Re: REVISED Web Page Proposal - Version 2.0

Jason White schrieb:
> On Thu, Nov 02, 2006 at 10:01:31AM +0100, Johannes Koch wrote:
>> So, speaking in XLink terms, everything that is referenced onLoad (in contrast 
>> to onRequest) is part of the web page.
> 
> That's exactly how I understand Gregg's proposal. The main difficulty that I
> find in it is that the definition is, indeed explicitly, relative to the user
> agent.

E.g. a reference in frame/@src is meant to be onLoad, but some user 
agents will create an onRequest reference.

> One way of solving this is to say that in the case of ambiguity, the more
> inclusive interpretation (encompassing more resources) should be preferred in
> deciding what is included in a "Web page" (or "Web unit" or whatever it is
> to be called).

Yep

> Also, suppose there are two or more "versions" of the content obtainable via
> content negotiation from the same URI. Which other resources are rendered
> simultaneously with them will depend on which of the alternatives is supplied
> in response to the request made by a user agent. Also, only one of the
> alternatives will be rendered in response to such a user agent request.
> 
> My suggested solution is that if two or more primary resources are identified
> by the same URI, each of them (along with any dependent resources associated
> respectively with it) should be considered a separate "page" for purposes of
> the guidelines.

That's the problem with the "identified by a single URI". Different 
results of POST forms are another thing that cannot be "identified by a 
single URI".

So the web page identifer in some cases cannot be just a URI, but must 
be able to include the necessary parameters (in HTTP: headers, message 
body) that the resulting response is based upon.

-- 
Johannes Koch
In te domine speravi; non confundar in aeternum.
                             (Te Deum, 4th cent.)

Received on Thursday, 2 November 2006 09:53:43 UTC