We present a new approach to inferring types in untyped object-oriented programs with inheritance, assignments, and late binding. It guarantees that all messages are understood, annotates the program with type information, allows polymorphic methods, and can be used as the basis of an op-timizing compiler. Types are finite sets of classes and subtyping is set inclusion. Using a trace graph, our algorithm constructs a set of conditional type constraints and computes the least solution by least fixed-point derivation.

This dissertation provides a framework for modularity in programming languages. In this framework, known as Jigsaw, inheritance is understood to be an essential linguistic mechanism for module manipulation. In Jigsaw, the roles of classes in existing languages are "unbundled," by providing a suite of operators independently controlling such effects as combination, modification, encapsulation, name resolution, and sharing, all on the single notion of module. All module operators are forms of inheritance. Thus, inheritance is not in conflict with modularity in this system, but is indeed its foundation. This allows a previously unobtainable spectrum of features to be combined in a cohesive manner, including multiple inheritance, mixins, encapsulation and strong typing. Jigsaw has a rigorous semantics, based upon a denotational model of inheritance. Jigsaw provides a notion of modularity independent of a particular computational paradigm. Jigsaw can therefore be applied to a wide variet...

This thesis develops a semantic model of inheritance and investigates its applications for the analysis and design of programming languages. Inheritance is a mechanism for incremental programming in the presence of self-reference. This interpretation of inheritance is formalized using traditional techniques of fixed-point theory, resulting in a compositional model of inheritance that is directly applicable to object-oriented languages. Novel applications of inheritance revealed by the model are illustrated to show that inheritance has wider significance beyond object-oriented class inheritance. Constraints induced by self-reference and inheritance are investigated using type theory and yield a formal characterization of abstract classes and a demonstration that the subtype relation is a direct consequence of the basic mechanism of inheritance. The model is proven equivalent to the operational semantics of inheritance embodied by the interpreters of object-oriented languages like Smalltalk. Concise descriptions of inheritance behavior in several object-oriented languages, including Smalltalk, Beta, Simula, and Flavors, are presented in a common framework that facilitates direct comparison of their features.

This paper presents a denotational model of inheritance. The model is based on an intuitive motivation of inheritance as a mechanism for deriving modified versions of recursive definitions. The correctness of the model is demonstrated by proving it equivalent to an operational semantics of inheritance based upon the method lookup algorithm of object-oriented languages.

In order to illuminate the fundamental concepts involved in object-oriented programming languages, we describe the design of TOOPL, a paradigmatic, statically-typed, functional, object-oriented programming language which supports classes, objects, methods, hidden instance variables, subtypes, and inheritance. It has proven to be quite difficult to design such a language which has a secure type system. A particular problem with statically type checking object-oriented languages is designing type-checking rules which ensure that methods provided in a superclass will continue to be type correct when inherited in a subclass. The type-checking rules for TOOPL have this feature, enabling library suppliers to provide only the interfaces of classes with actual executable code, while still allowing users to safely create subclasses. In order to achieve greater expressibility while retaining type-safety, we choose to separate the inheritance and subtyping hierarchy in the language. The design of...

We "unbundle" several roles of classes in existing languages, by providing a suite of operators independently controlling such effects as combination, modification, encapsulation, name resolution, and sharing, all on the single notion of module. All module operators are forms of inheritance. Thus, inheritance not only is not in conflict with modularity in our system, but is its foundation. This allows a previously unobtainable spectrum of features to be combined in a cohesive manner, including multiple inheritance, mixins, encapsulation and strong typing. We demonstrate our approach in a language (called Jigsaw, as in the tool, not the puzzle!). Our language is modular in two senses: it manipulates modules, and it is highly modular in its own conception, permitting various module combinators to be included, omitted, or newly constructed in various realizations.

While class-based object-oriented programming languages provide a flexible mechanism for re-using and managing related pieces of code, they typically lack linguistic facilities for specifying a uniform extension of many classes with one set of fields and methods. As a result, programmers are unable to express certain abstractions over classes. In this paper we develop a model of class-to-class functions that we refer to as mixins. A mixin function maps a class to an extended class by adding or overriding fields and methods. Programming with mixins is similar to programming with single inheritance classes, but mixins more directly encourage programming to interfaces. The paper develops these ideas within the context of Java. The results are 1. an intuitive model of an essential Java subset; 2. an extension that explains and models mixins; and 3. type soundness theorems for these languages. 1 Organizing Programs with Functions and Classes Object-oriented programming languages offer classes, inheritance, and overriding to parameterize over program pieces for management purposes and re-use. Functional programming languages provide various flavors of functional abstractions for the same purpose. The latter model was developed from a well-known, highly developed mathematical theory. The former grew in response to the need to manage large programs and to re-use as many components as possible. Each form of parameterization is useful for certain situations. With higherorder functions, a programmer can easily define many functions that share a similar core but differ in a few details. As many language designers and programmers readily acknowledge, however, the functional approach to parameterization is best used in situations with a relatively small number of parameters. When ...

We present the implementation of a type inference algorithm for untyped objectoriented programs with inheritance, assignments, and late binding. The algorithm significantly improves our previous one, presented at OOPSLA'91, since it can handle collection classes, such as List, in a useful way. Also, the complexity has been dramatically improved, from exponential time to low polynomial time. The implementation uses the techniques of incremental graph construction and constraint template instantiation to avoid representing intermediate results, doing superfluous work, and recomputing type information. Experiments indicate that the implementation type checks as much as 100 lines pr. second. This results in a mature product, on which a number of tools can be based, for example a safety tool, an image compression tool, a code optimization tool, and an annotation tool. This may make type inference for object-oriented languages practical.

We discuss denotational semantics of object-oriented languages, using the concept of closure widely used in (semi) functional programming to encapsulate side effects. It is shown that this denotational framework is adequate to explain classes, instantiation, and inheritance in the style of Sirnula as well as SMALLTALK-80. This framework is then com-pared with that of Kamin, in his recent denotational definition of SMALLTALK-80, and the implications of the differences between the two approaches are discussed. 1

We introduce simple object calculi that support method override and object subsumption. We give an untyped calculus, typing rules, and equational rules. We illustrate the expressiveness of our calculi and the pitfalls that we avoid. 1.