> My question is, do people consider this a valid interpretation of the
> draft recommendation, 

If it's in XSLT then no, and not really within the bounds of what can be
done via the extension mechanism either, ie something implementing that
would be a non-xslt/xpath system. (Not that there is necessarily
anything wrong with having a user option to become a system with
extended features, so long as people know that's what they have)

However I suppose if it's just an Xpath application then you get to
decide the host language embedding and at a stretch you could declare
that the initial  context includes a binding of every possible
QName to the value false.

However this seems to be against the spirit of the thing, are you sure
you want this?
>  or should I make it an optional feature?

Definitely don't make it the default behaviour,

David


-- 
http://www.dcarlisle.demon.co.uk/matthew

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list