Graduation Semester and Year
Summer 2026
Language
English
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Computer Science
Department
Computer Science and Engineering
First Advisor
Christoph Csallner
Second Advisor
Allison Sullivan
Third Advisor
Jeff Lei
Fourth Advisor
William Beksi
Abstract
Software can meet its functional requirements yet still fail to support the people who rely on it. Accessibility and usability gaps persist across software systems, affecting users with diverse needs. This dissertation investigates how software systems and development practices can better support those users. We developed and deployed Senior Fit, a standalone smartphone-based physical activity application for older adults, and conducted a six-month field deployment with 25 participants to examine how they incorporated the application into everyday routines. The study identifies how exercise guidance, feedback, privacy, social support, and flexibility shape self-guided use and informs the design of mobile health tools for older adults. We also developed Altify, a development-time tool that generates context-aware alt-text suggestions for mobile user-interface icons. A formative study with 52 developers informed its design, and an automated evaluation against human-authored descriptions assessed the semantic alignment of its suggestions. To examine accessibility within emerging development environments, we mined public issue reports and discussions across five AI developer-tool ecosystems and characterized reported visual accessibility barriers. We also examined repository-level response patterns in the GitHub-based issue-tracking systems. The rapid adoption of autonomous coding agents also motivates a separate empirical investigation of software quality practices in agent-generated contributions. We analyzed 33.5k agentic pull requests to examine how autonomous coding agents incorporate tests into software changes. Test-containing pull requests became more common over time, although testing patterns varied across agents. Together, these studies contribute practical approaches and empirical evidence for building more inclusive software and understanding emerging AI-enabled development practices.
Keywords
Software, Usability, Accessibility, Alt-text, Older adults, Visual accessibility, Software testing, Agentic coding
Disciplines
Computer Sciences
License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.
Recommended Citation
Haque, Sabrina, "Toward Inclusive and Reliable Software: User-Centered Design, Accessibility, and AI-Enabled Development" (2026). Computer Science and Engineering Dissertations. 15.
https://mavmatrix.uta.edu/cse_dissertations2/15