| Following the rapid development of computer technique, the implementation of large-scale software system becomes possible, but the continuous enlargement of software scale results in the gradually improvement, so it make the all parts of the interactional software generate some new characteristics as part of the whole, and they just contain the software structure. The complexity of software structure has become the main factor affecting the software quality. So far, traditional software metrics method has supported many ways to evaluate and control software quality, but it has some shortage in the system structure metrics, so for the quantification of software structure complexity, there is still no effective and complete metrics system.For the above problems, as to measure the characteristics and patterns in the system structure, and then control the structure quality of modern large-scale, this paper combines new emerging complex network theory, and makes the relationship between different modules composing a specific network structure become an abstraction, and analyses ripple effect the for nodes in the software network, put forward a metrics method about ripple degree. This paper combines the local ripple degree and the topology parameters of structure characteristic in software network, and summarizes its pattern, and puts forward ripple effect based on software network and the metric system of structure quality metric in software system, and makes experiment studies for verifying this metric system.The experimental verification result shows that it is effective to measure the quality fault of system structure, and it can be used as an effective method to measure the quality of software structure. In addition, the metric research method also has important reality meaning and application prospect in terms of fault-tolerance ability, maintainability control, and reconfiguration and iteration development of software, test of software and so on. |
PDF Full Text Request