Date of Award
2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Computer Science
Committee Chair
Feng Zhu
Committee Member
Haeyong Chung
Committee Member
Jacob Hauenstein
Committee Member
Chaity Banerjee
Committee Member
Jodi Price
Subject(s)
Eye tracking, Eye tracking--Analysis, Computer security, Computer programming
Abstract
Secure coders' experiences and their proficiency vary greatly, and any overlooked software security flaws in code can lead to costly repercussions in deployed software applications. The techniques that secure coders utilize to analyze source code and develop mitigation strategies for security flaws are not well understood. Gaining a proper understanding of how coders approach finding and mitigating security flaws can help us efficiently and accurately discover and resolve such issues. One potentially beneficial technique is to collect, analyze, and visualize eye gazes that capture their coding patterns and behaviors. Our systematic literature survey focused on published methods for multiple types of static and dynamic changing eye tracking stimuli, with a particular emphasis on techniques using multiple participant-editable types of stimuli presented simultaneously to simulate a realistic software coding experience. Our work proposes an eye tracking design and analysis framework that breaks down the various stages of software coding. Our decision matrix maps objectives for software programming to analyze techniques for comparing eye gazes among software developers. This involved investigating the limitations of current visualization methods, specifically for user-controlled dynamic stimuli. Our investigation involved using eye tracking technologies to capture how developers write code, use tools, and read natural language documents and instructions. The study encompassed a wide range of tasks, including simultaneously reading documentation, writing code, and using security source coding analysis tools. Software developer tasks and individual actions create complexity in designing eye tracking experiments and analyzing the collected eye gazes. Our approach allows us to explore behaviors across a range of tasks for a single secure coder and among different coders. New visualization techniques were developed to investigate behaviors during secure coding tasks including methods to present transitions among components within and between applications, as well as present coders' attention levels during secure coding. Our contributions include a literature survey, framework design, secure coding learning modules, scrollable and modifiable eye tracking stimuli analysis, pupil diameter changes analysis, and stimuli presented in different sequences based on individual participants' behavior. Our contributions focus on comparing and contrasting multiple visualization methods for eye tracking stimuli.
Recommended Citation
Davis, Daniel Kyle, "Eye tracking technologies to analyze and visualize the behavior of secure coders" (2023). Dissertations. 375.
https://louis.uah.edu/uah-dissertations/375