LP, the Larch Prover — Release 3.1b


The following bugs in Version 3.1a have been fixed in Version 3.1b.
95-014
The normalize command now properly normalizes a rewrite rule with a reducible left side. Formerly, it failed to normalize a rewrite rule when its left side could be reduced, but no further reductions were possible.
95-015
The operation of boolean equality is now recognized as an ac operator and not just as a commutative operator. Formerly, LP incorrectly (and unsoundly) eliminated some subterms when reducing formulas such as a = (r = (b = b')) that contain multiple occurrences of equality between boolean subformulas.
95-016
The experimental version of LP (invoked by the command lp -e) no longer crashes when a register command is issued during a proof.
98-001
The bound variable i1 in
f(i, j) <=> \A i1 \A j1 (i ~= i1 \/ j ~= j1 => P(i1, j1))
is no longer changed incorrectly to i2 when executing show normal-form f(i2, i1).
In addition, the following improvements have been made in Version 3.1b.