THE CHALLENGE OF INTRODUCING THE OBJECT-ORIENTED PARADIGM - AN EMPIRICAL-INVESTIGATION OF A SOFTWARE-ENGINEERING COURSE

This paper deals with teaching experiences in introducing the object-oriented (OO) paradigm in a software engineering (SE) course. The results are of interest not only for educational environments but also for practitioners who have decided to change from procedural to OO development. The course at hand lasts one academic year. It consists of a three-hour lecture per week as well as practical exercises in which students develop documents and programs in teams of five to seven persons. We experienced that students take OO development as just another coding paradigm. They regard it as business as usual with a slightly different procedural programming language. In comparison to previous non-OO approaches, we realized that the basic structure of most of the programs did not really change. Though encapsulation was obviously enhanced (or, better say, enforced). The structure of the projects did not come from modeling of the problem domain. Rather it came from copying ideas and examples taken from the database and interface OO tools (tool-orientation) creating objects according to a conventional modular design (functional orientation). The main consequence of our experiences is to enforce a much stronger focus on OO analysis and OO design. Without these precautions, software engineers tend to overdo the bottom-up approach, especially if they are inexperienced in object-oriented programming (OOP). This may sound like an obvious consequence; nevertheless, our experiences in industrial consulting show that this basic insight is often ignored when OOP is introduced in an industrial SE setting.