Re: Generic-Resources-53: URIs for representations

On Oct 11, 2006, at 12:52, ext noah_mendelsohn@us.ibm.com wrote:

> Patrick Stickler writes:
>
>> I this case, I'd argue that these URIs do not in
>> fact denote Representations, but rather resources
>> which are tightly related to the clock resource,
>> and corresponding to particular "modes of view",
>> and which share representations with the clock
>> resource -- such that at any given time, both the
>> URI of the clock and the URI of the mode of view
>> may resolve to the exact same representation, yet
>> neither of those resources are themselves
>> Representations.
>
> Yes, that's clearly one of the sensible formulations.  My concern  
> is that
> people tend to speak somewhat informally.  When they say:  "I want  
> a URI
> for the image/jpeg representation of the clock", they sometimes  
> mean for
> some particular bit stream, as you propose, and sometimes mean for the
> resource that will give the time-varying image/jpeg.  Sometimes  
> they don't
> notice the potential issue.   My point was that, given the  
> ambiguity in
> informal usage, it's worth explaining carefully which we mean when  
> writing
> Recommendations, TAG findings, etc.

I fully agree.

It should be clear to folks operating at the web layer, that
even though it is understood that URIs denote resources, they
will never ever see or touch those resources. All the web does
is provide a means to access and interchange representations.

To what degree the nature of the actual resources can be
determined based on those representations is entirely outside
the scope of the web layer itself, and is in the focus of the
semantic web and social interaction.

>
> Note that URIs for the time-varying abstraction are reasonably  
> common. For
> example, the  XHTML Recommendation 2nd Edition [1] has links to
> "Postscript Version" [2], "PDF Version" [3], etc.  Knowing what I  
> do about
> W3C URI-allocation policies it's a good bet that the octet stream  
> returned
> for these will change if there's a 3rd edition.  That doesn't prove we
> should call these URIs "for the PDF representation", but I bet that at
> least informally a lot of people will.

Which is why, for those that care, there should be some RDF or
similarly expressed knowledge accessible via the URI of the
resource in question -- ideally accessible in a manner independent
of actual representations and content negotiation machinery.

It is this very issue of users being able to discover, as much
as is possible/reasonable, what resource a given URI denotes and
the nature of that resource that I devised URIQA, and for that
purpose, it works great.

Cheers,

Patrick

>
> Noah
>
> [1] http://www.w3.org/TR/xhtml1/
> [2] http://www.w3.org/TR/xhtml1/xhtml1.ps
> [3] http://www.w3.org/TR/xhtml1/xhtml1.pdf
>
>
>
> --------------------------------------
> Noah Mendelsohn
> IBM Corporation
> One Rogers Street
> Cambridge, MA 02142
> 1-617-693-4036
> --------------------------------------
>
>
>
>

Received on Wednesday, 11 October 2006 20:13:29 UTC