> I think if you polled XSLT 2.0 users and asked them which features they find
> most valuable, grouping and regular expressions would come top of the list,
> followed by multiple output documents, date and time handling, user-written
> functions, ability to handle non-XML input, and atomic sequences (not per
> se, but the ability to do things like summing over computed values, which
> come as a consequence of atomic sequences).
Very near the top of my list would be the 'fail early' policy instead
of the 'never fail' policy of xslt 1.0, where the first-item-semantics
often meant you weren't aware of a problem until long after the
transform.
--
Andrew Welch
http://andrewjwelch.com
