RE: [xsl] Newbie question on XSL transformations: multiple sorts on element attr

["Michael Kay" <mike@xxxxxxxxxxxx>]

> Because I'm trying to obtain best performance in the business 
> logic and separate logic from presentation.
> For example if my business logic takes 300 msec and my xslt 
> trasform takes 30 msec, I'm happy because even if xslt 
> trasform takes 10% I have separated data from presentation. 
> The same logic in perl takes 600 msec, and xslt trasform 
> takes always 30 msec  or similar and so on.
> So I'd like to use saxon in c because it offers xslt2 support 
> but in the same way I'd like to continue to use c because my 
> business logic is faster.
> This my modest opinion.

So, where do you see the problem in calling an XSLT processor written in
Java? Is it because you think the transformation will take 50ms instead of
30ms (which may or may not be true), or is it because of the performance
overhead of the C-to-Java (or .NET) call, or is it because a
cross-environment call is just too much hassle? I can understand the last
argument - the more technologies you mix, the more hassle it becomes - but I
suspect the performance argument is spurious.

In any case, the answer to your question is no, there are no plans for a
version of Saxon written in C. Or assembler, for that matter. There are too
few hours in the day to write in low-level languages, and it's much better
to spend the time on writing better algorithms in a higher-level language.

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