Supporting collaborative software development over GitHub. (16th December 2016)

• Record Type: Journal Article
• Authors: Arora, Ritu; Goel, Sanjay; Mittal, Ravi Kant
• Is Part Of: Software, practice & experience. Volume 47:Number 10(2017)
• Page Start: 1393
• Page End: 1416
• Publication Date: 2016-12-16
• Subjects: collaborative software development -- computer‐supported collaborative work -- software configuration management systems -- workspace awareness; Computer software -- Periodicals; Computer programming -- Periodicals; Computer programs -- Periodicals; 005.3
• DOI: 10.1002/spe.2468
• Languages: English
• ISSNs: 0038-0644

Abstract

Summary: GitHub is a web‐based, distributed Software Configuration Management (SCM) system build over Git, which enables developers to host shared repositories over the Internet and access them from any location, at any time. It helps developers to effectively orchestrate their activities over shared codebases by capturing direct conflicts arising because of concurrent editing on the same shared artifact. However, SCM systems have limited support for capturing inconsistencies arising because of indirect conflicts which arise because of software dependency relationships that exist between related artifacts, and lead to the introduction of syntactic and semantic inconsistencies in codebases. In this paper, we propose a novel collaborative software development (CSD) tool named, Collaboration Over GitHub (COG), that provides real‐time information about arising direct and indirect conflicts among collaborative developers, working over GitHub, through a collection of workspace awareness widgets. These widgets provide people‐centric information about direct and indirect collaborators over GitHub. Resource‐centric information about current and conflicting activities of real‐time collaborators is captured and propagated to others, based on the dependency relationships between software artifacts being manipulated by them. COG uses dependency graphs to store and process dependency relationship information which is required to ascertain information about indirect conflicts. Notably, the most important novel contribution of COG is that it not only captures indirect conflicts that lead to the introduction of syntactic inconsistencies but also changes that lead to semantic inconsistencies in the codebase. It also does so at finer levels of granularity, with changes to individual method's body being traced, thereby capturing statement‐level conflicts as well. Copyright © 2016 John Wiley & Sons, Ltd.