Model Transformation Testing and Debugging: A Survey

Abstract

Model transformations are the key technique in Model-Driven Engineering (MDE) to manipulate and construct models. As a consequence, the correctness of software systems built with MDE approaches relies mainly on the correctness of model transformations, and thus, detecting and locating bugs in model transformations have been popular research topics in recent years. This surge of work has led to a vast literature on model transformation testing and debugging, which makes it challenging to gain a comprehensive view of the current state of the art. This is an obstacle for newcomers to this topic and MDE practitioners to apply these approaches. This paper presents a survey on testing and debugging model transformations based on the analysis of 140 papers on the topics. We explore the trends, advances, and evolution over the years, bringing together previously disparate streams of work and providing a comprehensive view of these thriving areas. In addition, we present a conceptual framework to understand and categorise the different proposals. Finally, we identify several open research challenges and propose specific action points for the model transformation community.

Advantages and disadvantages of (dedicated) model transformation languages

“…As mentioned by our interviewees, for debuggers there is a need for model transformation specific techniques. Troya et al (2022) showed that there are numerous advances in this area like those by Wimmer et al (2009), Hibberd et al (2007), and Ege and Tichy (2019) but none of them have led to well rounded debuggers yet. Further effort by researchers active in this area is therefore required.…”

Revisiting Fault Localization Techniques for Model Transformations: Towards A Hybrid Approach.

“…Our work builds on two existing fault localization approaches for model transformations. As categorized in a recent survey on model transformation testing and debugging (Troya et al 2022), these are spectrum-based fault localization (Troya et al 2018) and footprint-based fault localization (Burgueño et al 2015). While the former is a dynamic approach, i.e., the model transformation needs to be executed, the latter is static, meaning that the model transformation is not executed, but the transformation definition is directly analyzed.…”

“…There are several works which focus on testing and debugging of model transformations. A detailed survey on the topic has been recently published in (Troya et al 2022). With respect to the contribution of this paper, we focus on approaches that propose to locate bugs in model transformations in the following discussion on related work.…”

“…As model transformations are also the core mechanism to translate models to implementations realized in codepotentially applying several transformation steps, the correctness of software systems realized with MDE approaches highly depends on the correctness of the utilized model transformations. Therefore, it is critical to test and debug model transformations in the context of MDE (Troya et al 2022) as it is the case with compilers for programming languages (Chen et al 2020). However, several challenges are associated with testing and debugging model transformations as they are specific and complex software systems (Baudry et al 2006).…”

Ant-colony optimization for automating test model generation in model transformation testing

“…The correctness of software built using MDE techniques greatly relies on the correctness of model transformations, so it is of prime importance to test them. The core of model transformation testing is typically divided in three phases [72,85], namely test data generation, test suite assessment and oracle function. The first phase is the main focus of this research.…”