Have proof plans been used to teach students mathematics?
Not to my knowledge. This sounds like a promising project. Some of our methods do correspond to the kinds of proof techniques used by experts in the domain. Often students are expected to pick these techniques up by generalising examples. One difficulty with this is that students may not know the meta-level concepts from which the method preconditions are formed. This will prevent them constructing a sufficiently general method. Explicit teaching, especially explanations of the meta-level concepts, may help. • How can humans discover proof plans? This is an art similar to the skill used by a good mathematics teacher when analysing a student’s proof or explaining a new method of proof to a class. The key is identifying the appropriate meta-level concepts to generalise from particular examples. Armed with the right concepts, standard inductive learning techniques can form the right generalisation (see the answer about learning new proof plans for more information). Drawbacks and Limitati
