| In software engineering,requirement change process management has been one of the most important work for software project managers.Requirements change occur frequently in open source community software projects,for many a requirement change reports are committed to issue tracking system every day.Because of the limited time and energy,developers of open source communities’ software projects have to decide which issue to be solve first.Understanding the priorities of requirement changes helps developers identify significant issues results in a better process of requirement change management,leading to a more dependable evolution of the project.In this context,this paper did the following work: First,a novel complex-network-based approach is proposed to analyze requirement changes within issue tracking system.Issues and issue links between issues,within issue tracking systems,are mapped into Associated Network of Requirement Change（ANRC）.The nodes of the network are issues,while the edges are the issue links between issues,with corresponding weights set according to different dependency types and the definition of indegree and outdegree.After the Associated Network of Requirement Change is built,the importance of nodes in the network is later analyzed and sorted by selected centrality measures according to the characteristics of the network.Second,temporal dimension is also considered to remove the advantages of older nodes in the network result in a new rank of important nodes at the same period.The results show that developers are not yet able to evaluate issues from a macro,network-based perspective,lacking the awareness of the significant issues.The existence of small-world phenomenon in the network reflects a good information communication and close cooperation between nodes.This paper verified the small-world phenomenon of associated network of requirement change and found it significant.For a better requirement change management of the huge Associated Network of Requirement Change,this paper used community detection method to divide Associated Network of Requirement Change,and then analyzed the main theme of the community with nature language processing method,all leading to a better understanding of issues and the whole project. |
