Re: XSLT processor performance

[Sebastian Rahtz <sebastian dot rahtz at computing-services dot oxford dot ac dot uk>]

I will add that into my script, thanks, Oliver.

the results are predictable, I guess:

xt
	0m2.863s
Saxon
	0m2.983s
Oracle
	6413 Segmentation fault      
Sablotron
      Error [code:201] [URI:file:/home/rahtz/Home/xsltest/eratosthenes.xsl] 
      [line:56] [node:attribute 'test'] wrong expression syntax
Xalan
	0m4.480s
4XSLT
	0m24.417s

TransforMiiX just does nothing, though I don't understand why.

as ever,

 - XT zaps through  like a rocket, but isn't conformant
 - Saxon is fast and conformant, and accepts anything you throw at it
 - Xalan is nearly conformant, but a bit slow
 - Oracle is fast and conformant when it runs, but can go horribly wrong
 - TransforMiiX is a bit of a loose cannon
 - Sablotron is promising, but is not mature 
 - 4XSLT can be made to work, but is slow

I doubt that anyone would be surprised to read that. If we assume that

 - any remaining bugs in Xalan will soon be fixed
 - the Ginger Alliance people are proceeding nicely with Sablotron 
 - Oracle will presumably iron out their oddities soon

then even if all else fails we'll have 5 or 6 decent cross-platform
interchangeable engines by the anniversary of the XSLT spec. pretty
good, if you ask me!

Sebastian


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list