RE: Practical Suggestion for XSLT Performance Improvement

[Kay Michael <Michael.Kay@xxxxxxx>]

>   In this document a method is suggested for dramatically
>   improving XSLT processor performance (2-10x or more) 
>   for financial information or other 'high volume' 
>   information streams. It is shown how the time to 
>   produce a summary report for an 100K input stream 
>   changes from two and a half minutes to 14 seconds.  
>   It is also suggested how this number could be further 
>   reduced with an alternative XSLT processor design.

I'm slightly surprised the improvement is only ten-fold.

Have you tried a solution using SAXON's assignment statement? If not, I'd
like to try it, if you could send me your data files. Part of my thinking
when I provided saxon:assign was that if I didn't, people were going to
invent mutable variables by using external functions with side-effects,
which seems a much dirtier solution.

Mike Kay


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