An empirical evaluation of the impact of heuristics on the object-oriented design

This dissertation , based on empirical research, examines and evaluates the contribution, as well as the impact of design heuristics on the quality of object - oriented designs, pursuing a more systematic approach to software development. It analyses and examines the application of empirical research through experimentation, suggesting more efficient application ways. It contributes to the achievement of new assessment standards, based on the theory of metrics as well as on empirical research, thus making efficient the anticipation of software maintainability and effort. The following studies have been carried out: a)An investigation examining the way experimentation had been carried out in the past, also evaluating compliance to heuristics. Improved experimentation ways have been suggested, and heuristics application deficiencies have been identified. b) An observational study investigating the impact of the design heuristic is evaluated on software quality, on performance, on the efficiency for anticipating problematic situations, as well as the efficiency for further improving the design tools. c) A controlled experiment investigating the impact of an Object - Oriented feature of inheritance. It is concluded that there is sufficient empirical evidence suggesting relationships between object - oriented design heuristics and software quality, particular with respect to maintainability. Also, it suggests the building of models for the detection of cases where the heuristics are violated. Finally, it suggests improvement of design tools