least in the
context that it's used as a protocol element or
a semantic identifier/designator.

Received on Tuesday, 3 September 2002 11:05:22 UTC