基于波及度的软件结构风险预测方法的研究 Research on Forecasting Structural Risks of Large-Scale Software Based on Ripple Degree

由于软件网络中各节点之间的相互调用与关联，波及效应普遍存在于软件网络之中。本文针对大量的开源软件的网络结构进行研究，首先分析了正向波及度和逆向波及度在软件网络中的分布规律，然后对波及度较大的节点进行重点研究，提出了波及度显著性度量公式，并根据度量结果发现软件结构中的脆弱节点、僵化节点和“桥梁”节点，为软件结构设计与重构提供指导。 Because of the association between nodes in software network, the ripple effect exists in software network. In this paper, the distributions of forward and reversal ripple degree in many networks of open source software are analyzed firstly, and then a metric formula that evaluates the significance is presented after focusing on those nodes that have high ripple degree. According to the metric results, we can select the vulnerable nodes, rigid nodes and a “bridge” node in software structure and this method can provide guidance for design and remodeling of software structure.