Using Statistical Analysis of FLOSS Systems Complexity to Understand Software Inactivity
Authors
Abstract
Understanding how systems evolves can reveal important pieces of information that can help open source stakeholders to identify what can be improved in the software system’s internal organization. Once software complexity is one of the most important attributes to determine software maintainability, controlling its level in the system evolution process makes the software easier to maintain, reducing the maintainability costs. Otherwise, uncontrolled complexity makes the maintenance and enhancement process lengthy, more costly and sometimes it can contribute to the system abandonment. This work investigates the evolution of complexity in discontinued FLOSS projects, through statistical analysis with data obtained from analisis of SonarQube Software. SonarQube is an open-source software quality tool that analyzes the project’s source code and give the developers a feedback about the internal status of what is being developed. After several analyses, the outcome showed interesting results. A substantial portion of inactive FLOSS projects do not seem to be able to keep up with the extra work required to control the systems complexity, presenting a different behaviour of thesuccessful active FLOSS projects. Though, some inactive FLOSS projects do have a complexity evolution that resembles with the curves belonging to active projects.