BlackBoardNV: A System for Enabling Non-Visual Access to the Blackboard Course Management System
Vineet Enagandula, Niraj Juthani, I. V. Ramakrishnan, Devashish Rawal, Ritwick Vidyasagar · 2005 · Proceedings of the 7th International ACM SIGACCESS Conference on Computers and Accessibility (Assets '05) · doi:10.1145/1090785.1090839
Summary
This paper introduces BlackBoardNV, a system designed to make the Blackboard Learning Management System accessible to users with visual disabilities through non-visual (audio and keyboard) interaction. The core problem addressed is that screen readers cannot adequately convey the conceptual organization of Blackboard pages — they read content linearly without helping users understand the page structure, distinguish between topics, or avoid spending time listening to irrelevant information. BlackBoardNV tackles this by automatically partitioning Blackboard pages into logically coherent segments based on content analysis rather than HTML structural analysis. The Content Analyzer component examines page content and uses a Structural Analysis module with a Labeling system to assign meaningful labels to each partition. For example, all links from a "Tasks" navigation segment are grouped together, and all content from "My Announcements" is aggregated under that label. The system uses a content-driven approach rather than syntax-based HTML tag analysis, which makes it robust to structural variations across different Blackboard pages and scalable across the entire course management domain since Blackboard pages share consistent content semantics despite varying HTML structures.
Key findings
The BlackBoardNV architecture consists of several integrated components: a Content Analyzer that partitions and labels page segments, a Voicemarking Engine that creates navigable audio bookmarks (voicemarks) stored in a repository, a Dialog Interface Manager that generates VoiceXML-based dialogs for speech-driven exploration, and a Web Page Handler that manages Blackboard authentication and page retrieval. The system enables speech-driven guided exploration where users browse labeled content partitions and select which sections to listen to, making navigation of Blackboard content far less cumbersome. A key advantage of the content-driven approach over syntax-based methods (such as HTML wrappers) is that content analysis extends naturally to all Blackboard pages sharing similar domain semantics, whereas syntax-based solutions are brittle and page-specific. The system is described as "knowledge-free" and fully automatic — it does not require hand-crafted rules or semi-automated knowledge-driven approaches to associate meaning with page content. The Dialog Interface Manager also handles non-HTML content (PPT, PS, PDF, TXT) by invoking the screen reader for those document types.
Relevance
This research highlights a persistent challenge in educational accessibility: learning management systems are often built with complex, visually-oriented page layouts that are difficult for screen reader users to navigate efficiently. While Blackboard has evolved significantly since 2005, the fundamental problems identified — poor conceptual organization for non-visual access, time wasted listening to irrelevant content, difficulty distinguishing page sections — remain relevant to modern LMS platforms. The content-driven partitioning approach is notable because it addresses accessibility at a semantic level rather than relying on the underlying HTML structure, which is an approach that anticipates modern strategies for making complex web applications accessible. The work demonstrates that automated content analysis can bridge the gap between visually organized interfaces and non-visual access without requiring changes to the source application.
Tags: screen readers · web accessibility · education · non-visual access · content analysis · learning management system · assistive technology