Profiling Learners with Special Needs for Custom E-Learning Experiences, a Closed Case?
Paola Salomoni, Silvia Mirri, Stefano Ferretti, Marco Roccetti · 2007 · Proceedings of the 2007 International Cross-Disciplinary Conference on Web Accessibility (W4A) · doi:10.1145/1243441.1243462
Summary
This paper from the University of Bologna tackles the problem of creating comprehensive learner profiles that capture both user accessibility needs and device capabilities — two dimensions that existing standards addressed separately but not together. The authors argue that no single profiling specification available at the time could fully describe a learner's context. IMS ACCLIP (Accessibility for Learner Information Package) excels at describing user accessibility preferences — display settings like cursor size, font preferences, screen reader configuration (speech rate, pitch, volume), control preferences (alternative keyboards, voice recognition, mouse emulation), content format preferences (alternatives to visual, auditory, or text), and eligibility for accommodations — but it ignores the device being used. Conversely, W3C's CC/PP (Composite Capabilities/Preference Profiles) thoroughly describes device hardware (screen dimensions, audio capability, Braille display presence), software (operating system, installed assistive technologies like JAWS), and browser capabilities (supported formats, CSS versions, HTML versions) but says nothing about the user's personal needs. The authors' key insight is that the union of these two specifications — ACCLIP Profile + CC/PP Profile = Complete Profile — covers the full range of characteristics needed to adapt e-learning content. They identify where the specifications share common definitions (particularly around assistive technologies) and show how to resolve potential conflicts when both profiles describe the same feature.
Key findings
The paper demonstrates its profiling approach through three detailed scenarios involving a SCORM-packaged SMIL video lecture composed of video, audio, slides, captions, and textual descriptions. Scenario A shows a non-disabled learner on a PC that happens to have a screen reader installed — the system correctly ignores the screen reader because the ACCLIP profile does not request it, even though CC/PP reports its availability. Scenario B describes a deaf learner on a PDA with limited capabilities: ACCLIP specifies visual alerts instead of audio, mouse emulation via keypad, and captioning preferences, while CC/PP reveals the small screen (240x320), Pocket Internet Explorer, JAWS 6.20, and no SMIL support — the system adapts by providing visual content with captions and no audio-dependent elements. Scenario C profiles a blind learner on a fully equipped PC with screen reader, Braille display, and SMIL player — the system activates all assistive technologies and provides audio-based alternatives to visual content. The profiling system was integrated into a content adaptation platform with three components: a Media Broker that receives requests and retrieves profiles, a Profile Manager that stores and manages user/device profiles (supplemented by WURFL, a database of 7,000+ device descriptions), and a Transcoding Sub-System that uses web services to transform learning objects based on the combined profile.
Relevance
This paper addresses a challenge that remains fundamentally unsolved in accessible e-learning: how to automatically adapt educational content based on a complete understanding of both the learner's needs and their technology environment. While the specific standards discussed (ACCLIP, CC/PP, UAProf) have evolved or been superseded, the core problem persists. Modern learning management systems still struggle to combine user accessibility preferences with device capabilities for dynamic content adaptation. The paper's insight that assistive technology presence on a device does not necessarily mean the user needs it (Scenario A) remains an important design consideration — systems should not assume that detecting a screen reader means the user is blind. The three scenarios effectively illustrate the complexity of real-world adaptive delivery and serve as useful test cases for any content adaptation system. The work also highlights the importance of standardized, interoperable profiling approaches over proprietary solutions, a principle that continues to drive efforts like the W3C's personalization specifications.
Tags: e-learning · user profiling · content adaptation · assistive technology · device capabilities · accessibility standards · learning objects · mobile accessibility
Standards referenced: IMS ACCLIP · CC/PP · IMS LIP · ACCMD · SCORM · SMIL · UAProf