W3C home > Mailing lists > Public > www-tag@w3.org > July 2002

Re: TB16 Re: Comments on arch doc draft

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Wed, 03 Jul 2002 17:03:02 +0300
To: ext Jonathan Borden <jonathan@openhealth.org>, ext Tim Bray <tbray@textuality.com>, WWW TAG <www-tag@w3.org>
Message-ID: <B948DEC6.17CD5%patrick.stickler@nokia.com>

On 2002-07-03 15:53, "ext Jonathan Borden" <jonathan@openhealth.org> wrote:

> Patrick Stickler wrote:
> 
>> 
>> A term in some vocabulary is an abstract resource, and if the
>> URI denoting that term dereferences to anything, then that is
>> a bug, since it is impossible to actually obtain a representation
>> of an abstract resource. Similarly, a namespace is an abstract
>> resource, and thus, if a URL is used for the namespace name,
>> having it resolve to anything is IMO a bug, just as for any
>> abstract resource.
>> 
> 
> This argument is a wonderful example for which the phrase "begging the
> question" is properly applied. Such perfectly circular examples are not seen
> frequently. I applaud Patrick on the construction of this paragraph.

Thank you, though I don't see that it is circular. You seem to
be having some difficulties with vocabulary. See below...

> We give various names to representations or serializations of vocabularies
> such as "schema" or "ontology" which are perfectly transportable over the
> Web. Since the antecedent to your argument is clearly false**, there is no
> reason to evaluate the truth of its consequence.
> 
> ** a) a term in some vocabulary may or may not be either abstract or a
> resource. 

No. A vocabulary term is always abstract. It is also a resource, if it
is named by a URI.

> For example if a term is identified by a URI reference, then it is
> not a _resource_ in the context of the term "abstract resource" e.g. RFC
> 2396 (resources you recall are identified by URIs not URI references).

That's not the view of RDF, and thus not the view of the Semantic Web.

Though I really don't see your point anyway. A URI or URI reference
does not determine whether something is or is not abstract or
is or is not web-accessible.

Using an http: URL to denote a rock in my back yard or the concept of
'spicy' does not make either of those things web-accessible or
the latter concrete.

And if those URLs really *do* denote those non-web-accessible things,
then putting anything else at that URL which would be accessible
is a bug.

I consider it a fundamental constraint of the Semantic Web that
a given URI or URI reference denote only one thing, even though
more than one URI or URI reference may denote the same thing.

Thus, using a mailto: URL to denote a person, or using a web
page URL to denote a company, or using a namespace
URI to denote both the set of terms and some document
describing things related to that set of terms are all errors.

If the resource is not web-accessible, then it should not be denoted
by a URI that is dereferencable -- or if it is dereferenced, nothing
should be returned.

> b) if a term in some vocabulary is identified by a URI and the URI resolves
> without error, then the resource is not abstract in that sense of the term
> "abstract resource".

Of course it is abstract. Again, the behavior encountered from
the dereferencing of some URI denoting an abstract resource
does not affect its nature. You might get *something* back,
but it's not a representation of the abstract resource. It doesn't
make the term concrete. And if you do get something back, that's
a bug. Eh?

Just as you might denote the city of Paris with a URL, and then
put a picture of Paris at that URL such that when you dereference
the URL, you get a representation of the picture. That's a bug,
since the URL was not meant to denote the picture, but the city,
and you're not getting the city, you're getting a picture of
the city, which is not the same thing.

But we've been over this before, Jonathan, many times.

> c) the typical situation when a term is represented by a URI reference,

Who ever said anything about URI references? Terms can be denoted
by URIs as well. Heck, just have a look at the Dublin Core
vocabulary. Nope, no URI references there.

> is
> that the URI resolves to a representation of the _vocabulary_ either a
> schema, ontology, directory etc. and the fragment identifier is used by the
> interested agent to identify the description of the desired _term_.

A schema is not a representation of a vocabulary. Sorry. Nope. It's
something that uses the terms of a vocabulary, but the vocabulary itself
is abstract, just as are the terms.

I think you are confusing XML instances and element IDs with abstract
vocabularies and terms.

A URI reference into a particular XML instance, such as an RDF Schema,
does indeed correspond to a relationship of instance to element ID, but
that URI reference does not denote the term, but a particular element
within a particular XML instance. That element may describe the term,
but it is not the term itself.

If you try to use that URI reference to denote both the abstract
term and a particular element in a particular XML instance that
describes that term, again, that's a bug.

Patrick

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com
Received on Wednesday, 3 July 2002 10:06:18 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:09 GMT