partitioning instead of matching so much



I've written up some programs that are largely alternatives to matching. 
I show how the "derivative divides" symbolic integration in Joel Moses' 
SIN program,
one of the initial methods in Maxima's program, can be written in half a 
page or so,
probably more effectively than already in Maxima.  It might be 
interesting especially
for fans of table-driven programs for integration or other tasks.

comments are of course welcome.
http://www.cs.berkeley.edu/~fateman/papers/partition.pdf

The text, if you want to extract it and use it yourself, is in the same 
place, partition.tex.

RJf