基于复杂网络的大型软件系统度量  

Metrics for Large-scale Software Systems Based on Complex Networks

作　　者：王树森[1] 顾庆[1] 陈焘[1] 陈道蓄[1] 

机构地区：[1]南京大学计算机科学与技术系软件新技术国家重点实验室,南京210093

出　　处：《计算机科学》2009年第2期287-290,302,共5页Computer Science

基　　金：国家"863"项目(编号:2006AA01Z177);江苏省自然科学基金基础研究项目(编号:BK2006115)资助

摘　　要：随着软件系统规模的不断增大,如何度量面向对象大规模软件系统成为一个亟待解决的问题。近年来迅速发展的复杂网络理论为解决该问题提供了一个新的视角。介绍了大规模软件系统中的复杂网络现象,从元素级、模块级、网络级3个不同的粒度提出基于复杂网络的软件系统的各种度量;实现了大型Java程序复杂网络描述和度量工具JPAC。JPAC可用于分析大型Java系统的结构,并计算基于复杂网络的各度量值。With the aggrandizement scale of software systems, how to measure the complexity of large-scale object- oriented software systems has been a challenge for software engineers. The theory of complex networks recently studied has offered a new perspective on solving this problem. We studied the phenomenon of complex networks in large-scale software system, and proposed metrics from three different levels, i. e. , unit-level, module-level and network-level. We built JPAC, a tool for characterizing and measuring the complexity of Java systems. JPAC can be used to analyze the structure of large-scale Java systems, and compute software metrics based on complex networks.

关 键 词：复杂网络 软件系统 软件度量 JPAC 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论] Q2-0[自动化与计算机技术—计算机科学与技术]