Delivering high-quality software is important for any high-tech company, like ASML. Assessing the quality of software through metrics on code-level is a well-known and well-studied area. Applying these principles to a Model-Driven Engineering context is not trivial. The ideas behind common complexity metrics, such as Cyclomatic Complexity, are still applicable to models; however, their computations are not, since the computations are expressed using code constructs and not model constructs. In this work we established a framework to define and implement metrics for control models. As a case study, we defined a set of metrics according this framework. Metrics are not usable if there is no tooling to compute them. We developed a tool that computes metrics on control models developed as state machines. The tool can be extended with new metrics and new state machine specifications. Additionally, the tool can be used in multiple development environments, such as ASD:Suite and ASOME. Furthermore, we present an approach to empirically validate the metrics.
|Award date||28 Sep 2018|
|Place of Publication||Eindhoven|
|Publication status||Published - 28 Sep 2017|