RE: [httpRange-14] What do HTTP URIs Identify?

It merely needs to state the role of the interpretant with 
regards to establishing meaning.
 
The rest of it is I think, as Jonathan suggests, technique 
with regard to interpreting the media type.  URIs don't 
have meaning beyond naming the association between the 
authority and the assertion that a relationship exists. 

The system architecture is simply a way of saying that 
where such a relationship is stated by the assignment of 
the URI, one can, if the URI is assigned and processed 
correctly (no spoofing, no redirecting, no futzing), 
that the same relationship is always being signified 
regardless of point of view of the observer.  A URI 
is like the speed of light as a constant: a systemic 
way of keeping the information space from collapsing 
because one cannot ascertain the identity (singularity) 
of a given particle or location with certainty, so it 
confers an artificial certainty.

If as Fielding says, URIs are the words of the web, then 
we should understand that as the linguists say, words 
have no meaning.

I apologize for my use of metaphors.  I am a good 
sheepdog but a lousy shepherd.

len

-----Original Message-----
From: Joshua Allen [mailto:joshuaa@microsoft.com]

> "The meaning of a URI(ref) as intended by its authority is defined by
the
> set of assertions obtained when the URI is referenced.

Do you mean "dereferenced"?

> When the media type of such a representation is: application/rdf+xml,
then
> the meaning of the represented URI is given by the graph as per the
RDF
> model theory."

While this is one *possible* scenario that could occur, there are many
other ways that meaning can be associated with a URI.  It's debatable
whether this particular approach ever stands a chance of succeeding.

I strongly recommend that TAG avoid endorsing any particular approach to
associating meaning with URIs.  

I suggest sticking with the simple affirmation that URIs *do* have
meaning.  

Assignation of meaning to symbols is a complex and fluid social process
(in which ownership of a DNS server can play a big part, of course).
The Web Architecture doesn't need to codify this process, and risks
looking foolish by trying.

Received on Thursday, 8 August 2002 16:49:08 UTC