A theory of abstract data types for program development: Bridging the gap?

This paper outlines a logical approach to abstract data types, which is motivated by, and more adequate for (than algebraic approaches), the practice of programming. Abstract data types are specified as axiomatic theories and notions concerning the former are captured by syntactical concepts concerning the latter. The basic concepts of nambility, conservative extensions and interpretations of theories explain implementation, refinement and parameterisation. Being simple, natural and flexible, this approach is quite appropriate for program development.