Re: non-HTTP URIs in HTTP requests

hello noah.

> I'm not sure that my interpretation is what matters here, but FWIW: yes, 
> it applies to any URI.  First of all, the original httpRange-14 issue [1] 
> was "What is the range of the HTTP dereference function?", and there's 
> essentially no mention of URI schems.

this actually is an interesting exercise is web archeology. my search 
got me to where the issue was raised 
(http://lists.w3.org/Archives/Public/www-tag/2002Mar/0273) which pointed 
to an earlier message at 
http://lists.w3.org/Archives/Public/www-tag/2002Mar/0092 and this one 
very specifically talks about HTTP URIs only, even going as far as saying:

"Why do you want to extend the range of http URI dereference to cars? 
plate://us/ma/123yui  could still be defined to identify cars - I don't
object to other URI schemes identifying cars.  uuid schemes can as far 
as I know now."

but then again this does not really describe what the issue covers in 
terms of URI schemes, but more what the starting point of the discussion 
was (what HTTP URIs are supposed to identify).

maybe i was mislead by the "Dereferencing HTTP URIs" title, which should 
have been "Dereferencing URIs with HTTP" then. but i think this 
discussion alone proves the point that the range of httpRange-14 (pardon 
the pun) is not all that well-defined.

> So, to the extent you take [2] as representative of the TAG's position on 
> httpRange-14, it is grounded in the definition of HTTP status code 200, 
> the semantics of which are independent of URI scheme.  If you're returning 
> a 200, you're saying that the "entity corresponds to the resource", and I 
> think it's fair to say that's what the TAG says can't be the cases for 
> resources that aren't information resources.
> [2] http://www.w3.org/2001/tag/doc/httpRange-14/2007-05-31/HttpRange-14#sec-http-rep-assoc

which brings me back to my point that the whole concept of an 
"information resource" is not well-defined and conceptually weak. 
"Information resources are resources identified by URIs and whose 
essential characteristics can be conveyed in a message" is the 
definition (given in AWWW). i would argue that this is not a definition 
at all. "essential" must refer to something/somebody (the 
entity/authority deciding that a feature is essential or not), and 
regardless of what's identified by a URI, there is always a scenario 
where some entity/authority regards something as essential that cannot 
easily be conveyed in a message. or, looking at it from the other side, 
the entity/authority deciding on what is essential could always decide 
that whatever it uses/designs as representation does indeed cover all 
essential characteristics. or to put it in terms of philosophy: we can 
never make references to things "per se", we always refer to systems of 
description and communications, or, as science fiction writers often put 
it more succinctly, "everything is information". classifications can be 
made based on pragmatics, but these are contextual and not self-evident 
as assumed by the AWWW definition.

cheers,

dret.

Received on Friday, 8 January 2010 00:05:00 UTC