[xsl] select text of current element and not identical child element for word count

[Karen McAdams <kemcadams@xxxxxxxxx>]

I am currently collecting word counts of an entire
element and identical child elements in one sum. What 
I need now is to get the count for only the current
element without the identical child elements. 

My templates match the  element based on a class
attribute that contains "topic/topic." 

The structure is like this (its dita structure
language)

element [class = "topic/topic other/stuff"]
element title 
element body
element ref links

 and there can be up to 3 more similar elements nested
child elements like the this element.

I could try an approach in which i copy the contents
of the element sans its child elements of the same
type or create a mode similar to the way they are used
in this example.
http://www.dpawson.co.uk/xsl/sect2/N3235.html#d4106e196

not exactly sure how to write these templates - any
suggestions?




--------------------
Karen McAdams (650) 302 - 0830