Technical debt in procedural model transformation languages

Model transformations (MT) are a significant element of model-based and model-driven engineering approaches. Hence their quality is important for their effectiveness for use in such developments. The concept of technical debt (TD) estimates the impact of quality flaws on the maintenance costs of software artefacts. In this paper we define a quality model for model transformation languages, and define TD indicators for MT based on this model. We apply the indicators to evaluate the TD levels in 40 transformation cases in three procedural MT languages: QVT-O, Kermeta and EGL, and compare these levels with results for the declarative QVT-R language.