W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > June 2001

Re: #rdfms-difference-between-ID-and-about

From: Sergey Melnik <melnik@db.stanford.edu>
Date: Fri, 15 Jun 2001 15:58:07 -0700
Message-ID: <3B2A92FF.60CA6705@db.stanford.edu>
To: Dan Connolly <connolly@w3.org>
CC: guha@alpiri.com, Aaron Swartz <aswartz@upclink.com>, Brian McBride <bwm@hplb.hpl.hp.com>, rdf core <w3c-rdfcore-wg@w3.org>
Currently, the concept of relative URIs is present in the M&S 1.0
serialization syntax only, and is not reflected in the model in any way.
That is, relative URIs is a just another "abbreviation feature" of the
M&S syntax, and I'd like to emphasize that in the revised spec to avoid
misconceptions. As abbreviation features, relative URIs have nothing to
do with the fragment IDs described in
http://www.w3.org/DesignIssues/Model.html#Fragement (for example, a
fragment ID in an RDF document does *not* refer to a portion of this
document). Therefore, IMO the axioms and benefits summarized by TimBL
for HTML at above link do not seem to hold in the RDF world.

I feel strongly about not using relative URIs, especially given the
parsing/editing complication (see prev. postings in this thread), and
the problems associated with moving RDF documents from one location to
another. However, I think keeping cleaned-up rdf:ID, rdf:about and
relative URIs in the revised M&S syntax is essential for backward
compatibility and is the least bloody compromise (sigh...) I'd just
suggest to word it so the developers understand the pros and cons.

Sergey

Dan Connolly wrote:
> 
> "R.V.Guha" wrote:
> >
> > Ok, Aaron, you hit the nail on the head.
> >
> > RDF absolutely has to make sense even outside the context of
> > an enclosing document which can be given a uri. so ...
> 
> So... what? That doesn't make any sense to me.
> 
> An RDF document is an XML document. Each XML document
> has a base URI (cf the infoset spec).
> If you copy the contents from one
> place in the web to another, you get a different XML
> document, and hence a difference RDF document; if
> it uses relative URI references, the resulting triples
> may be different.
> 
> This is by design.
> 
> This design does allow users to goof,
> but it also allows folks to manage collections of
> documents and by and large, it has succeeded over the course
> of the last 10 years.
> 
> Noone is forced to use relative URI references; anyone
> who uses them does so by choice. Surely the consequences
> of that choice for RDF documents should be the same
> as the consequences for HTML, XML, PDF,
> and other document formats in the Web, no?
> 
> Or rather: surely the the consequences *are* the same for
> RDF as for XML in general; we're not designing RDF 1.0 today;
> we're just clarifying the spec; and the spec is already
> pretty clear on this:
> 
>   [[[ The value of the about attribute is interpreted as
>   a URI-reference per Section 4 of [URI]. The corresponding
>   resource identifier is obtained by resolving the
>   URI-reference to absolute form as specified by [URI].
>   ]]]
> 
>   --        Resource Description Framework (RDF) Model and Syntax
> Specification
>   http://www.w3.org/TR/REC-rdf-syntax/
>   Wed, 24 Feb 1999 14:45:07 GMT
> 
> The only question is about this sort of fuzzy text:
> 
>   [[[ The ID attribute signals the creation of a new resource ... ]]]
> 
> But this text in particular suggests pretty strongly
> that rdf:ID="foo" means the same thing as rdf:about="#foo" :
> 
>   [[[ The ID attribute, if specified, provides the URI
>   fragment identifier for c.
>   ]]]
> 
>   -- section 6. Formal Grammar for RDF
> 
> --
> Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Friday, 15 June 2001 18:32:02 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:37:10 EDT