Understanding Screen Readers' Plugins
Farhani Momotaz, Md Touhidul Islam, Md Ehtesham-Ul-Haque, Syed Masum Billah · 2021 · Proceedings of the 23rd International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3441852.3471205
Summary
This paper presents the first empirical study focused specifically on understanding the ecosystem of screen reader plugins — small pieces of code that extend screen reader functionality. The researchers conducted semi-structured interviews with 14 blind screen reader users (ages 24-52, 10 male, 4 female) recruited from diverse professional backgrounds across North America and South Asia. Most participants were expert users of JAWS and NVDA, the two dominant screen readers. The study addressed two research questions: why blind users rely on plugins, and how plugins are developed, maintained, and distributed from the users' perspective. The research is motivated by the observation that while plugin ecosystems are well-studied for browsers and other software (Firefox has over 450,000 add-ons), screen reader plugins remain virtually unstudied despite being widely used. The authors note that screen reader quality plateaued in the mid-2000s and has not kept pace with the post-PC era, making plugins an increasingly important mechanism for addressing accessibility gaps that screen readers alone cannot solve.
Key findings
Users rely on plugins for four main reasons: making inaccessible applications accessible, modifying screen reader behavior for better usability (e.g., reducing NVDA's excessive verbosity in Zoom), adding audio feedback on keystrokes, and creating custom shortcuts. Installation is straightforward, but finding the right plugin is ad hoc and frustrating — there is no central repository for most screen readers (only NVDA has a community add-ons page, and it is considered inadequate). Plugin development is concentrated among a handful of well-known individuals recognized by first name in the blind community. Most users want to develop their own plugins but face prohibitive barriers: inaccessible programming IDEs, steep learning curves, and 20-30 page developer guides. Critically, maintenance is essentially nonexistent — once deployed, plugins rarely receive updates and become obsolete when screen readers or target applications change. One participant reported refusing to update NVDA to preserve plugin compatibility. Security is a serious concern: plugins are distributed without security checks, and one participant lost data to ransomware disguised as a plugin. Non-native English speakers face additional challenges finding plugins that support non-Latin scripts and languages.
Relevance
This study has direct implications for accessibility practitioners and screen reader vendors. The finding that plugins serve as a critical stopgap for application-level accessibility failures highlights a systemic problem: developers are not building accessible software, and blind users must rely on community-created patches to use everyday applications. The recommendations — a centralized plugin store with ratings and security checks, financial incentives for developers, and step-by-step plugin creation tools — offer a practical roadmap for improving the ecosystem. For accessibility researchers, the plugin-based distribution model represents an underutilized channel for deploying accessibility interventions directly to users rather than waiting for application developers to act. The study's limitation is its focus on expert users who were already familiar with plugins, potentially missing the experiences of the many screen reader users who are unaware plugins exist.
Tags: screen readers · plugins · blind users · assistive technology · NVDA · JAWS · extensibility · user study · accessibility tools