RE: Do resources have representations?

> -----Original Message-----
> From: ext Benja Fallenstein [mailto:b.fallenstein@gmx.de]
> Sent: 28 July, 2003 15:07
> To: Stickler Patrick (NMP/Tampere)
> Cc: www-rdf-interest@w3.org
> Subject: Re: Do resources have representations?
> 
> 
> 
> Hi Patrick,
> 
> Patrick.Stickler@nokia.com wrote:
> > I saw nothing in your summary that I would disagree with. In
> > fact, I think it captures the key issues. It's very similar
> > to a recent post of mine on the TAG list
> > 
> > http://lists.w3.org/Archives/Public/www-tag/2003Jul/0172.html
> > 
> > And I see nothing wrong with your summary. 
> > 
> > You seem to think there is a problem lurking in there somewhere,
> > and I'd like to understand why?
> 
> It may just be my mental model that was wrong so far.
> 
> Somehow I've assumed that all that a URI "does" is:
> 
> - it denotes a resource

Right.

> - it is associated with some authoritative information

Not necessarily.

One can use a URI to denote some thing in the universe without
ever providing either representation or description of that
thing.

That would be pretty useless. But certainly within the
bounds of the specs.

Note, we're just talking about URIs here. Not URIs that
are significant for some process or protocol, such as
HTTP or a particular application. There will usually be
additional requirements about what one can (or should
be able to) do with a given URI within on a given 
application context.
 
Within the context of an XHTML href attribute being
evaluated by a web browser, one expects that that URI
can be used to resolve to one or more representations.
But that's not a requirement placed on the URI, but
on its use in a given context.

And its use in a given context does not in any way
affect the meaning of the URI -- i.e. what it denotes.

> Further, I assumed that representations are properties of a resource, 
> that is:
> 
>      <http://example.org> hasRepresentation "<http><head>..." .
> 
> When I type <http://example.org> in my browser I would see one of the 
> representations associated with it.
> 
> But in fact, the actual model is different: It's not the 
> resource that 
> is associated with a (set of) representation(s), it's the 
> URI, because 
> two URIs can denote the same resource yet have different sets of 
> representations.

Right. As I suggested in my other post, dereferencing a given
URI can be thought of as working from a particular intensional
viewpoint, and the representations provided in terms of that
intension will differ from other intensional viewpoints, per
other URIs, even if the denotation of the URIs is the same.

> In fact, two people can create two URIs, denoting the same 
> resource, but 
> having different representations, and yet think exactly the 
> same things 
> about a resource: 
> If the W3C and I both create a URI denoting 
> the W3C, 
> we can have exactly the same idea of what the W3C is, yet our 
> URIs can 
> have different representations (web pages).

Sure. Why not.

> So a URI does *three* things:
> 
> - it denotes a resource

Yes.

> - it's associated with authoritative information

Not necessarily. Though URIs which contain a web authority
component are more useful for obtaining authoritative
representations and descriptions than those which have no
clear authority component.

> - it is associated with a set of representations

Again, not necessarily. Though distinct URIs can reflect
particular intensions of the thing denoted.

--

There is no requirement that different URIs that denote
the same thing resolve to different sets of representations.

There is no a priori basis for presuming that two URIs
taht resolve to the identical set of representations
denote the same thing.

Such relationships, or lack thereof, must be stated
explicitly to be the basis of SW agent behavior. One
cannot make reliable deductions based on observation 
alone (though one might concieve of various heuristics
for guessing possible relationships between resources
based on the HTTP observable nature of their URIs).

Patrick

--
Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com

Received on Monday, 28 July 2003 11:22:58 UTC