The issue tracking system (ITS) is a rich data source for data-driven decision making. Different characteristics of bugs, such as severity, priority, and time to fix may be misleading. Similarly, these values may be subjective, e.g., severity and priority values are assigned based on the intuition of a user or a developer rather than a structured and well-defined procedure. Hence, we explore the dependency graph of the bugs and its complexity as an alternative to show the actual project evolution. In this work, we aim to overcome uncertainty in decision making by tracking the complexity of the bug dependency graph (BDG) to come up with a bug resolution policy that balances different considerations such as bug dependency, severity, and fixing time for the bug triaging. We model the evolution of BDG by mining issue tracking systems of three open-source projects for the past ten years. We first design a Wayback machine to examine the current bug fixing strategies, and then we define eight rule-based bug prioritization policies and compare their performances using ten distinct internal and external indices. We simulate the behavior of the ITS and trace back the effect of each policy across the history of the ITS. Considering the strategies related to the topology of the BDG, we are able to address bug prioritization problems under different scenarios. Our findings show that the network-related approaches are superior to the actual prioritization task in most cases. Among the selected open-source projects, LibreOffice triagers are the only ones who disregard the importance of the BDG, and that project is faced with a very dense BDG. Although we found that there is no single remedy that satisfies all the expectations of developers, the graph-related policies are found to be robust and deemed to be more suitable for bug triaging.
Download