Operation Based Model Representation: Experiences on Inconsistency Detection

Jerome Le Noir (1) , Olivier Delande (1) , Daniel Exertier (1) , Marcos Aurélio Almeida da Silva (2) , Xavier Blanc (3)

1 Thales Research and Technology [Palaiseau]
2 MoVe - Modélisation et Vérification
3 LaBRI - Laboratoire Bordelais de Recherche en Informatique

Keeping the consistency between design models is paramount in complex contexts. It turns out that the underlying Model Representation Strategy has an impact on the inconsistency detection activity. The Operation Based strategy represents models as the sequence of atomic editing actions that lead to its current state. Claims have been made about gains in time and space complexity and in versatility by using this kind of representation when compared to the traditional object based one. However, this hypothesis has never been tested in an industrial context before. In this paper, we detail our experience evaluating an Operation Based consistency engine (Praxis) when compared with a legacy system based on EMF. We evaluated a set of industrial models under inconsistency rules written in both Java (for EMF) and PraxisRules (the DSL - Domain Specific Language - for describing inconsistency rules in Praxis). Our results partially confirm the gains claimed by the Operation Based engines.

Domaines: Ingénierie assistée par ordinateur

Jerome Le Noir, Olivier Delande, Daniel Exertier, Marcos Aurélio Almeida da Silva, Xavier Blanc. Operation Based Model Representation: Experiences on Inconsistency Detection. ECMFA 2011 - 7th European Conference on Modelling Foundations and Applications, Jun 2011, Birmingham, United Kingdom. pp.85-96, ⟨10.1007/978-3-642-21470-7_7⟩. ⟨hal-00626824⟩