Re: Proposed issue: What does using an URI require of me and my software?

From: Dan Connolly <connolly@w3.org>
Subject: Re: Proposed issue: What does using an URI require of me and my software?
Date: Wed, 24 Sep 2003 08:43:18 -0500

> 
> On Wed, 2003-09-24 at 07:13, Peter F. Patel-Schneider wrote:
> > From: Dan Connolly <connolly@w3.org>
> > Subject: Re: Proposed issue: What does using an URI require of me and my software?
> > Date: Tue, 23 Sep 2003 12:41:24 -0500
> > 
> > [...]
> > 
> > > > 6/ I believe that one of the issues that needs to be resolved is what
> > > >    information is implicit in the use of a URI reference with optional
> > > >    fragment identifier, particularly in the case where removing the
> > > >    fragment identifier results in the URI that can be used to retrieve an
> > > >    RDF (or OWL) document.
> > > 
> > > Hmm... I don't see how that's any smaller than the
> > > whole rdfURIMeaning-39 issue.
> > 
> > I see that there is a broad reading of this, but I meant a much more narrow
> > meaning along the following lines:
> > 
> > 	If the notions of how information is recorded and gathered and how
> > 	the meaning/denotation/... of a URI reference is determined in a
> > 	particular context are given then how does one determine the
> > 	context in which to determine the meaning/denotation/... of a URI
> > 	reference.
> > 
> > I think that this is actually quite a bit smaller than rdfURIMeaning-39, as
> > it excludes issues of differing languages, inaccessible information, etc.
> > 
> > My solution, of course, is that under a representation regime context is
> > determined from an initial set of information solely by the explicit
> > importing constructs of that representation regime that import
> > web-accessible documents that are compatible with the representation
> > regime.
> 
> After reading that several times, I can now parse it as English
> (a comma ala "under a representation regime, context..." is sorta
> implicit, yes?). But I'm still having a lot of trouble understanding.
> How about a few examples? 

See below.

> What's a representation regime?

I was trying to produce a neutral term.   Examples would include ``the
model theory of RDF'', ``abstract ontologies in OWL'', ``first order
logic'', ``a unified theory for RDF, OWL abstract ontologies, and XML
Schema datatypes''.  

Under the first regime, context for an RDF/XML document is exactly what the
RDF (not RDFS) model theory gets you from the RDF graph for that document.
Under the second (not-quite-extant) regime, the context for a collection of
abstract OWL ontologies includes not only the information in the ontologies
but also involves following imports links and looking for other abstract
OWL ontologies there.  Under the third (definitely-not-extant) regime, the
context for a collection of documents could include all of the above plus
the results of some process that uses information in the documents to look
for XML Schema documents that provide definitions for the datatypes in the
documents.

> I'm not sure what "excludes issues of differing languages" means,
> but if it means disregarding things like WSDL and XML Schema,
> I doubt that solution is very interesting to the TAG.

No, it means that someone is going to have to state what sort of languages
can be used to write information.  For example, can XML Schema documents be
used to provide information for RDF?

> > Peter F. Patel-Schneider
> -- 
> Dan Connolly, W3C http://www.w3.org/People/Connolly/

Peter F. Patel-Schneider

Received on Wednesday, 24 September 2003 12:01:06 UTC