A Comparative Study of Model Transformation Approaches through a Systematic Procedural Framework and Goal Question Metrics Paradigm

摘要

Model Driven Engineering has become a key Software Engineering approach, which aims at improving the cost-effectiveness and reusability of software by capturing the essential semantics of systems in models. By means of model transformations, these models can be analysed, improved, and mapped to executable implementations in a variety of languages and platforms. A large number of different transformation languages and tools, ranging from graph theoretic to relational, hybrid and imperative exist across the research community. A key problem in the current state of Model Driven Engineering is the lack of guidelines and techniques for measuring or improving transformation quality.This thesis addresses this problem by defining a transformation quality framework based on the ISO/IEC 9126 international software quality standard. The framework is validated on different transformation languages using diverse case studies. The case studies highlight the problems with the specification and design of particular categories of model transformation, and provide challenging examples by which model transformation languages and approaches can be compared. The evaluation procedure provides clear guidelines for suitability of selected transformation approaches on specific transformation problem by identifying the advantage and disadvantage of each approach.