Training students in software engineering should attempt to mimic industry practices. Thus, student teams develop non-trivial software products, which includes interacting with collaborative tools deployed as web applications. The interaction may be mechanistic or organic, and occur for different durations. Collaboration studies tightly control these factors, relying on manual activity logging, very specific software requirements, surveys and interviews. Since these tools allow simultaneous interaction and capture revision histories, collaboration may be more objectively measured. This paper investigates social media conversations, revision histories, and commit logs from undergraduate student teams performing software development. The objective is to examine how this form of data could be translated into collaborative activities and whether the same performance relationships are achieved in a class setting. A small pilot study shows that the translation methodology did not produce the exact relationships from other studies, but it does shed light on a team’s perception of collaborators.
展开▼
