Ted,
Consider the expression comm(a,b).c - c.comm(a,b) + b . comm(a,c) - comm(a,c).b + comm(b,c).a - a . comm(b,c).
The Jacobi identity tells us that this expression vanishes, but I don't think your simplification rules crunch
it to zero:
(%i32) comm(a,b).c - c.comm(a,b) + b . comm(a,c) - comm(a,c).b + comm(b,c).a - a . comm(b,c)$
(%i33) expand(subst('comm = lambda([a,b],a.b - b.a),%));
(%o33) 0
--Barton