[Context and motivation] The Model-Driven Engineering (MDE) paradigm remains an active niche in information systems engineering and research. Given their primary role, assuring the quality of models and modelling languages is crucial. In a previous literature review, we have confirmed that there exist different conceptions on model and modelling language quality. [Question/problem] Also, many approaches tackle a narrow perspective and practitioners often feel that such practices burden them with extra effort. A wider overview of the field is needed, taking into account the constraining and pressing conditions of industrial practice. [Principal ideas/results] This paper recalls the essence of, current trends and pitfalls of modelling language quality evaluation within MDE. Some of the current unsolved issues are related to a growing system complexity, evaluating the quality of a complementary set of languages, the need to bring model transformations in the picture, etc. [Contribution] Through this evaluation work, we define a research agenda that reveals the main challenges in modelling language quality evaluation in the context of MDE.