> > Ultimately i am trying to check if the items in $s1 can 
> also be found in $s2.  
> > For example - 
> 
> as these are both subsets of the same set of nodes you presumably mean
> node identity here rather than value equality. You don't need 
> any loops
> at all for that
> 
> count($s1|$s2)=count($s2) 
> 
> is true just if all the items in s1 are in s2.

I read the problem as being based on equality rather than identity of nodes,
but I might not have read carefully enough...

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