W3C home > Mailing lists > Public > www-tag@w3.org > February 2008

Re: [httpRange-14] What is an Information Resource?

From: Tim Berners-Lee <timbl@w3.org>
Date: Fri, 1 Feb 2008 11:38:14 -0500
Cc: "Booth, David (HP Software - Boston)" <dbooth@hp.com>, Ed Davies <edavies@nildram.co.uk>, "Sean B. Palmer" <sean@miscoranda.com>, "www-tag@w3.org" <www-tag@w3.org>
Message-Id: <017C53CE-5822-4361-96B8-846ADF55F3F3@w3.org>
To: Ian Davis <lists@iandavis.com>

On 2008-02 -01, at 10:19, Ian Davis wrote:
> [...]
> To follow on from your dog example, the missing information I need is
> the assertion that:
>
> ex:Dog owl:disjointWith awww:InformationResource .


> it's intuitive that if ex:Dog is the class of real-world dogs then the
> above is true,

Yes.   I have a few statements like that kicking around


> but it's less clear that
>
> ex:RdfGraph owl:disjointWith awww:InformationResource .
>

This is because ex: hasn't defined RDFGraph completeley,
and specifically, hasn't defined the relationship to InformanResource.

In the cwm world, thee is for example a property log:semamantics
(don't argue about the name of it) which relates
an information resource to an n3:Formula.
It is backed up by code.
There are built-ins which operate on formulae, and do not operate
on InformationResources.   It is clear that Formula is disjoint from
InformationResource.  The code won't let you take the log:semantics of  
a Formula.

This is Philosophical Engineering, not Experimental Philosophy.
We are at core designing (have designed)  a system, not observing the  
world.
We define that an N3 Formula is disjoint from IR, we build code
around it, systems work.


> or
>
> ex:XmlNamespace owl:disjointWith awww:InformationResource .
>

Similarly, the notion in common parlance of XmlNamespace is not
tied down to a class in any engineered architecture.   There are  
various related things people have discussed: the namespace document,  
the set of URIs starting with the given string, the set of such URIs  
which are documented in the namespace document, the set of resources  
with URIs starting with that string, and so on and so on. Until  
someone needs an ontology of the space including how things relate,   
and is going to write code which uses it, the discussion about what we  
should chose to use the term Namespace for, (often misstated  "but  
what actually IS a namespace?") is not useful.


> And the everyday website operator needs to know those kinds of  
> things to
> configure their web server. How do we help them?

We write code which asserts that foaf:Person owl:disjointWith  
awww:InformationResource, and which concludes <a> a  
InformationResource when GET a  returns 200. And complains when  
something is the member of two disjoint classes.   So if someone  
writes a FOAF file about themselves using the URI of the file as a URI  
for themselves, it gets caught.

(We have that intent in the tabulator, but we don't do disjoint yet.   
It does warn one when document uses a URI which returns a 301 Moved.)



> Ian
Received on Friday, 1 February 2008 16:38:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:52 GMT