RE: HTTP Endpoints and Resources

>Rhys,
>
>I like your recognition that 200 and 303 URIs have something in common,
>but please don't refer to the 303 case has having an "http endpoint"
>that "responds", because doing so would introduce the unnecessary
>confusion of having the same URI denote two different things: an "http
>endpoint" and the thing the URI was intended to denote -- a person, for
>example.
>
No. This is confused because of the now venerable confusion, which 
continuously dogs all these discussions, between "identifying" in the 
sense of providing a functional Web-mediated connection to something, 
and referring to an object, aka denoting it. These are (forgive the 
shout) NOT THE SAME THING. Please don't get them confused. A name can 
denote without being attached (like all names off the Web) and it can 
be attached without denoting. This awful word "identifies" sometimes 
means one thing, sometimes another; but the two concepts are 
distinct. The fact that a URI provides access to something (a 
resource? I neither know nor care) which performs some meaningful 
action (such as a 303 redirect) does NOT imply that the URI, 
considered as a referring name, must therefore denote that thing 
(resource?). These two relationships, of accessing and naming, are 
COMPLETELY DISTINCT. Now, the TAG is suggesting, and I agree that 
this makes good sense, that *in the case where the accessed resource 
emits a 200 response* (and perhaps in similar cases for other xxTPs, 
i.e. where the access works 'normally'), then we should all agree 
that the accessing URI does indeed denote the resource: but this is a 
*convention* we are being asked to adopt, not a fact of nature or an 
architectural requirement. And this convention does not say ANYTHING 
about the relationship between denotation and access in other cases: 
nor should it, indeed, because there is nothing to be said in those 
cases. If you get a 404 error or if you get a 303 redirect, you can 
infer NOTHING AT ALL about what, if anything, the URI denotes.

>What the 200 and 303 cases have in common is that the server's response
>indicates that the URI owner has associated the URI with a resource,
>i.e., the URI owner has "minted" or "allocated" the URI.  (Slight
>digression: hence the server's response can be viewed as "declaring"[3]
>that URI.)  But in both cases it is the *server* that sends back the
>response -- not the URI.  When the response is 200 we may colloquially
>speak of "the URI responding", and this sloppiness in language is
>harmless in that case because the URI denotes the resource that is
>(conceptually) "responding".  But when the URI denotes a person, and the
>response is 303, it is *not* the person that is conceptually responding.
>Hence, the sloppiness of talking about "the URI responding" becomes
>quite misleading.

But nobody said that the URI responded. In the 200 case, what 
responds is the resource Web-attached to, and denoted by (assuming 
the above-mentioned convention) the URI. In the 303 case, what 
initially responds is the thingie (resource?) which is Web-attached 
to the URI - which the URI accesses - but which (might but more 
probably) might not be denoted by it. In the 404 case, all bets are 
off.

BTW, I don't think calling any of these a 'declaration' helps in any 
way at all, and if anything is only going to make things more 
confusing.

Pat Hayes

>3. URI declarations: http://dbooth.org/2007/uri-decl/
>
>
>David Booth, Ph.D.
>HP Software
>+1 617 629 8881 office  |  dbooth@hp.com
>http://www.hp.com/go/software
>
>Opinions expressed herein are those of the author and do not represent
>the official views of HP unless explicitly stated otherwise.
>


-- 
---------------------------------------------------------------------
IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes

Received on Wednesday, 26 September 2007 16:57:08 UTC