Designing an Adaptive Web Navigation Interface for Users with Variable Pointing Performance
Aqueasha Martin-Hammond, Foad Hamidi, Tejas Bhalerao, Christian Ortega, Abdullah Ali, Catherine Hornback, Casey Means, Amy Hurst · 2018 · Proceedings of the 15th International Web for All Conference (W4A) · doi:10.1145/3192714.3192818
Summary
This paper investigates how to design adaptive web navigation interfaces for people who experience variable pointing ability — difficulty using a mouse to click targets on screen that fluctuates over time due to conditions like early-stage Parkinson's disease, age-related decline, arthritis, fatigue, or stroke recovery. Unlike users with severe permanent motor impairments who typically adopt dedicated assistive technologies, this population often experiences intermittent or gradually changing difficulties and may not self-identify as needing assistive technology. The authors conducted semi-structured interviews with 27 participants across three groups: 5 people with early-stage Parkinson's, 10 older adults (65+), and 12 younger adults (18-34) who all experience intermittent pointing problems. Using a high-fidelity Chrome extension prototype as a technology probe, they explored preferences around three key dimensions: notification design (how to inform users about detected errors), information needs (what users want to know about their pointing performance), and assistance activation modes (manual, automatic, or mixed-initiative). Common pointing challenges included unintentional selection, slipping off targets, overshooting or missing targets, and cursor loss from erratic movement. The findings directly informed the design of PINATA (Pointing Interaction Notifications and AdapTAtions), a Chrome extension that tracks pointing performance over time and provides dynamic assistance.
Key findings
Three core themes emerged from the qualitative analysis. First, participants expected the system to recognize and integrate their individual goals and preferences for how pointing tasks should be completed — not just detect errors, but understand their unique patterns and adapt accordingly. Second, participants strongly valued transparency: they wanted to understand what the system detected, why it recommended a particular assistance, and what data it was using. Abstract indicators (like a simple red bar) were universally disliked; 63% preferred Dialog Box notifications and 30% preferred Bar+ notifications because they provided meaningful information and choices. Third, 78% of participants preferred a mixed-initiative approach where they could review and confirm suggested assistance before it was deployed, rather than fully automatic or fully manual control. All participants wanted to be involved in determining whether assistance was activated. These findings led to the PINATA design framework with three layered principles: Keep Users Informed (KUI), Put the User in Control (PUC), and Help Users Trust the System (HUTS). PINATA implements four components: a User Preferences Manager, Pointing History Browser (with error-type and website-specific visualizations), Interactive Notifications, and an Adaptive Bubble Cursor that dynamically increases its selection area in response to more frequent pointing errors.
Relevance
This research fills a critical gap in adaptive interface design by centering the preferences and expectations of users with variable — not just severe — motor impairments. The key insight for practitioners is that adaptive systems must prioritize user agency, transparency, and trust-building over pure automation accuracy. Users want to understand why an adaptation is happening, have the ability to reject it, and see the system learn their individual preferences over time. The design framework (KUI/PUC/HUTS) provides actionable guidance for anyone building adaptive accessibility features. The work also highlights an underserved population: people whose pointing difficulties are intermittent or gradual, who may not identify as assistive technology users and thus fall through the cracks of traditional accessibility solutions. The PINATA Chrome extension demonstrates that lightweight, browser-based adaptive assistance is technically feasible and could be deployed without requiring users to install specialized software. A key limitation is that the study captured initial expectations from a prototype interaction; longitudinal studies are needed to understand how preferences evolve with actual use.
Tags: adaptive user interface · motor impairment · pointing · assistive technology · web navigation · older adults · Parkinson's disease · user study · personalization · browser extension · bubble cursor · mixed-initiative design · user preferences · transparency · trust