Re: [xsl] // expanding to descendant-or-self::node()

[Evan Lenz <evan@xxxxxxxxxxxx>]

Also, I've occasionally run into an unfortunate, seemingly arbitrary 
consequence of this design, due to the syntax-oriented restrictions of 
XSLT patterns: you can use "//" in patterns but not "descendant::" 
(except in predicates).

I know there have been situations where this annoyed me, but I'm having 
trouble thinking of an example at the moment.

Evan

Michael Kay wrote:
> > Why does // expand to descendant-or-self::node() rather than 
> > to descendant::x where x is the next item in the path?
> >     
>
> I've always imagined it was primarily to make //@x work; though I'm not at
> all convinced it was a smart decision. Certainly in practice it's very rare
> to see any axis other than "child" after "//" (for example //.., or
> //following-sibling::x is never seen in practice and never required; and
> //@x is something one could live without).
>
> I think making "//" expand to "/descendant::" might well have been a better
> design.
>
> Michael Kay
> http://www.saxonica.com/