Distributed-pair programming can work well and is not just distributed pair-programming

Author(s): J. Schenk, L. Prechelt, S. Salinger
Venue: International Conference on Software Engineering
Date: 2014

Type of Experiement: Case Study
Sample Size: 2
Class/Experience Level: Professional
Participant Selection: Personally selected by the authors
Data Collection Method: Observation


In this article, Schenk, Prechelt and Salinger explore the effects of distributed-pair programming and conclude that awareness deficits and concurrent code editing, which are both absent in traditional pair programming, do not affect skilled programmers who are familiar with each other and have focused, successful communication. According to their study, distance collaborative programming are traditionally done through screen-sharing (Remote Pair Programming or RPP) or through a distributed IDE (Distribute Pair Programming or DPP and eDPP). They observed the latter by conducting a case study on two seasoned German software developers who both used Saros, an Eclipse plugin for eDPP, specifically focusing on awareness and editing freedom. From the case study, they find that experienced programmers who are not co-located “have much lower needs for physical awareness than previously assumed” and can conduct eDPP with hardly any trouble. In addition, rather than the initial hypothesis that concurrent code editing might be cumbersome for developers, using Saros actually improved the two developers’ work process.