Incremental Needed Narrowing


M. Alpuente, S. Escobar, S. Lucas


Needed narrowing is currently the best complete strategy for executing inductively sequential functional logic programs. Its optimality properties and the fact that inductively sequential programs are a subclass of strongly sequential programs support the claim that needed narrowing must be considered the functional logic couterpart of Huet and Lévy's strongly needed reduction. In this paper, we show how a pre-eminent property of reduction in (a distinguished subclass of) strongly sequential programs, namely the incrementality of the evaluation, can be inherited by needed narrowing. We give an incremental definition of needed narrowing and show that the original optimality properties are kept. Moreover, we experimentally demonstrate that the incremental refinement can lead to substantial improvements in the overall evaluation process.


Functional logic languages, Implementation, Incrementality, Needed Narrowing.