Object-oriented software engineering: problems and perspectives
Mohamed E. Fayad · 1994
The purpose of this thesis is to explore object-oriented software engineering problems and perspectives. This thesis presents a road map to evolve an ad-hoc development process into a managed, efficient object-oriented approach. This road map provides both technical and managerial perspectives. The contributions of this thesis are: (1) An effective transition plan to OO software engineering (OOSE) is presented. (2) Three case studies (real projects) and lessons learned are presented. These case studies are dealing with applying three different object-oriented techniques on real software developments projects and describe lessons learned from different object-oriented software developments. These techniques are: Colbert's object-oriented software development (OOSD) method, Shlaer-Mellor's object-oriented analysis (OOA), and Rumbaugh et al.'s object modeling technique (OMT). Solution(s) and strategies to some existing problems with OO techniques will be presented. These lessons are related to the following issues: technical; modeling and notations; standards and processes; and managerial. (3) The Object-Engineering Technique (O-ET) is introduced. O-ET is a new object-oriented approach to software development, encompassing engineering-oriented analysis (E-OA) and engineering-oriented design (E-OD). It is based on a set of notations, processes, and guidelines. O-ET provides extensions for the following properties: multiple views; tailorability; scalability; testability; maintainability; traceability; reusability support; documentation aid; packaging and repackaging support; and user's reconfiguration support. A complete case study will be provided with illustrated examples. (4) Comparison of four object-oriented techniques (e.g., Colbert's OOSD, Fayad's O-ET, Rumbaugh et al.'s OMT, and Shlaer-Mellor's OOA) based on technical and managerial perspectives is presented.