W3C

Using RDF to model XML structures

XML is a relatively simple technology, by itself. But the collection of specifications built on XML at W3C and elsewhere is growing at an extreme pace. W3C has launched a Quality Assurance activity in recognition of the need for, for example, more uniform testing and translation practices. Formal methods are being applied to clarify specifications such as RDF (cf RDF Model Theory) and XML Schema (XML Schema: Formal Description). But these post-hoc clarification efforts are surely not optimal.

Semanticists should be obstetricians, not coroners of programming languages.

-- John Reynolds

In the Advanced Development component of the W3C Semantic Web Activity, we're testing the hypothesis that application of XML and RDF to a wide variety of W3C's operations will lead to an increase in quality and efficiency. In this paper, we explore the application of RDF to modelling of XML structures.

The XML Information Set complements the XML 1.0 specifiction with an abstract model of the information in an XML document. The model of an XML document is a collection of information items; named properties associate items with values.

The item/property/value structure of the XML information set is the same shape as the subject/predicate/object structure used in the Resource Description Framework (RDF). An RDF Schema for the XML Information Set is an April 2001 W3C Note that takes the terms from the XML information set, considers them as RDF properties, and constructs an RDF Schema as a result.

But the resulting RDF Schema does not capture many of the interesting features of the structure of the infoset; for example, that elements have many children but (at most) one parent.

Using additional RDF vocabulary from DAML+OIL, we can capture the structure of the XML information set much more completely. A transcription from the text of the infoset spec into RDF using DAML+OIL terms follows point by point; it was an afternoon's work.

The resulting schema is usable with generic RDF tools; ising some tools for visualizing RDF as circles-and-arrows diagrams, we have produced a UML-like SVG diagram of the schema.

Background/Fodder

@@more about how XML Schemas talk about documents; RDF documents talk about all sorts of stuff, including XML documents. (mention DSD?)

simpler infoset diagram. hmm... same as xpath data model?

@@xml schema component model is also a DLG. diagram. vs diagram from spec

Related work

Dan Connolly, Dan Brickley@@Apr 2002
$Revision: 1.16 $ of $Date: 2002/07/15 22:21:38 $ by $Author: connolly $