[xquery-talk] RE: Future of XQuery and XQuery Update Faclilty

[Michael Kay]

> On at least the surface it seems many of those problems are 
> elegantly solved by just using XML & namespaces. For 
> instance, by representing an XQuery function with an XML 
> fragment(in a standardized namespace), lets the XQuery 
> language itself do introspection, serialization and it fits 
> naturally into the data model.

That's certainly the right approach in some cases - in the XSLT world there
are countless applications that exploit the fact that XSLT stylesheets have
an XML representation, making them amenable to introspection; and I think
the same approach can also be used with schema information.

However, it has to be said (a) that there are some data structures that are
not easy to represent in XML, because of the restriction to hierarchies, and
because of unwanted baggage like node identity and document order, and (b)
that once you start representing objects such as "functions" in XML, you
really start to need some kind of encapsulation to separate what is visible
to applications from private data that should only be seen by the
implementation.

Michael Kay
http://www.saxonica.com/