Compilers for higher-order programming languages like Scheme, ML, and Lisp can be broadly characterized as either "direct compilers" or "continuation-passing style (CPS) compilers", depending on their main intermediate representation. Our central result is a precise correspondence between the two compilation strategies. Starting from

The semantics of imperative languages are often expressed in terms of a store-passing translation and an algebra for reasoning about stores. We axiomatize the semantics of several typical imperative languages via equational axioms by "inverting" the storepassing translation as well as the algebraic axioms for reasoning about the store. The inversion process is simple and systematic and results in theories that are similar to equational theories for imperative languages that have been derived in more complicated ways, and is likely to be applicable to languages other than those presented here.

Program fragments in functional languages may be observationally congruent in a language without effects (continuations, state, exceptions) but not in an extension with effects. We give a generic way to preserve pure functional congruences by means of an effects delimiter. The effects delimiter is defined semantically using the retraction techniques of [14], but can also be given an operational semantics. We show that the effects delimiter restores observational congruences between purely functional pieces of code, thus achieving a modular separation between the purely functional language and its extensions. 1 Introduction Functional programming is a powerful paradigm, but it has long been recognized that purely functional programs are often inefficient and cumbersome. Many modern functional languages, e.g., SML [9], build in control and state features that strictly fall outside the functional paradigm. For example, SML of New Jersey includes a "call-with-current-continuation" opera...

Formal calculi of record structures have recently been a focus of active research. However, scarcely anyone has studied formally the dual notion---i.e., argument-passing to functions by keywords, and its harmonization with currying. We have. Recently, we introduced the label-selective -calculus, a conservative extension of -calculus that uses a labeling of abstractions and applications to perform unordered currying. In other words, it enables some form of commutation between arguments. This improves program legibility, thanks to the presence of labels, and efficiency, thanks to argument commuting. In this paper, we propose a simply typed version of the calculus, then extend it to one with ML-like polymorphic types. For the latter calculus, we establish the existence of principal types and we give an algorithm to compute them. Thanks to the fact that label-selective -calculus is a conservative extension of -calculus by adding numeric labels to stand for argument positions, its...

Imperative Lambda Calculus is a type system designed to combine functional and imperative programming features in an orthogonal fashion without compromising the algebraic properties of functions. It has been noted that the original system is too restrictive and lacks the subject reduction property. We define a revised type system that solves these problems using ideas from Reynolds's Syntactic Control of Interference. We also extend it to handle Hindley-Milner style polymorphism and devise type reconstruction algorithms. A sophisticated constraint language is designed to formulate principal types for terms. 1 Introduction The recent research in programming languages has greatly clarified the interaction between imperative and functional programming. The conventional notion that functional programming and imperative state-manipulation are in conflict has been dispelled. Several programming languages have now been designed which combine functions and assignments without destroying algeb...

We present atype reconstruction algorithm for SCIR [10], atype system for a language with syntactic control of interference. SCIR guarantees that terms of passive type do not cause any side e ects, and that distinct identi ers do not interfere. A reconstruction algorithm for this type system must deal with di erent kinds (passive and general) and di erent uses of identi ers (passive and active). In particular, there may not be a unique choice of kinds for type variables. Our work extends SCIR typings with kind constraints. We show that principal type schemes exist for this extended system and outline an algorithm for computing them. 1

Record calculi have recently been a very active field of research, but its reciprocal, i.e. the use of keywords in functions, is still ignored. Selective -calculus is a conservative extension of lambda calculus which, by labeling abstractions and applications, enables some form of commutation between arguments. It is an enhancement for both clarity, thanks to labels, and power with the possibility to commute. We propose here a simply typed version of this calculus, and show that it extends to second order and polymorphic typing. For this last one there exists a most generic type, and we give the algorithm to find it. This, combined with the fact selective -calculus extends naturally lambda calculus, giving numeric labels their intuitive meaning, provides us with a keyword extension for polymorphically typed languages like ML, compatible with the original syntax, where unlabeled abstractions and applications can be seen as labeled by 1. 1 Introduction The idea of introducing labels in ...

This paper introduces a succinct intermediate representation for compiling high-level, data-parallel languages. It is shown how inferences applied to this representation can infer the necessary information to compile it into efficient code. Further inference procedures are being investigated in order to automatically distribute and align data objects on distributed memory multi-processors. The inferences shown in this paper are just those which infer the shape, type, rank and `sort' of data parallel operations. Shape inference and type inference create expressions (in the same representation) which evaluate to the type and shape at run-time. Thus the intermediate representation is polymorphic in terms of shape (the shape of data-parallel operations and variables) and type. Inference on an access sub-typing structure will eventually provide the mapping information. Compilers from Eval to F-code and from F-code to C exist and a new front-end compiler from FORTRAN 90 is being developed, a...

Some algorithms make critical internal use of updatable state, even though their external specification is purely functional. Based on earlier work on monads, we present a way of securely encapsulating stateful computations that manipulate multiple, named, mutable objects, in the context of a non-strict, purely-functional language. The security of the encapsulation is assured by the type system, using parametricity. The same framework is also used to handle input/output operations (state changes on the external world) and calls to C.