Comparison between original rippling and case-analysis technique, and
an extended version which allows non-rippling goals after splitting to
be conjectured as lemmas.

4 extra theorems can be proved by the extended technique, but not by
the original:
--------------------------------------------------------
drop n (xs @ ys) = drop n xs @ drop (n minus (len xs)) ys
len butlast xs = (len xs) minus (suc 0)
count n t + count n [h] = count n (h # t)
take n (xs @ ys) = take n xs @ take (n minus (len xs)) ys


2 theorems proved by original technique times out for extended
version:
--------------------------------------------------------
max (max a b) c = max a (max b c)              
min (min a b) c = min a (min b c)