Model transformation is a recognized cornerstone in software engineering. Any nontrivial software system involves some kind of model transformation (for example, a large and heterogeneous system or an automotive system). Insuring that a transformation is correctly performed is challenging. Spectrum-based fault localization (SBFL) is a commonly used technique that employs probability techniques to search for software bugs. The paper is about applying such a fault localization technique to model transformation.
The paper is easy to read thanks to its gentle introduction to model-driven engineering, model transformation, and SBFL. It makes a good case for applying classical SBFL techniques to model transformation rules, showing how it is both viable and beneficial. The paper is suitable for both experts in the field and casual readers who wish to know more.