Testing M2T Transformations - A Systematic Literature Review

Context: Model-Driven Development (MDD) is about to become a reality in the development of enterprise information systems due to its benefits, such as reduction of development and maintenance costs, and support for controlled evolution. Consequently, testing model transformations, considering their high complexity particularly regarding Model-to-Text (M2T) transformations, plays a key role to increase the confidence in the produced artefacts. Objective: this paper aims to characterize testing approaches and test selection criteria that focus on M2T transformations, in particular white-box approaches. Method: the objective is accomplished through a systematic literature review. We defined research questions regarding the testing of M2T transformations and extracted and analyzed data from a set of primary studies. Results: we identified a variety of incipient white-box testing approaches for this context. They mostly rely on mapping strategies and traceability of artefacts. Most of them focus on well-formedness and correctness of models and source code, although we could notice a change of focus in most recent research. Conclusions: current solutions for testing M2T transformations have begun to change the initial focus on well-formedness and correctness of models. Some approaches involve techniques that establish coverage criteria for testing, whereas others try to solve the testability across many transformations languages.