Software reuse in an object oriented framework: Distinguishing types from implementations and objects from attributes

Almost no object oriented programming languages offer distinct language constructs for the definition of types and their implementations; instead these are united into a single class concept. Similarly object oriented programming languages do not normally distinguish between object types, which may be independently instantiated, and attribute types, which may not. The paper shows how these distinctions can be used to develop both a specialized and a generalized bracket technique, and how the ideas lead to interesting possibilities for reusing code in a flexible and modular way.