xmlhack: Does XQuery fit all?

While XQuery is usually considered an inappropriate tool for querying RDF collections or Topic Maps, Jonathan Robie demonstrated at XML 2001 that using XQuery functions over their normalized forms can be practical and not especially complex.

In his presentation, challengingly entitled "The Syntactic Web", Robie tries to reconcile the abstraction of the semantic world with its concrete syntactic representation:

This title was not chosen as a criticism of the semantic web, but to emphasize that there should be a straightforward syntactic representation of logical models so that general purpose tools can leverage the content of the semantic web. Syntax and semantics are not separate enterprises - they go hand in hand. Every syntax should have a clear underlying semantics, and every semantics should have a syntactic form that clearly expresses its underlying model. This should include schema-level information as well as instance-level information; for instance, our RDF internal format represents the entire RDF schema with assertions, making it possible for schema-level constructs to be evaluated in XQuery.

Having defined normalized XML representations for both RDF and Topic Maps, Robie created some simple XQuery functions to manipulate high level concepts of each language and to perform the "joins". He showed that when using these functions over the normalized representations XQuery queries were practical, readable and had roughly the same level of complexity as those written with specialized languages.

Acknowledging the help of a number of RDF and Topic Map experts, Robie noted: "they do not all agree with the conclusions I have drawn".