Computers in Human Behavior Reports (Aug 2022)
Analysis of software developers’ coding behavior: A survey of visualization analysis techniques using eye trackers
Abstract
The coding behavior of a software developer is not easily gathered and investigated. One accurate and objective approach is to utilize eye tracking techniques to capture how developers' write code, use tools, and read natural language documents and instructions. A beneficial insight is to visualize the eye gazes that capture coding behavior. A software developer tasks, and individual actions create complexity to design eye tracking experiments and to analyze the eye gazes. Our systematic literature survey focuses on published methods for multiple types of static and dynamic changing eye tracking stimuli, especially techniques that use multiple participant-editable types of stimuli presented at once to allow for a realistic coding experience. We propose an eye tracking design and analysis framework, which break down the various stages of software coding that require using different programming skills. Our decision matrix maps objectives for software programming to analysis techniques for comparing the eye gazes among software developers'. Finally, we discuss the limitations of current visualization methods, specifically for user controlled dynamic stimuli consisting of existing source code stimuli. We propose additional visualization techniques to help researchers investigate behavior of software developers' while they work software coding tasks.
