Re: ISSUE-81 (resource vs representation)

On Sun, 27 Sep 2009, Julian Reschke wrote:
> Ian Hickson wrote:
> > >
> > > It appears that you're saying that the phone number does not really 
> > > identify a person then.
> > 
> > It gives enough information to obtain a person. (Not enough 
> > information to unambiguously obtain a _particular_ person.)
> 
> So phone numbers identify people (*), but not particular people? That 
> doesn't sound like a helpful concept.

Well, I'm no fan of phones personally, but several billion people seem to 
consider phone numbers remarkably useful despite this.


> > > So then, if a resource is a bag-of-bits, what do you call the thing 
> > > you address with a POST request?
> > 
> > A POST request returns a bag of bits also, so "resource".
> 
> The response to a POST request is a bag of bits, but that doesn't *make* 
> the resource a bag of bits.

Right. What makes the response to a POST a "resource" is that the 
definition of "resource" is "bag of bits" (amongst other things, as you 
point out below).


> Almost everybody who deals with resources is totally aware that the 
> resource is an abstract concept, and, in the case of an HTTP resource, 
> can be a static document, a Java servlet, a CGI script, whatnot; but 
> certainly not the same thing as you get when you do a GET request.

I hold the opposite viewpoint; I don't have any supportive evidence one 
way or the other though.


> The distinction of the abstract concept of a resource, and the thing a 
> GET request returns (call it bag-of-bits, entity, representation, 
> whatever) is present in *all* relevant specs (HTTP, URI, WEB-ARCH...).

I agree that HTTP, URI (and IRI), and AWWW use this terminology.

I also think that HTTP, URI (and IRI), and AWWW happen to be the three 
main documents in this space that are out of touch with reality these days 
(for many reasons, not just this particular minor terminology issue). So I 
don't really feel that they are especially relevant here.

(Larry is doing good work to address this as far as the IRI spec goes.)

Most of the other relevant specs (HTML4, CSS, DOM, SVG, XMLHttpRequest, 
etc) use the word "resource" in the far more generic sense understood by 
most people (e.g. referring to "XML resources", referring to the 
dimensions of resources, referring to resources as images, referring to 
resources having encodings, etc).


> if the "resource" is a bag-of-bits, what is the thing you send a POST 
> request to?

You send the POST request to an HTTP server, and the HTTP server responds 
with a resource.


> PS: is a resource in the "tel" URI scheme a bag-of-bits as well? Is a 
> "websocket" resource a bag-of-bits?

I don't think most people would refer to any resources in those contexts, 
though I suppose one could argue that the phone connection or TCP 
connection obtained when one uses those URLs are resources, in the same 
sense that memory, CPU, and disk space are resources.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Monday, 28 September 2009 00:28:57 UTC