RE: xslt 2.0 xsl:copy copy-namespaces

> Surely it's a common enough requirement? Strip namespaces 
> comes up here
> often enough.
> Omitted for purist reasons? Not requested?

If one tried to design an instruction that does a "copy this tree putting
all elements and attributes in the null namespace", you would quickly want
to extend it to put elements and attributes in a different namespace, to map
namespace A to namespace B, to apply to elements but not attributes, to
change the elements into attributes while you're about it, and so on. All
this and more can already be achieved using the identity template with
variations.
> 
> (Do you agree about the header variances wrt your xslt 1.0 book Mike?
> Heck of a usability difference)

Yes, the lack of alphabetical running heads in the book is a big mistake. I
spotted it at the page proof stage and was told it was too late to correct
it. The root cause, I think, was that the editorial and production processes
at Wiley's weren't well enough coordinated.

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