Educational program visualizations using synthetized execution information

Abstract

Visualization is a powerful tool for explaining, understanding, and debugging computations. Over the years, several visualization tools have been developed for educational purposes. Most of these tools feed visualization engines using the raw program state data available provided by the debugger API. While this suffices in certain contexts, there are situations where additional relevant information could aid in building up more comprehensive visualizations. This paper presents two novel visualizations of Paddle, an educational programming environment based on synthesized program execution information. We generate execution traces and relevant program states through static and dynamic analysis of the execution data. The synthesized information captures program behaviors that facilitate the creation of comprehensive and rich visualizations involving arrays that depict position reads, writes, moves, and swaps.