Re: resources and URIs

pat hayes wrote:

>
> And BTW, it is exactly that phrase "identified by URIs" which is
> giving me problems: I *still* don't know what it means: but Im pretty
> sure that what Dan Connolly takes it to mean isn't exactly what Roy
> Fielding means by it.  Dan's version is much more inclusive and
> all-encompassing than Roy's. For example, RFC 2396 says that once the
> resource has been identified from the URI, operations can be
> performed on it (the resource). Good luck performing operations on an
> active galaxy or an imaginary character.

A GET operation on either with an Accept: image/jpeg might return a picture.

Are you worried about not being able to PUT or DELETE?

...
>
> >  You, on the other hand, are profoundly disturbed by the notions
> >that (a) the weather in Oaxaca can be a resource, i.e. an object in
> >a networked information system
>
> Yes, that seems like complete nonsense to me. Or at any rate, the
> only way I can make sense of it requires me to treat 'object in a
> network' as essentially meaningless. If anything can be 'in a
> network', including remote galaxies, tomorrow's weather and imaginary
> characters, then we aren't talking about network architecture in any
> sense I am aware of.

Pat, you really need to get out more :-))

The term "network" has in some contexts been used as synonymous with
"graph", for example "neural network" which dates back approximately to the
1950's. Or the ?German "feltwork"
http://www.fasthealth.com/dictionary/f/feltwork.php which dates back to the
1800s.

The implication is that there are rich connections between the elements.

>
> >, and (b) that an HTML page can be a representation of it.
>
> I have absolutely no problem with the HTML being a representation, or
> its being a representation of the weather: in fact I think this is
> perfectly correct. I even don't have too much trouble with its being
> a representation of a resource (though I don't think that can
> actually make strict sense if taken literally). What I can't do is
> swallow all this together at once, particularly given all the other
> stuff said about resources..
>
> >Do I have this straight?
> >
> >There is no way to know if it is correct or not since it doesnt seem
> >to mean anything that can be tested for truth.  Am I identified by a
> >URI? How could anyone possibly tell?
> >
> >You're right; the current web architecture provides no way to test
> >this condition.  I had (perhaps naively) understood that the
> >Semantic Web was going to give us the machinery to make
> >machine-usable assertions about semantic classes of resources and
> >their relationships.
>
> The semantic web cannot possibly decide what the words you write are
> supposed to mean, surely?

Well to some extent: WordNet, but at some point we look to what is within
the <rdf:comment>

...
>
> They are manifestly NOT adequate to this task when the software is
> obliged to process the representations *as representations*, ie in
> ways which reflect their intended meanings. There is a large body of
> technical literature, much of it in computer science, which uses
> words like "representation of" and "network" with some precision. All
> I have been doing is using that technical understanding to read your
> documents, and they don't then make overall sense.

I think we have a rather classic example of the document having a set of
interpretations. You are interpreting the phrase "information network" in a
certain fashion, which excludes the interpretations that other reasonable
people have. Perhaps the document should define this phrase so that it
entails what we consider a reasonable range of interpretations, i.e. if we
intend to say that a galaxy might be a component of an information network
then let's make this more explicit, and if not, then let's make that more
explicit.

Jonathan

Received on Friday, 18 July 2003 19:13:08 UTC