Geppetto is a C library which assists in writing genetic programming
applications, just like SGPC.  I *think* it's got pretty much everything
SGPC has (I haven't tried multiple populations, but it *should* work...)

In addition it has:
	Typed objects (boolean, short, int, long, float, double, list)
	All data types can be combined in a single program
	Data types can be used to control the way programs are built
	Erroneous programs can be discovered and labelled as unfit

I built a symbolic regression application using both SGPC and Geppetto and
compared the speed and accuracy of 100 runs of each in order to see how much
speed I was giving up for all these extra features.  To my surprise, I
discovered that Geppetto was roughly twice as fast!

This directory contains the source code for the library and a few sample
applications, as well as a user's guide, a bunch of tools which will probably
be useless (but are included "just in case") and the code for the SGPC
application I used to compare the two systems.  The 'TODO' file also lists
some problems with Geppetto that I plan to fix someday.

I play with Geppetto in my free time, so I can't guarantee that I'll
be able to fix any problems you have, but feel free to send them to
dglo@CS.Berkeley.EDU and I'll see what I can do...
