Abstract
The distributed development model has been successfully adopted in such open source projects as the GNU/Linux operating system, Apache HTTP server and so on. However, the poor handling of the quality and customer support prohibits the progress of open source software (OSS). We focus on the problems of software quality. OSS systems are key components of critical infrastructures in society and continue to expand even now. Many OSS is developed in all parts of the world; for example, Firefox, Apache HTTP server, Linux and Android. In particular, a large-scale open source solution composed of several OSS is now attracting attention as a next-generation software development paradigm because of the cost reduction, quick delivery and reduced amount of work. In this paper, we propose a new approach to software quality assessment based on stochastic differential equations in order to consider the interesting aspect of the collision status in the binding phase of OSS. Notably, we derive several quality assessment measures considering the software service performance. In addition, we analyze actual software fault-count data to show numerical examples of software quality assessment considering the component collision for several OSS. Moreover, we show that the proposed method can assist in improving the quality of the large-scale open source solution.
