W3C home > Mailing lists > Public > www-tag@w3.org > September 2002

Re: Using URIs as language instead of as protocol element

From: Tim Bray <tbray@textuality.com>
Date: Sun, 01 Sep 2002 09:27:47 -0700
Message-ID: <3D724003.2030404@textuality.com>
To: Larry Masinter <LMM@acm.org>
Cc: www-tag@w3.org

Larry Masinter wrote:

> I don't think you can get far without acknowledging
> that using URIs as semantic identifiers carries more
> ambiguity than as a protocol elements.

Well yes, but so what?  Like Phil Karlton used to say, "There are only 
two hard things in CS: naming and cache invalidation".  Naming things is 
hard, and naming them in a decentralized planetary-scale networked data 
system is really hard.  Our experience over the last ten years is that 
operationally, naming things with HTTP-class URLs seems to work pretty 
well, and many people believe that the generalization from URL to URI 
adds value.

Hmm... is the concrete suggestion that the webarch document contain some 
discussion of the fact that naming things via URI carries the 
possibility of ambiguity?  As in your www.w3.org example.

Or is the suggestion that the webarch principle "Absolute URI references 
are unambiguous: Each absolute URI reference unambiguously identifies 
one resource" is simply wrong?

Speaking for myself, I'm much less comfortable with that principle than 
with the others in the webarch doc.  -Tim
Received on Sunday, 1 September 2002 12:27:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:11 GMT