Leveraging Multi-Level Modeling for Multi-Domain Quality Assessment

Quality Evaluation Systems (QESs) are a class of software systems which receive evaluation requests and quality requirement specifications as inputs and produce results as outputs of an assessment process. This class of systems usually work with a quality model including quality definitions and metrics, and produce the output as a quantitative evaluation of a subject. QESs can be implemented using model-driven techniques and dedicated languages, for domain-specific evaluation of different subjects. However, applying traditional two-level metamodeling techniques for this scenario entails that every time a QES is required, a new modeling framework, with consequent engine for interpreting the newly defined models, must be re-developed from scratch. To overcome this repetitive process, in this paper, we propose a Multi-Level Modeling (MLM) approach for realizing the artifacts involved in the development phase of a QES which are reusable across multiple domains. We demonstrate the approach with running examples from three different application domains comprising different evaluation scenarios.