From Developer Networks to Verified Communities: A Fine-Grained Approach

Abstract

Abstract-Effective software engineering demands a coordinated effort. Unfortunately, a comprehensive view on developer coordination is rarely available to support software-engineering decisions, despite the significant implications on software quality, software architecture, and developer productivity. We present a fine-grained, verifiable, and fully automated approach to capture a view on developer coordination, based on commit information and source-code structure, mined from version-control systems. We apply methodology from network analysis and machine learning to identify developer communities automatically. Compared to previous work, our approach is fine-grained, and identifies statistically significant communities using order-statistics and a community-verification technique based on graph conductance. To demonstrate the scalability and generality of our approach, we analyze ten open-source projects with complex and active histories, written in various programming languages. By surveying 53 open-source developers from the ten projects, we validate the authenticity of inferred community structure with respect to reality. Our results indicate that developers of open-source projects form statistically significant community structures and this particular view on collaboration largely coincides with developers' perceptions of real-world collaboration.

On Using Network Science in Mining Developers Collaboration in Software Engineering: A Systematic Literature Review

“…However, for communication and social networks where an edge is a sent email, and for the collaboration networks where the edge is a repository fork or a comment added to the code, the diameter, for example, is just meaningless and does not provide a value to the analysis. In addition, only 9 PSs, [45,44,11,60,58,38,30,16,8], contributed a new model or a new measure based on the specifics of the used networks, the other 26 PSs only used an existing measure in the literature. Also, 11 primary studies did not provide an explanation and interpretation over the used measure.…”

“…The rest of the primary studies considered a static network. The PSs that considered directed networks were 13 PSs, [55,17,10,60,57,48,38,30,9,8,5,53,50] and the rest PSs considered an undirected network. We found that 60% of the PSs, 21 PSs, used a proxy edge.…”

Developer Experience Considering Work Difficulty in Software Development*

“…Some studies have suggested methods to determine experts in a particular domain [14], their roles [6], and developer relationships [16]. These studies are helpful to understand the main contributor in a specified domain and assign jobs.…”

On the Relationship Between Feature Dependencies and Change Propagation

“…simultaneous change in distinct program elements) within in a specific commit. Similar to several work [CATALDO et al 2008, BIRD et al 2011, JERMAKOVICS et al 2011, JOBLIN et al 2015, we assume that co-changes of distinct program elements in a single commit suggests the execution of a complete maintenance task. In other words, within a commit it is more likely to find program elements that cooperate to perform some high-level function in part of one or more feature dependencies.…”

“…Specifically, we analyse both feature dependencies and change propagation. Similar to several work [HASSAN and HOLT 2004, CATALDO et al 2008, BIRD et al 2011, JERMAKOVICS et al 2011, GEIPEL and SCHWEITZER 2012, JOBLIN et al 2015, in this study we use simultaneous changes (see Section 2.3) as indicators of change propagation. Simultaneous The rationale is that changes made simultaneously in different features during a single evolution means that those features are related somehow.…”