> As far as I can tell, you will have a real problem here.
> In XSLT (1.0) alone, you cannot get the character code from a 
> given char.
> 
> Therefore, you will have to try out every possibility of a given
> character. This is barely feasible using only ASCII characters, but if
> you need more, then you are really bad off.
> 
> You will have to go recursively through the URI in question, escaping
> the characters that are not acceptable as they are. The escaping
> itself being done by a huge choose-when sentence covering all
> possibilities.

It's certainly not a problem I would want to tackle in XSLT 1.0, but I don't
think it's as bad as that. You can classify characters using the translate()
function.

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