RE: test if null

> Also, an aside -- I'm sure there's a reason why some people prefer 
> "string-length(.)!=0" to "string()" (or more robustly, 
> "normalize-space()") 
> -- but I've never been able to guess what it could be.

Slight unease, I think, at relying on implicit type conversion. It feels a
bit sneaky, like using 0 to mean false in C.

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