For software developers, source code is the static blueprint that is transformed into running behavior. In modern applications, this behavior is often facilitated through distributed systems. Thus, behavior developers observe on their machine is usually not the same as it is perceived by end users. We propose a system where source code is mapped with information gathered at runtime to make programs fell “alive”. We implemented a system that maps runtime performance to source code artifacts. In this demo, we show how performance problems can be prevented in different scenarios.

PhD student at the University of Zurich, Research intern at IBM T.J. Watson Research Center interested in how we can make the software development process more data-driven by giving structure to runtime data and providing guidance to developers on how to use it