Towards Self-Adaptive IDEs

Abstract

Developers use Integrated Development Environments (IDEs) to maintain and evolve software systems. IDEs facilitate development activities such as navigating, reading, understanding, and writing source code. Development activities are composed of many basic events, such as browsing the source code of a method or editing the body of a method. We call these actions "interaction data". We believe that collecting, processing, and exploiting these interactions at run-time can potentially augment the productivity of developers. Our goal is to create self-adaptive IDEs: IDEs that collect, mine, and leverage the interactions of developers to better support the developers' workflow. We envision a development environment that automatically and seamlessly adapts itself to support developers while maintaining and evolving software systems. To reach our goal, we will develop means to reshape the user interface of the IDE, interaction-based recommenders, and integrate live and adaptive visualizations inside the IDE. As a first step towards our vision, we have developed DFlow, a tool that non-intrusively records all IDE interactions while a developer is programming. At the moment DFlow collects all the interactions between the developer and the IDE, and enables retrospective analysis by means of software visualizations.