Re: following-sibling and xsl:sort

On 4/30/05, Wendell Piez <wapiez@xxxxxxxxxxxxxxxx> wrote:
> Karl,
>
> I think you may be misinterpreting what Dimitre said. At any rate, I took
> "should have a solution without using it" to mean "will also be solvable
> without it". No warrant is given as to whether that solution is "good" in
> any sense -- it could be very long, or require heaps of time and memory to
run.
>
> I don't believe Dimitre intended to imply that because you always have an
> alternative, you should not use xxx:node-set(). There is a set of cases for
> which the alternative you have may be theoretically possible, but
> prohibitively difficult in practice.

Thank you, Wendell, this is correct.

For example, FXSL for XSLT 1.0 uses heavily the exslt:node-set()
extension function -- the only extension function used in FXSL.

In fact, this made me implement a subset of EXSLT(of course containing
the node-set() extension)  for MSXML4 two years ago.

Cheers,
Dimitre Novatchev