The type and effect discipline is a new framework for reconstructing the principal type and the minimal effect of expressions in implicitly typed polymorphic functional languages that support imperative constructs. The type and effect discipline outperforms other polymorphic type systems. Just as types abstract collections of concrete values, effects denote imperative operations on regions. Regions abstract sets of possibly aliased memory locations. Effects are used to control type generalization in the presence of imperative constructs while regions delimit observable side-effects. The observable effects of an expression range over the regions that are free in its type environment and its type; effects related to local data structures can be discarded during type reconstruction. The type of an expression can be generalized with respect to the variables that are not free in the type environment or in the observable effect. 1 Introduction Type inference [12] is the process that automa...

We present the first algorithm for reconstructing the types and effects of expressions in the presence of first class procedures in a polymorphic typed language. Effects are static descriptions of the dynamic behavior of expressions. Just as a type describes what an expression computes, an effect describes how an expression computes. Types are more complicated to reconstruct in the presence of effects because the algebra of effects induces complex constraints on both effects and types. In this paper we show how to perform reconstruction in the presence of such constraints with a new algorithm called algebraic reconstruction, prove that it is sound and complete, and discuss its practical import. This research was supported by DARPA under ONR Contract N00014-89-J-1988. 1

this paper is to marry effects to monads, writing T for a computation that yields a value in and may have effects delimited by oe. Now we have that ( is

. This paper presents a new dialect of Scheme aimed towards concurrency and distribution. It offers a few primitives, including first-class continuations, with very simple semantics. Numerous examples are given showing how to program the classical concurrent control operators such as future, pcall and either. The implementation is sketched and presented along the lines of a metacircular interpreter. This paper presents the idiom of Icsla 1 , a language belonging to the Lisp family and more precisely a descendant of Scheme. This dialect has been designed with respect to the following main objectives: -- It should have a very simple and understandable semantics, with few but powerful and unrestrictively combinable concepts; -- It should offer concurrency, distribution and some other modern features such as sophisticated control features while not sacrificing the variety of styles traditionally offered by Lisp. These goals are rather general and deserve further comment. Following Sche...

It is well known that adding side effects to functional languages changes the operational equivalences of the language. We develop a new language construct, encap, that forces imperative pieces of code to behave purely functionally, i.e., without any visible side effects. The coercion operator encap provides a means of extending the simple reasoning principles for equivalences of code in a functional language to a language with side effects. In earlier work [36], similar coercion operators were developed, but their correctness required the underlying functional language to include parallel operations. The coercion operators developed here are simpler and are proven correct for purely sequential languages. The sequential setting requires the construction of fully abstract models for sequential call-by-value languages and the formulation of a weak form of "monad" suitable for expressing the semantics of call-by-value languages with side effects. 1 Introduction Two pieces of code are...

General-purpose programming languages are a Good Thing, but domain-specific languages can be better. Such languages provide a natural vocabulary for concepts that are fundamental to the problem domain, whilst with general-purpose languages one is reduced to idiom: idiom, moreover, that is of its very nature alien to the host language --- not to mention uncheckable. In addition, domain knowledge often inspires a range of optimizations that would not otherwise suggest themselves, let alone be practical. Domain-specific languages necessarily differ from general-purpose languages in their purpose, and therefore design, but we claim that the underlying principles can and should be the same. There need be no contradiction here: one is merely enriching the conceptual basis upon which the language is grown. The principled introduction of new concepts may have profound --- and sometimes unanticipated! --- consequences, however, so language designers (and implementors) must be prepared to follow...

First-class continuations are a powerful computational effect, allowing the programmer to express any form of jumping. Types and effect systems can be used to reason about continuations, both in the source language and in the target language of the continuation-passing transform. In this paper, we establish the connection between an effect system for first-class continuations and typed versions of continuationpassing style. A region in the effect system determines a local answer type for continuations, such that the continuation transforms of pure expressions are parametrically polymorphic in their answer types. We use this polymorphism to derive transforms that make use of effect information, in particular, a mixed linear/non-linear continuation-passing transform, in which expressions without control effects are passed their continuations linearly.

Numerous high-level control operators, with various properties, exist in the literature. To understand or compare them is difficult since their definitions use quite different theoretical frameworks; moreover, to our knowledge, no implementation offers them all. This paper tries to explain control operators by the often simple stack manipulation they perform. We therefore present what we think these operators are, in an executable framework derived from abstract continuations. This library is published in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. For instance, we do not claim our implementation to be faithful nor we attempt to formally derive these implementations from their original definitions. The goal is to give a flavor of what control operators are, from an implementation point of view. Last but worth to say, all errors are mine. Among the many existing control operators, w...

. Safety of interoperation of program fragments written in different safe languages may fail when the languages have different systems of computational effects: an exception raised by an ML function may have no valid semantic interpretation in the context of a Safe-C caller. Sandboxing costs performance and still may violate the semantics if effects are not taken into account. We show that effect annotations alone are insufficient to guarantee safety, and we present a type system with bounded effect polymorphism designed to verify the compatibility of abstract resources required by the computational models of the interoperating languages. The type system ensures single address space interoperability of statically typed languages with effect mechanisms built of modules for control and state. It is shown sound for safety with respect to the semantics of a language with constructs for selection, simulation, and blocking of resources, targeted as an intermediate language for optimization o...

Abstract. This paper studies continuations by means of a polymorphic type system. The traditional call-by-name continuation passing style transform admits a typing in which some answer types are polymorphic, even in the presence of first-class control operators. By building on this polymorphic typing, and using parametricity reasoning, we show that the call-by-name transform satisfies the eta-law, and is in fact isomorphic to the more recent CPS transform defined by Streicher. 1