Liu, C., van Dongen, B. F., Assy, N., & van der Aalst, W. M. P. (2019). Detecting behavioral design patterns from software execution data. In E. Damiani, G. Spanoudakis, & L. A. Maciaszek (Eds.), Evaluation of Novel Approaches to Software Engineering – 13th International Conference, ENASE 2018, Revised Selected Papers (pp. 137-164). (Communications in Computer and Information Science; Vol. 1023). Cham: Springer. DOI: 10.1007/978-3-030-22559-9_7
Design pattern detection techniques provide useful insights to help understand the design and architecture of software systems. Existing design pattern detection techniques require as input the source code of software systems. Hence, these techniques may become not applicable in case the source code is not available anymore. Large volumes of data are recorded and stored during software execution, which is very useful for design pattern detection of software systems. This chapter introduces a general framework to support the detection of behavioral design patterns by taking as input the software execution data. To show the effectiveness, the proposed framework is instantiated for the observer, state and strategy patterns. The developed pattern detection techniques are implemented in the open-source process mining toolkit ProM. The applicability of the proposed framework is evaluated using software execution data containing around 1.000.000 method calls that are generated by running both synthetic and real-life software systems.