Web-ALAP: A Web-based LaTeX Editor for Blind Individuals
Safa Arooj, Shaban Zulfiqar, Muhammad Qasim Hunain, Suleman Shahid, Asim Karim · 2020 · Proceedings of the 22nd International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '20) · doi:10.1145/3373625.3417009
Summary
This paper presents Web-ALAP (Web-Accessible LaTeX-based Authoring and Presentation), an open-source, web-based LaTeX editor designed specifically for blind and visually impaired users. The system addresses a critical gap in STEM accessibility: while LaTeX is widely used in academia for producing mathematical and scientific documents, existing web-based editors like Overleaf provide poor screen reader support, no auditory feedback on compilation, and no way for blind users to verify the semantic correctness of mathematical expressions. Web-ALAP was developed based on initial interviews with visually impaired LaTeX users, including a STEM professor with over 15 years of experience who identified the lack of math accessibility and debugging support as the major barriers. The system uses a client-server architecture with MiKTeX for PDF generation and ShareLaTeX for compilation. Its interface is a split-screen with an accessible code editor and a PDF preview or error display pane. Three key features distinguish Web-ALAP: a fully integrated text-to-speech engine that narrates each character and word during editing without requiring an external screen reader; assistive debugging that automatically narrates the first error message with its line number upon compilation; and a "Math Mode" that provides natural language descriptions of LaTeX mathematical expressions — for example, translating \frac{3}{4} as "fraction numerator 3 over denominator 4 end Frac." The tool requires zero installation and runs in Google Chrome.
Key findings
A comparative user study with 10 visually impaired participants (average age 23.4, average 9.8 years of screen reader experience) evaluated Web-ALAP against Overleaf paired with JAWS across five tasks involving plain text writing, debugging, and mathematical expression coding. While average task completion time was slightly higher for Web-ALAP (17m 8s vs. 15m 6s for Overleaf) — attributable to the additional time spent listening to audio feedback and Math Mode narration — Web-ALAP showed dramatically better error recovery rates: an average of 3.8 errors fixed per user compared to 1.0 for Overleaf. In the debugging task specifically, 8 of 10 participants successfully fixed errors with Web-ALAP compared to only 1 of 10 with Overleaf. The Math Mode feature enabled participants to catch semantic errors that compilers cannot detect, such as writing the wrong variable in a fraction. The Technology Acceptance Model survey showed Web-ALAP scoring substantially higher across all dimensions: System Accessibility (5.6 vs. 2.65), Perceived Ease of Use (5.2 vs. 3.03), Perceived Usefulness (5.6 vs. 2.6), and Self-Efficacy (5.2 vs. 3.1) on a 7-point scale. Incomplete tasks were far fewer with Web-ALAP (5 vs. 19).
Relevance
This work highlights a significant and often overlooked accessibility barrier in STEM education: the inability of blind students and professionals to independently create and verify mathematical documents. While LaTeX's text-based nature theoretically makes it more accessible than visual equation editors, the lack of auditory feedback for compilation results and mathematical content verification means blind users cannot work independently. The Math Mode feature — converting LaTeX notation to natural language descriptions — is particularly innovative because it addresses semantic verification, not just syntactic correctness. For accessibility practitioners, the paper demonstrates that making an interface screen-reader compatible is insufficient when the content itself (mathematical notation) requires specialized interpretation. The zero-configuration web-based approach also addresses a common barrier where assistive tools require complex local setup, creating additional dependencies and points of failure for users who already face technology access challenges.
Tags: blindness · STEM accessibility · mathematics accessibility · screen readers · web accessibility · text-to-speech · education accessibility
Standards referenced: MathML