>  Hmm, I wonder, however, if the notion of node *set* is really
>  appropriate for most of the use cases here. Why did they not
>  call it node-sequence? The sequence of stuff is what we're
>  mostly interested in. Well, instead of wondering about the
>  specs, I discovered a related problem with your topological
>  sort. As beautiful as it is, it disturbs document order where
>  it isn't warranted by the dependencies. For example, let's
>  say I have 3 cliques of dependent nodes and no link between
>  them. What happens with top sort is that those three cliques
>  are all intermingled. The trick is to keep the document sort
>  order if it doesn't need to be changed. 

There's another algorithm possible, which traverses "cliques" by "cliques" (I think
these are actually called connected components).

But it's Christmas time now -- not healthful to work hard during the holidays...

To you and all xsl-listers -- My Seasonal Greetings.

Dimitre Novatchev.

__________________________________________________
Do You Yahoo!?
Send your FREE holiday greetings online!
http://greetings.yahoo.com

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