A Deductive Approach for Fault Localization in ATL Model Transformations - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2017

A Deductive Approach for Fault Localization in ATL Model Transformations

Résumé

In model-driven engineering, correct model transformation is essential for reliably producing the artifacts that drive software development. While the correctness of a model transformation can be specified and checked via contracts, debugging unverified contracts imposes a heavy cognitive load on transformation developers. To improve this situation, we present an automatic fault localization approach, based on natural deduction, for the ATL model transformation language. We start by designing sound natural deduction rules for the ATL language. Then, we propose an automated proof strategy that applies the designed deduction rules on the postconditions of the model transformation to generate sub-goals: successfully proving the sub-goals implies the satisfaction of the postconditions. When a sub-goal is not verified, we present the user with sliced ATL model transformation and predicates deduced from the postcondition as debugging clues. We provide an automated tool that implements this process. We evaluate its practical applicability using mutation analysis, and identify its limitations.
Fichier principal
Vignette du fichier
FASE.pdf (531.66 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01435977 , version 1 (16-01-2017)

Identifiants

Citer

Zheng Cheng, Massimo Tisi. A Deductive Approach for Fault Localization in ATL Model Transformations. FASE 2017 : 20th International Conference on Fundamental Approaches to Software Engineering, Apr 2017, Uppsala, Sweden. pp.300-317, ⟨10.1007/978-3-662-54494-5_17⟩. ⟨hal-01435977⟩
542 Consultations
383 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More