RE: XSLTPROC performance

> I don't agree though - if you are comparing say Gestalt Win32 
> version with Saxon running on Windows, then surely it's only 
> really fair to make the comparison after the Java bytecode 
> has become native machine code - after JVM startup.

The only fair comparison, quite honestly, is the one that best simulates the
actual workload you need to perform. And that's highly variable.

The JVM startup cost can be very irritating for certain kinds of work. But I
tend to reckon that waiting a second or two is rarely more than an
irritation; for workloads where performance is critical, there are ways of
building your application to avoid this cost.

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