There is a lot to be done.  Here are some open areas that need to 
be addressed.

We hope that the users of KEIM will contribute some of their expertise and
code to the KEIM library.

1.  Bound variables are currently implemented using names, so when we do
    lambda-contraction, we potentially have to do a lot of renaming.
    It would be nice to have a DeBruijn-style implementation of abstractions,
    or better, something like a hybrid, so that only in printing do we
    generate names of variables, but based on the original parsed input.

2.  Unification is a little weak.  We need better, faster, unification
    algorithms.  In particular algorithms for different types of sort
    systems.

3.  Speaking of sort systems, we plan to extend KEIM to allow sorts. 
    The details have not yet been decided. 

4.  It would be nice to have some small sample implementations using KEIM.
    See the example resolution prover in keim/prog/examples.
    
5.  Overall efficiency needs to be examined.

6.  The extraction of documentation should be easier to extend.

