> Obviously this suitability test can be stretched into something absurd,
> e.g.:
> 
>        input:
> 
>                <number-of-primes>6<number-of-primes>
> 
>        output:
> 
>                <primes>1,2,3,5,7,11</primes>
> 
> Using XLST to do that computation would not be a great idea :-)...

Why do you think so?

See the XSLT solution of the Google's "Find the first sequence of 10
consecutive digits of e, which is a prime number"

at:
    http://www.xsltblog.com/archives/code/code_of_the_day/index.html


The code is very compact and understandable, using such great XSLT 2.0
and Saxon features as xsl:function and the saxon:memo-function
attribute.

The solution is amazingly fast and no doubt can be further improved.


Cheers,
Dimitre.