Compilers for monomorphic languages, such as C and Pascal, take advantage of types to determine data representations, alignment, calling conventions, and register selection. However, these languages lack important features including polymorphism, abstract datatypes, and garbage collection. In contrast, modern programming languages such as Standard ML (SML), provide all of these features, but existing implementations fail to take full advantage of types. The result is that performance of SML code is quite bad when compared to C. In this thesis, I provide a general framework, called type-directed compilation, that allows compiler writers to take advantage of types at all stages in compilation. In the framework, types are used not only to determine efficient representations and calling conventions, but also to prove the correctness of the compiler. A key property of typedirected compilation is that all but the lowest levels of the compiler use typed intermediate languages. An advantage of this approach is that it provides a means for automatically checking the integrity of the resulting code. An important

The recent years have seen a number of proposals for extending statically typed languages by dynamics or generics. Most proposals --- if not all --- require significant extensions to the underlying language. In this paper we show that this need not be the case. We propose a particularly lightweight extension that supports both dynamics and generics. Furthermore, the two features are smoothly integrated: dynamic values, for instance, can be passed to generic functions. Our proposal makes do with a standard Hindley-Milner type system augmented by existential types. Building upon these ideas we have implemented a small library that is readily usable both with Hugs and with the Glasgow Haskell compiler.

In this paper, we explore an extension to Haskell type classes that allows a type class declaration to define data types as well as values (or methods). Similarly, an instance declaration gives a witness for such data types, as well as a witness for each method. It turns out that this extension directly supports the idea of a type-indexed type, and is useful in many applications, especially for self-optimising libraries that adapt their data representations and algorithms in a type-directed manner.

We introduce System FC, which extends System F with support for non-syntactic type equality. There are two main extensions: (i) explicit witnesses for type equalities, and (ii) open, non-parametric type functions, given meaning by toplevel equality axioms. Unlike System F, FC is expressive enough to serve as a target for several different source-language features, including Haskell’s newtype, generalised algebraic data types, associated types, functional dependencies, and perhaps more besides.

We present a new foreign-function interface for SML/NJ. It is based on the idea of data-level interoperability the ability of ML programs to inspect as well as manipulate C data structures directly. The core component of this work is an encoding of the almost complete C type system in ML types. [Variable-argument functions are the only feature of the C type system that we do not handle very well yet.] The encoding makes extensive use of a folklore typing trick, taking advantage of ML's polymorphism, its type constructors, its abstraction mechanisms, and even functors. A small low-level component which deals with C struct and union declarations as well as program linkage is hidden from the programmer's eye by a simple program-generator tool that translates C declarations to corresponding ML glue code.

Haskell overloading poses new challenges for compiler writers. Until recently there have been no implementations of it which have had acceptable performance; users have been adviced to avoid it by using explicit type signatures. This is unfortunate since it does not promote the reusability of software components that overloading really offers. In this paper we describe a number of ways to improve the speed of Haskell overloading. None of the techniques described here is particularly exciting or complicated, but taken together they may give an order of magnitude speedup for some programs using overloading. The techniques fall into two categories: speeding up overloading, and avoiding overloading altogether. For the second kind we borrow some techniques from partial evaluation. There does not seem to be a single implementation technique which is a panacea; but a number of different ones have to be put together to get decent performance. 1 Introduction Haskell, [Hud92], introduces a new ...

We present a general framework HM(X) for type systems with constraints. The framework stays in the tradition of the Hindley/Milner type system. Its type system instances are sound under a standard untyped compositional semantics. We can give a generic type inference algorithm for HM(X) so that, under sufficient conditions on X, type inference will always compute the principal type of a term. We discuss instances of the framework that deal with polymorphic records, equational theories and subtypes.

A Hindley-Milner type system such as ML's seems to prohibit type-indexed values, i.e., functions that map a family of types to a family of values. Such functions generally perform case analysis on the input types and return values of possibly different types. The goal of our work is to demonstrate how to program with type-indexed values within a Hindley-Milner type system. Our first approach is to interpret an input type as its corresponding value, recursively. This solution is type-safe, in the sense that the ML type system statically prevents any mismatch between the input type and function arguments that depend on this type. Such specific type interpretations, however, prevent us from combining different type-indexed values that share the same type. To meet this objection, we focus on finding a value-independent type encoding that can be shared by different functions. We propose and compare two solutions. One requires first-class and higher-order polymorphism, and, thus, is not implementable in the core language of ML, but it can be programmed using higher-order functors in Standard ML of New Jersey. Its usage, however, is clumsy. The other approach uses embedding/projection functions. It appears to be more practical. We demonstrate the usefulness of type-indexed values through examples including type-directed partial evaluation, C printf-like formatting, and subtype coercions. Finally, we discuss the tradeoffs between our approach and some other solutions based on more expressive typing disciplines.

9706572, and the US Air Force under grant F19628-95-C-0050 and a generous fellowship. The views and conclusions contained in this document are those of the author and should not be interpreted as representing the official policies, either expressed or implied, of any sponsoring institution, the U.S. government or any other entity.

with constraints. The basic idea is to factor out the common core of previous extensions of the Hindley/Milner system. I present a Hindley/Milner system where the constraint part is a parameter. Speci c applications can be obtained by providing speci c constraint systems which capture the application in mind. For instance, the Hindley/Milner system can be recovered by instantiating the constraint part to the standard Herbrand constraint system. Type system instances of the general framework are sound if the underlying constraint system is sound. Furthermore, I give a generic type inference algorithm for the general framework, under sucient conditions on the speci c constraint system type inference yields principal types.