S Woodside wrote:

It's so wonderfully backwards. Whoever thought of a variable that
contains a choose statement? :-)

Backwards, pshaw!

XSLT is a declarative language. You aren't instructing something to name a
memory location 'x' and then store value 'y' in it; you're asserting that
there is a identical relationship between 'x' and 'y', period. It's just like
math.

In XML, the element/content hiearchy makes expressing such relationships a
matter of containment: <x><y/></x>. <y/> can be written as <xsl:foo/> because
an XSLT instruction is a type of function; it can be used interchangably with
its evaluated result, since f(x)=y: f of x *is* y.

So I'm curious... which XSLT processors cache and don't lazily evaluate
extension functions? And aren't extension elements, in theory, in the same
boat? (If no one knows what I'm talking about, never mind...)

simon
k
---
www.simonwoodside.com