Using metrics for assessing the quality of ATL model transformations

Model transformations play a pivotal role in model-driven engineering. Since they are in many ways similar to traditional software artifacts, they have to be treated similarly. Therefore, it is necessary to assess their quality. We propose to use metrics to assess various quality attributes of model transformations In this paper, we focus on model transformations created using ATL, which ATL is currently one of the most widely used model transformation formalisms. We have collected metrics data from a heterogeneous collection of seven model transformations. The quality of the same transformations has been evaluated manually by nineteen ATL experts. We assess whether the metrics are appropriate predictors for the quality attributes by correlating the metrics data with the expert data. To support or refute the correlations, we also acquired qualitative statements from the ATL experts. Although the study was intended as a first exploration of the relation between metrics and quality attributes, some significant correlations were found that are supported by statements of the participants.