> I suppose we could produce the desired output by making a 
> second pass through
> this output using following-sibling to find unique directory 
> names. but I'd much
> rather do this in one pass.

Why? I think this a problem that is crying out for a two-pass solution.

Mike Kay


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