Exploring possibilities to analyse microblogs for dependability information in variability-intensive open source software systems

Abstract

Many open source software systems are variability-intensive because they are frequently adapted to different customer needs or deployment environments. Variability in these systems not only occurs in functionality, but also in quality attributes (e.g., security, reliability, dependability). Furthermore, variability in functionality and variability in quality attributes affect each other. Recently, open source communities have adopted microblogging to document and share software engineering knowledge. Microblogging dissolves boundaries between developers and other stakeholders (e.g., end users). This facilitates the involvement of many different non-technical stakeholders in the software development process. In this paper, we investigate the feasibility and suitability of utilizing microblogs to identify and analyze dependability information about open source software systems. We use the example of the currently most popular microblogging infrastructure (Twitter) and a widely used open source content management system (Drupal). Our results indicate that even though open source communities frequently share information on Twitter, challenges exist for extracting and analyzing dependability information. Furthermore, a better understanding is needed to describe the relation between issues reported by user communities in microblogs, and issues and actions documented by developer communities in change logs and release information.