Analyzing programming behavior to support self-reflection for improving source code quality

Improving the quality of source code is an important aspect of reducing software development cost in industry. This position paper postulates that developers in a software team conduct different activities during software development like closing the source code editor for compilation as opposed to leaving it open that either improve or degrade the quality of resulting source code. We propose to combine mining of contextualized attention metadata with developers' quality reputations to identify activities that seem to be inhibitors of quality. By analyzing the behavior of an entire team, adverse activities can be detected automatically, and can be reported back to developers so that they can learn to avoid adverse behavior.