Software metrics for complex networks

The complexity of software structures can be characterized by software structure metrics based on complex network theory. The parameters indicate the complexity degree and vertex strength. This study uses a classic object-oriented software system for statistical analyses of two parameters using numerical fitting to analyze correlations between the two parameters. The results show that the distributions of the complexity degree and vertex strength decay exponentially, especially the distributions of the out-degree and the out-vertex strengths, which shows that software experience stochastic growth in the software network. The functional relationship between the degree and vertex strength is also given. A software network growth mechanism is given based on the stochastic growth to determine the growth degree distribution and when the growth is exponential.