> Using the document() function was my previous implementation step: as 
> I'm on the optimization stage now and I need to apply the same 
> transformation to *multiple* XML sources, I'd prefer NOT to 
> load *each* 
> time that *same* nodeset but to pass it as a parameter.

Don't assume that if you call document() twice on the same URI, the document
is automatically loaded and parsed twice. Details are
implementation-dependent.
> 
> As I need such a parameter, IMHO I think that would be much 
> more elegant 
> to serve a nodeset already purged from ignorable whitespaces, 
> instead of 
> having to mind *existing* whitespaces...
> 
> Is there a way to force the XPath parser to spit a resulting Node 
> *without* ignorable whitespaces?

Try building the document using a transformation that strips spaces and does
nothing else; or write a simple SAX filter to do it.

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