The Semantic Challenge of Object-Oriented Programming

Motivation

Object-oriented programming is based on an informal concept of object as an entity or thing whose identity persists over time. The object concept is immediately mean ingful to programmers, and has proven to be a useful and flexible organizational de vice in the analysis, design, and maintenance of complex systems. But though ob jects are attractively simple and intuitive in their initial conception, programming languages that support object-orientation are subtle and pose significant challenges for researchers.

Research on the Foundations of OOP has largely concentrated on operational semantics and type theories. On the other hand, research in denotational semantics has moved from the study of purely functional languages to include questions of local state and interaction, both of which are integral to the essence of the object concept. But these latter advances, while related to the concerns of object-oriented programming, have often taken place for languages that do not directly support objects. They do not evidently apply to OOP, and do not address some of the requirements that shape object-oriented languages.

The purpose of this seminar is to bring together researchers from the two camps. On one hand, OOP provides a great challenge for current semantic methods, and at tempting to apply them will likely bring up new problems and give new insight on the methods themselves. On the other hand, a deeper semantic analysis of object-oriented languages can potentially impact program specification, type systems, and static analysis. Further, the object concept itself carries an inherent interest, which demands a semantics that reflects and supports the programmer's informal conception.