AbstractCase tools dedicated to the design of databases have appeared on the market for some years. Those softwares only had a small penetration in the development of programs. This failure is partly due to the lack of support given in critical phases of conception, particularly in the maintenance and evolution of databases.
In this thesis, we think about the maintenance and evolution of relational databases design problems. This design is very often divided in three main parts: conceptual, logical and physical level. When a designer modifies the specifications in one of these three abstract levels, these modifications must be propagated in the other ones. How could we help the designer to propagate the modifications made in a program ' This is the question we try to answer in this thesis.
We work as follow. Through case studies, we examine closely the different types of modifications made to a database as well as their possible propagation at all levels of design. A better understanding of the situation is given by these analyses. After, we formalize strategies that are enable to solve the problems underscored during the examination of case studies. Further on, we set up methodological recommendations, which help the design of flexible databases.
The setting up of those strategies requires a maintenance environment. This environment must enable the representation of data structures at any abstraction level, to establish relations between the different level structures and to keep an image of the transformations made on the structure. In this area, a number of models and theorical concepts already exists. However, the environment is improved with a certain amount of elements for the evolution.
Finally, we build on the basis of developed strategies, assistant tools that support the analyst work carrying out maintenance activities. These implementations are build within the scope of the DB-MAIN project.
|Date of Award||2001|
|Supervisor||Jean-Luc Hainaut (Supervisor), Jean Fichefet (Jury), Naji Habra (Jury), Jacques Kouloumdjian (Jury) & Esteban Zimányi (Jury)|
- reverse engineering
- CASE tools
- schema transformation
Student thesis: Doc types › Doctor of Sciences