My 2 cents on these issues, with some repetition of earlier comments.
On Mon, 2004-01-26 at 13:38, Roland Schwaenzl wrote:
> Dears,
>
>
> well there is quite some trouble with the notions in the model paper.
I think you are exaggerating quite a bit. Any unclear language can of
course be fixed. Let's focus on trying to improve the document.
> Ex: resource URI: a URI that identifies a single resource
>
> Question: Does there exist URI's which identify none or more than one resource?
None: probably yes, but is that interesting from a resource description
point of view? It won't be described, so there wont be any data about it
to handle.
More than one? In practice, this will probably happen. However, it only
affects the poor confused person using the name to identify two things
:-)
I think the assumption is that [URI -> resource] is actually a function
(with unknown and evolving domain).
>
> Question: Does the paper treat "URI" and "URI reference" as synomyms?
I think we need to talk exclusively of URI references. For those not in
the know, the URI RFC (number?) defines a URI reference as
URI [ # [fragment]]
that is, a URI with an optional fragment identifier (so a URI is a URI
reference, too). On the web, the fragment ID is never seen by the web
server. It is only used by the client.
RDF uses only URI references, not URIs, and I think we need follow that.
> Question: Is there an intrinsic criterion, which makes a "value URI" a value URI?
No. As has been said, it is a resource URI (the resource in this case
being the value) appearing in a certain context.
> Question: Is there an example of a gadget, which is NOT a resource?
For all practical purposes, no. Again, is this important? We deal only
with resource descriptions, so everything is already a resource.
/Mikael
--
Plus ça change, plus c'est la même chose
The more things change, the more they stay the same
|