A Vision for Interactive Suggested Examples for Novice Programmers

Abstract

Many systems aim to support programmers within a programming context, whether they recommend API methods, example code, or hints to help novices solve a task. The recommendations may change based on the user's code context, history, or the source of the recommendation content. They are designed to primarily support users in improving their code or working toward a task solution. The recommendations themselves rarely provide support for a user to interact with them directly, especially in ways that benefit the knowledge or understanding of the user. This poster presents a vision and preliminary designs for three ways a user might learn from interactions with suggested examples: describing examples, providing detailed relevance feedback, and selective visualization and tinkering.