Une approche pour la maintenance et la ré-ingénierie globale des logiciels
Jean-Marie Favré · HAL (Le Centre pour la Communication Scientifique Directe) · 1995
While computing turns to the future, this thesis focuses on the past; legacy software systems are and will continue to be maintained. Three themes are essential in the context of this study: Software Maintenance, Software Reengineering and Programming-In-The-Large (PITL). The main goal of this thesis is to explore the intersection between Reengineering and Programming-In-The-Large, a domain that we coin "Reengineering-In-The-Large". The key idea is to prove that Programming-in-the-small concepts can be used to solve practical Programming-In-The-Large problems. We studied PTIL according to 4 different views: architecture, manufacture, variation and evolution. We also defined an abstract model based on set theory concepts to formalise PITL entities. The PITL technology is studied. The differences between the State-Of-The-Art and the State-Of-The-Practice emphasizes the need for Reengineering-In-The-Large. Preprocessors are studied as special cases of PITL tools. These low-level tools are traditionally used to describe program families. In order to make their comprehension easier, we use techniques like partial evaluation, inter-procedural data-flow analysis and slicing. A prototype which combines these different techniques was implemented to demonstrate that an approach based on theoretical Programming-In-The- Small concepts can be used to solve practical Programming-In-The-Large problems.