Losing Functions without Gaining Data – a new method for defunctionalisation

We describe an automated transformation which takes a higher-order program, and a produces an equivalent first-order program. Unlike Reynolds style defunctionalisation, it does not introduce any new data types, and the results are more amenable to subse-quent analysis operations. Our transformation is implemented, and works on a Core language to which Haskell programs can be re-duced. Our method cannot always succeed in removing all func-tional values, but in practice it is remarkably successful