Martin Hofmann’s case for non-strictly positive data types

We describe the breadth-first traversal algorithm by Martin Hofmann that usesa non-strictly positive data type and carry out a simple verification in anextensional setting. Termination is shown by implementing the algorithm inthe strongly normalising extension of system F by Mendler-style recursion.We then analyze the same algorithm by alternative verifications in anintensional setting, in a setting with non-strictly positive inductivedefinitions (not just non-strictly positive data types), and one by algebraicreduction. The verification approaches are compared in terms of notions ofsimulation and should elucidate the somewhat mysterious algorithm and thusmake a case for other uses of non-strictly positive data types. Except forthe termination proof, which cannot be formalised in Coq, all proofs wereformalised in Coq and some of the algorithms were implemented in Agda andHaskell