Predicting build failures using social network analysis on developer communication

Abstract

A critical factor in work group coordination, communication has been studied extensively. Yet, we are missing objective evidence of the relationship between successful coordination outcome and communication structures. Using data from IBM's Jazz TM project, we study communication structures of development teams with high coordination needs. We conceptualize coordination outcome by the result of their code integration build processes (successful or failed) and study team communication structures with social network measures.Our results indicate that developer communication plays an important role in the quality of software integrations. Although we found that no individual measure could indicate whether a build will fail or succeed, we leveraged the combination of communication structure measures into a predictive model that indicates whether an integration will fail. When used for five project teams, our predictive model yielded recall values between 55% and 75%, and precision values between 50% to 76%.

Stakeholders' social interaction in requirements engineering of open source software

“…In order to create a visual represen-tation for stakeholders' socio-technical relationship, Sarma et al [169] considered both emails among developers and comments in Bugzilla. Posting and reading comments by contributors were also considered by Wolf and colleagues as a means to represent communication flow [194]. Bird et al [20] considered email as a communication medium among stakeholders and studied latent social structure in open source projects based on email communication.…”

“…For example, comments or activities in issue tracking systems, email, and IRC are widely considered a means of social interaction among different stakeholders [20,36,169,194]. In order to create a visual represen-tation for stakeholders' socio-technical relationship, Sarma et al [169] considered both emails among developers and comments in Bugzilla.…”

“…We need to consider the edge weights to calculate the metrics that will help us identify the structural holes. In order to define the edges, we adopt the approach presented by Wolf et al [194]. Let X and Y be two stakeholders and T be a task that both X and Y contribute to.…”

Impact of Stakeholder Type and Collaboration on Issue Resolution Time in OSS Projects

“…Bird et al showed that the sociotechnical network of software module and developer are able to predict if an entity is failure prone with greater accuracy than other methods [7]. Wolf et al formed a developer-task network to explore the impact of developer communication on predicting a build integration fail [32]. Pinzger et al constructed a developer-module network to predict the software failures [30].…”

“…There are three centrality measures in SNA, namely Freeman degree centrality, closeness and betweeness. In the scope of this study, we investigate Freeman centrality degree since this metric is successfully applied in relevant studies as in [14][30] [32]. In our network, the Freeman degree centrality of an issue represents the number of unique stakeholders that are involved in the issue.…”