> >> self::node() selects the context node regardless what kind 
> it is, the
> >> principal node kind doesn't come into it. So the above 
> expression is 
> >> fine.
> >
> I just can't match that up with the XPath 2.0 draft as of yesterday. 
> Sect. 3. 2. 1. 1. makes the principal kind of the self axis element. 
> 3.2.1.2 says that  for a name test be true, the pricipal kind of the 
> axis is the kind of the node. There's got to be something wrong 
> somewhere 

node() isn't a name test!

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