RE: Information resources?

> -----Original Message-----
> From: www-tag-request@w3.org 
> [mailto:www-tag-request@w3.org]On Behalf Of
> ext Dan Connolly
> Sent: 07 September, 2004 17:30
> To: Norman Walsh
> Cc: www-tag@w3.org
> Subject: Re: Information resources?
> 
> 
> 
> On Tue, 2004-09-07 at 09:17, Norman Walsh wrote:
> > The notion of "resources" and "information resources" is, from my
> > perspective, a compromise designed to allow two world views 
> to achieve
> > consensus. Consensus is a good thing and I don't object to 
> the notion
> > of "information resources" because they don't appear to do 
> any harm. I
> > must admit, however, that it is not a distinction that I 
> understand at
> > any intuitive level.
> > 
> > One of the problems that I have is, I don't know how to tell if
> > something's an information resource or not. And if I can't tell, I
> > don't understand the value of the distinction.
> > 
> > Here are some URIs. Which of these are information resources?
> > 
> >   http://norman.walsh.name/files/2004/09/car
> >   http://norman.walsh.name/files/2004/09/truck
> >   http://norman.walsh.name/files/2004/09/bicycle
> >   http://norman.walsh.name/files/2004/09/cars#dbw
> >   http://norman.walsh.name/files/2004/09/trucks#ndw
> >   http://norman.walsh.name/files/2004/09/widgets#car
> > 
> > They're all 404 today, so you'll have to decide based on something
> > else.
> 
> Well, they're your resources, so ultimately you'll have to decide.
> 
> The rest of us aren't licensed to.
>   http://www.w3.org/TR/webarch/#pr-uri-opacity

Well, that depends on what the term "information resource" is
intended to mean, versus what is understood to mean by the masses.

AWWW says "Any resource that has a representation is an information resource."
so if any of the above URIs resolve to representations, then they are,
technically, according to AWWW, information resources. Period.

However, I think that alot of folks (including some Tag members) will expect 
information resources to be particular *kinds* of resources, i.e. corresponding
to "bodies of information", and not simply resources which have web-accessible 
representations. And such an interpretation clearly goes beyond what is
specified in AWWW.

It is true that Norm must decide about his resources, and he could say things like

 <http://norman.walsh.name/files/2004/09/cars#dbw> rdf:type ex:InformationResource .

(whatever ex:InformationResource means) but saying such things about his resources 
doesn't require any particular definition in AWWW -- and any such statement would
anyway be outside the scope of the web machinery.

Hence my recommendation to (a) leave the term out and (b) define any RDF classes
for any particular kinds of resources that folks feel it's important to be
able to identify as such and talk about (though not in AWWW).

--

I reiterate Dan's reference to URI opacity, suggesting that any classifications
of particular kinds of resources is out of scope for the web architecture (layer).
Rather, that's the job of the SW layer. The web machinery itself cannot reveal
anything explicit about the nature of the resources denoted by the URIs. It can
only provide access to representations, and the interpretation of those representations
is also outside the scope of the web machinery proper.

The SW allows us to say what we want/need to say about the resources denoted by
those URIs. The web gives us access to representations of the resources denoted
by those URIs. Thus the web and SW are complementary -- two sides of the same
coin, and that coin is the shared URI space denoting the resources of interest.

Leave the classification of "information resources" to the SW layer, and leave it
out of AWWW.

Respectfully,

Patrick

--

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

Received on Tuesday, 7 September 2004 15:38:47 UTC