On Mon, Sep 8, 2008 at 23:25, Wendell Piez <wapiez@xxxxxxxxxxxxxxxx> wrote:
> James,
>
> If this were my code, I'd be sorely tempted to encapsulate the logic in a
> function, if only to make it more legible and maintainable. This would also
> provide you a nice neat place to document it and its rationale.

A good suggestion. I'm trying to do that more and more.  (Though if
I've asked a question I always like to send out a copy of what I chose
in case anyone else later reads the thread. And once or twice that
later person has turned out to be me!)
> Something like:

function snipped
-- though makes me curious... how hard would it be to change this so
that the groupings were based on the count() of items in the nodeset?
i.e. make divisions that were basically equal regardless of the input?
 So if there were almost no entries in the first half of the alphabet,
the first division would be a-r, then s-t (if a lot there), then u-w
(if few there), etc.  maybe passing how many sets to split it into
total?

> BTW, why the double-escaping of '&amp;amp;'? I can't see how that would do
> anything for you. The processor will see only the string '& a m p ;' here,
> which translate() will decompose into characters.

I have no idea how that ended up like that in my post, it isn't like
that in my actual XSLT.... weird. Unless I mistyped it and then
corrected it after pasting, or pasting into gmail did something weird.
More likely the former.

-James