Object orientation is not always best!

Abstract

In recent years many new object oriented techniques have appeared and great things promised of them. We have been following these developments closely and have concluded that the best approach is not to cast out all the traditional functional methods and replace them completely with a set of new object oriented methods, but to select the most appropriate technique for each phase of a project. In this paper we argue that functional requirements analysis is far superior to object oriented requirements analysis, that object oriented design is very valuable and that object oriented programming is not appropriate to all systems.

We believe what we are discussing are ‘generic’ features of functional and object oriented methods rather than just limitations or benefits of particular methods and cannot be resolved simply by modifications to particular methods. Our conclusion is that the current rush to adopt any method with the phrase ‘object oriented’ in it should be examined very carefully and propose a mixed solution which we believe provides the best of both worlds.