RE: [xml-dev] DOM or SAX: Sense and Sensibility

["Michael Kay" <michael.h.kay@ntlworld.com>]

> Is anyone working on any XSLT implementations that work with
> Jaxen? That
> would be very cool.

In recent releases of Saxon I've been following a similar direction, of
making the XPath engine able to work with multiple document representations.
The current release includes a JDOM driver to illustrate the principle. I
decided to make the interface between the XPath engine and the document
representation work at the level of the XPath axes: essentially, the driver
has to support a Node object, plus iterators over the 13 XPath axes. This
seems to offer a high enough abstraction layer to work efficiently with a
wide variety of different stored data structures. Interfacing Saxon to DOM4J
shouldn't be at all difficult.

The two things that are hard to do are (a) efficently sorting nodes into
document order, and (b) implementing the XSLT whitespace stripping rules.

Incidentally, some of the things DOM4J and Jaxen are doing look very similar
to the original "Java only" Saxon library before it changed direction to
become an XSLT engine.

Mike Kay