The modern software development processes enable evolving software systems and refining models across software life cycle. However, these evolution attitudes may lead to some consistency problems among models at different levels of abstraction. Hence, it is required to discover and detect the potential inconsistencies occurring in models when developing a system. This paper focuses on checking the vertical consistency of UML models using an approach based on defining constraints at the meta-level. These constraints are expressed using EVL (Epsilon Validation Language) to ensure the consistency of models. Representative examples of constraints for checking vertical inconsistencies between class and sequence diagrams are proposed to illustrate our contribution.

Software systems are often modeled as a set of related UML diagrams. Due to the overlapping multi-view nature of UML and due to the incremental and iterative nature of the used software development process, these diagrams might contain inconsistencies. Thus, it is of utmost importance to detect, analyze and fix these inconsistencies before implementing the system. In this paper, we elaborate a transverse view of the aforementioned activities. More explicitly, this work proposes a Systematic Literature Review (SLR) that evaluates the existing techniques for managing inconsistencies using different research questions. The ultimate objective of this review is to come up with new recommendations that should be taken into consideration when conceiving a new proposal for checking inconsistencies.

Nowadays, every business is heavily depending on software. However, designing and developing software is a very serious engineering challenge since software systems are growing in size and complexity. This gives rise to many inconsistencies in software design. To deal with this issue, researchers have been working for many years on different model inconsistency management activities, namely the detection, the diagnosis and the handling of those inconsistencies. This work focuses on handling UML model inconsistencies by proposing an AHP-based method aiming to help modelers choosing the right repair action that fits well their modeling objectives and strategies. The proposed method helps structuring the selection of the most appropriate repair action among a set of alternatives. It also represents and quantifies the criteria elements of this decision problem, relates these elements to overall goals and evaluates and ranks alternative solutions. The efficiency of the method is evaluated using different inconsistency examples through different situations. The obtained results show that modelers can have credible propositions to make decisions in accordance with their initial objectives.

No abstract

No abstract