Michael Kay has just released the Saxon 8.5 parser, in time for eXtreme Markup, too! (wish I could be there...)
Saxon is by far the most powerful, and speedy XSLT parser I've ever used. Sure, I started with Xalan, but after working a lot with the DocBook stylesheets, Saxon is the best way to go. To be fair, I have also used xsltproc, but Saxon is usually my first "go to" parser.
Oleg Tkachenko's blog has a much deeper explanation of new features, but basically Saxon 8.5 provides:
- hash join optimization for both XSLT and XQuery
- binary disk representation of validated source documents
- sequential XSLT processing of input documents without reading the whole document into memory for serial transformations
- integrated processing of an entire directory of files using John Cowan's Tag Soup parser
You can download the latest version of Saxon at: http://saxon.sourceforge.net/
See also: XSLT
