> It's not the key() function that's reducing it to a single node, it's the 
> generate-id() function you are running on the nodes returned by key(). .4]
  
[snip]

> Beyond this I think Tom S. explained it.

Yes, what I meant was that perhaps I should be a little more explicit in
what I'm passing into the key function when I generate the  keys.  However,
I find that I actually do understand what's going on now, although my
understanding on how to get keys to always return the results I expect
remains a little foggy (more experimentation needed).

>But I'm confused not having the examples in front of me. Would you be able 
>to snip them out (again) and recapitulate the question if I'm not getting 
>what you're asking?

At this point I don't think its worth you spending more time on it.  I just
made one other reply a bit ago that may help explain the bit you where
"missing".  It does have some open questions that I could use some opinions
on, but I'm pretty much ok with the way the grouping is currently working. 

The exception is the open question on  why I can't get the key to replace
the //dataset reference to work.  That's what I'll play with over the next
couple of days.  I may end up posting back to the list again, but I'm
starting to suspect there's a way to restructure the whole set of templates
to deal with the data with a different line of attack that may make the
whole issue go away...



 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list