XSchema Question 1: RDF

[Dan Brickley]

On Mon, 1 Jun 1998, Simon St.Laurent wrote:

> We need to get the RDF discussion moving again so we can move on to specific 
> syntax.  Dan Brickley suggested that I take a look at the W3C's P3P proposal 
> for one example of how RDF/XML specifications can look like plain old XML 
> without incurring much pain.  (I also hope he'll feel free to kick me if I get 
> this discussion wrong.  Everyone else is welcome to join in as well.)
> 
> The proposal is at http://www.w3.org/TR/WD-P3P10-syntax.  Like nearly all of 
> the W3C documents we're discussing here, this is a _Working Draft_, not a 
> final spec, and is subject to change.

A related angle on this is that it would be possible for XSchema instances
to be legitimate RDF-in-XML documents (ie. with an agreed mapping from
XML elements-and-attributes to RDF nodes-and-arcs) without committing to
being a full RDF application. 

The 'syntax' bit of the RDF Model and Syntax paper defines _two_
syntaxes for representating RDF graphs in XML: the normal 'serialisation'
syntax and an 'abbreviated' syntax. RDF-aware software needs to be ready
to parse both syntaxes. Given the time scale, this aspect of RDF could be
a little daunting for the XSchema project. 

However, this doesn't mean that XSchema  couldn't be 'RDF-friendly'. It
might be, for example, that an RDF model reprenting XSchema concepts gets
mapped out, and a single XML representation using one of the RDF-in-XML
syntaxes is defined. This scenario would differ from XSchema being a full
RDF application as follows:

1) XSchema documents  would look exactly like serialised RDF graphs using 
  a syntax compliant with the RDF Model and Syntax spec 

2) XSchema documents would be mappable into an RDF graph that
 represents DTD/XSchema concepts in terms of
 resource/propertytype/value tripes

3) Not all RDF serialisations of the graph described in (2)
 would count as legal XSchemas.  [ this is the halfway house bit! ]

In other words, we agree an RDF model representation, pick our
favourite XML syntax for this and then say that XSchema-aware software is
_not_ expected to be capable of understanding of the other equivalent
representations defined by the RDF Model and Syntax spec. This would 
be a compromise similar to that sketched in the P3P draft.

As Tim Bray points out, the RDF model is at core rather simple, even if
the gory details of the spec takes a bit of grappling with. It
shouldn't be unfeasibly hard to represent XSchema ideas in terms of
assertions framed as RDF triples:
 "the ____ property of the resource known as ____ is _____"...

I suspect this would be a useful exercise in itself, even if it was
decided that XSchema syntax couldn't (for whatever reason) be RDF
friendly. 

The only other potential stumbling block is namespace prefixing:
RDF uses namespaces, so an RDF-friendly syntax would hopefully do
likewise. So a concrete question to ask might be: do we agree that it's
acceptable to spend those extra bytes namespace-prefixing the elements and
attributes in an XSchema document?
 
Dan



--
Daniel.Brickley at bristol.ac.uk                           
Research and Development Unit                    tel: +44(0)117 9288478
Institute for Learning and Research Technology   http://www.ilrt.bris.ac.uk/
University of Bristol,  Bristol BS8 1TN, UK.     fax: +44(0)117 9288473 



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)