Many current software development projects are developed by teams that are distributed across the world. This has paved the way for an increasing number of tools that are specifically tailored toward helping developers collaborate in such distributed environments. Common examples include code repositories, documentation repositories, discussion forums, voice and/or video chats, and so on.

Unfortunately, collaborative debugging has been paid very little attention. When a developer needs a colleague to collaborate with him or her on a debugging session, he usually has to ask him or her to use a remote desktop tool and voice and/or video chat. Although this is a common practice, Estler et al. have found that this technique is not effective enough. In this paper, they present CDB, a system that helps developers debug their projects collaboratively. CDB is integrated into CloudStudio, a web-based collaborative development environment. Simply put, CDB provides a web-based interface that allows several developers to work on the same debugging session by setting breakpoints, stepping through the code, and inspecting the values of the variables.

CDB itself is an interesting technical contribution to the state of the art in debugging, but what shines from a scientific point of view is the empirical study in which the authors explore which debugging features are the most effective in a collaborative environment, the impact of control policies while debugging (for example, a single developer versus multiple developers), and how using CDB versus remote desktop debugging compares. Their study is supported by a sound statistical analysis of the results.

Overall, the paper is well written. The authors have managed to provide a good motivation for collaborative debugging, which is a largely unexplored research field. What I appreciate the most is that they have carried out a comprehensive experimental study to prove that their contribution is actually worthy.