An Empirical Study on the Maturity of the Eclipse Modeling Ecosystem

Abstract

Since the early days of Model-driven Engineering (MDE), our community has been discussing the reasons why MDE had not quickly became mainstream. It is now clear the answer is a mix of technical and social factors, but among the former, the lack of maturity of MDE tools is often mentioned. The goal of this paper is to explore the question of whether this lack of maturity is actually true. We do so by comparing the maturity of over a hundred modeling and non-modeling projects living together in the Eclipse ecosystem. In both cases, we use the word project to refer to a variety of tools, libraries and other artefacts to build and manipulate software components, either at the model or code level. Our maturity model is based on code-centric and community metrics that we evaluate on the repository data for both kinds of projects. Their incubation status is also considered in the assessment. Results show that there are indeed differences between modeling and non-modeling projects, though less than we expected when setting up the study. Moreover, while the incubation status clearly separates non-modeling projects, the same is not true for modeling projects which seem to remain much more stable across their lifespan. We believe our results help to have a better perspective on maturity of modeling support nowadays and provide ideas for further analysis towards their improvement.