Detection of Generic Micro-architectures on Models

Cédric Bouhours (1) , Hervé Leblanc (1) , Christian Percebois (1) , Thierry Millan (1)

1 IRIT-MACAO - Modèles, Architectures, Composants, Agilité et prOcessus

Existing pattern detection methods generally use code information obtained during reengineering process. However, none of these methods exclusively works with design information. In this paper, we propose a novel pattern detection method based on structural properties of UML models. This technique allows the detection of any kind of generic micro-architecture, like design patterns or spoiled patterns. Since a generic architecture is context-free, the structure of the searched fragments depends on the use context. So, our technique uses a structural concordance paradigm to identify all possible instantiations of a generic micro-architecture. To increase the precision of the detection, authorized, prohibited, and optional relations can be directly precised into the micro-architecture model.

Domaines: Génie logiciel [cs.SE]

Cédric Bouhours, Hervé Leblanc, Christian Percebois, Thierry Millan. Detection of Generic Micro-architectures on Models. PATTERNS 2010, The Second International Conferences on Pervasive Patterns and Applications, Nov 2010, Lisbon, Portugal. ⟨hal-02042297⟩