Binary Information Press
Key classes have many implications for resource allocating. However, identifying them is not an easy task, especially when working with an unfamiliar and complex software system. To the best of the authors' knowledge, few works has been done on this topic. This paper proposes to use the centrality metrics widely used in complex network theory to evaluate the importance of class, and further to identify the key classes in a specific software system. To fulfill this task, they first use a software networks model to describe the internal dependency structure of software. Then some centrality metrics widely used in complex network theory are introduced to quantify the importance of classes. The commons and differences among these centrality metrics are revealed.