Event-Based Metric for Computing System Complexity

In order to increase quality and productivity of software, weaknesses must be identified in the software development process. Software metrics play a central role in identifying such weaknesses. High complexity of software is also one of the weaknesses. Identifying it in the early analysis stage can reduce cost, time and efforts. In this paper, a model-based approach is proposed to derive software metrics for measuring system complexity. The proposed metric measures the complexity of system on the basis of event flows and their inter-dependencies. Events taking place in the system are represented using Event Templates. An Event-Flow Model is then constructed from Event Templates and is represented as Event-Flow Graph. The proposed Event-Flow Complexity metric for analysis model is derived from Event-Flow Graph. The metric has also been evaluated in terms of Weyuker's properties. Results of evaluation show that it satisfies 8 out of 9 Weyuker's properties.