Chapter 3 - Model clone detection and its role in emergent model pattern mining: Towards using model clone detectors as emergent pattern miners – Potential and challenges

Abstract

Model-based software engineering approaches continue to gain traction in both industry and research. Accordingly the size, complexity, and prevalence of the models themselves are increasing. Model analysis and management thus becomes an essential task within any model-based process. One form of analysis that can support model-based approaches during the software engineering life cycle is pattern extraction, whereby tooling identifies emergent model patterns. These patterns can be used by analysts to ensure adherence to standards, software quality assurance, and library generation and optimization. In this chapter, we discuss the plausibility of using model clone detection as a form of emergent pattern mining for model-based systems. After a brief primer on the field of model clone detection and model pattern detection, we propose a conceptual framework, MCPM, centered on model clone detection that analysts can employ to detect emergent patterns in their models. In describing our framework concept, we illustrate our ideas using Simulink, and our Simulink model clone detector, Simone, as an example. However, we also consider other model clone detectors' potential within the MCPM framework. This includes how existing research and tooling can be applied to each step within the framework. We identify open challenges for researchers in realizing model clone detection as a model pattern mining tool, as well as the potential benefit that can be experienced by practitioners in the application of MCPM.