Speaker: Kim Herzig , Microsoft, USA
Area: Mining Version Archives
When and where: Tuesday, July 19, 14:05 – 15:35 at DFKI
Abstract: Mining software archives made its way to industry. At Microsoft, we collect and link nearly every artifact produced by engineers to guide decision making and to provide insights into root causes of issues. However, mining one subset of software repository artifacts leaves users and data analysts with an incomplete picture of the development process. In order to ensure to solve a problems root cause, it is important to capture the entire development process from coding over testing and code reviews to build, deployment, and telemetry data collected from customers. This can easily become a challenge, not only due to the size of the collected data but also to manage the heuristics or triggers that links different artifacts across different version archives. Making mistakes or being sloppy can have severe consequences, not only for individuals but also for the entire company.
In this talk, I want to share some lessons learned and approaches that work at scale for a company like Microsoft.
More on Kim:
Kim Herzig is as Software Development Engineer and Researcher for the Tools for Software Engineers team at Microsoft Redmond. He is closely collaborating with the Empirical Software Engineering (ESE) group at Microsoft Research. His work is focused on optimizing development and testing processes. Currently, Kim is analyzing build, test, and verification processes of large Microsoft product groups, such as Windows, OneDrive and Bing, aiming to improve and optimize the effectiveness and reliability of these processes.