From Don Smith:

I am working on a backward chaining intuitionisitic prover with
focussing.  I've written and tested, in Prolog, a propositional prover
with focussing.  (Because assumptions are copied and not removed in rule
focusL, the prover would loop, even for propositional formulas, if not
for iterative deepening.  Already, loop detection would be useful.)  I
plan to extend the prover to the first-order case over the next week or
two.
 
As for extensions, I need to read more to decide where to go. And of
course I'm open to suggestions. I expect to spend a lot of time over the
next weeks reading papers.  Induction would be cool.  Equality is
desirable for realistic problems. Figuring out a way to optimize for
Horn theories would be cool.
 
Once I'm sure I understand the issues and the design is good, I could
reimplement in ML or maybe Mercury.