Assistive Debugging to Support Accessible LaTeX Based Document Authoring
Ahtsham Manzoor, Murayyiam Parvez, Suleman Shahid, Asim Karim · 2018 · Proceedings of the 20th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '18) · doi:10.1145/3234695.3241013
Summary
This demonstration paper presents ALAP (Accessible LaTeX-based Authoring and Presentation), an open-source extension to the TeXlipse Eclipse plugin that makes LaTeX document authoring more accessible for blind researchers and writers. LaTeX is widely used for scientific and technical documents, but since writing in LaTeX is akin to programming, the lack of accessible debugging tools presents a major barrier for blind users. The authors found that existing LaTeX editors — ChattyInfty, MathType, ShareLaTeX, and TeXlipse — all lacked user-friendly debugging tools for blind users. ALAP integrates Microsoft's text-to-speech engine with TeXlipse and provides two key features: when errors occur during compilation, the TTS speaks the first error message along with its line number, and then the cursor automatically navigates to the exact error position. This eliminates the need for blind users to manually search through code to find errors. The system handles both compile-time and runtime errors, and announces "no error found" on successful builds. Additional keyboard shortcuts maximize source navigation accessibility.
Key findings
The system was evaluated with 12 blind or partially blind computer-literate individuals, divided into LaTeX novices (8 participants, mean age 22) and LaTeX experts (4 participants, mean age 32). Using a Technology Acceptance Model (TAM) framework with six metrics on a 7-point Likert scale, results showed a clear split based on prior experience. LaTeX novices significantly preferred MS Word with JAWS for continuous text authoring (perceived ease of use: 5.8 vs. 3.5, p<0.01), finding the familiar Word/JAWS combination easier. However, LaTeX experts found ALAP equally useful as Word for continuous text (6.3 vs. 6.2) and significantly better for mathematical/non-continuous content (productivity: 6.0 vs. 3.1, p<0.01; perceived usefulness: 6.8 vs. 3.9). Notably, even LaTeX novices rated ALAP as equally proficient as Word for non-continuous mathematical content (4.5 vs. 3.9). Qualitative feedback highlighted the assistive debugging as the most valued feature — participants praised the automatic error redirection, noting they no longer had to manually navigate each line to find errors. Ten of 12 participants agreed the assistive debugging would help them code independently in LaTeX.
Relevance
This paper addresses a specific but important barrier in STEM accessibility: the ability for blind researchers to author technical documents containing mathematical content. While MS Word with JAWS works well for continuous prose, it handles mathematical notation poorly — a limitation that pushes researchers toward LaTeX, which then presents its own accessibility challenges around code-like syntax and debugging. The finding that prior experience dramatically affects tool preference (novices prefer familiar tools; experts prefer specialized ones) has practical implications for training and tool deployment. For practitioners, the paper demonstrates that accessible developer tools — particularly debugging with speech-based error navigation — can meaningfully bridge the gap between blind and sighted users in technical workflows. The work from LUMS in Pakistan also represents accessibility research from the Global South, where resources for accessible technology development may be more constrained but needs are equally pressing.
Tags: mathematical accessibility · LaTeX · blind and low vision · text-to-speech · document authoring · STEM accessibility · debugging · screen readers · programming accessibility