Skip to content

Home

Jump to bottom
Ian-J-S edited this page Dec 15, 2025 · 18 revisions

A11yTagger

a11ytagger is a tool for adding accessibility tags to PDFs, with the ultimate goal of providing full PDF/UA support.

UX Team Members

  • Colin Jamison - Wireframe, Generation of Personas, Prototype interactivity, Methodologies of Phase II and III.
  • Lars Bartels - Generation of Personas, Wireframe, Phase 1 draft
  • Ian Stewart - Added personas, scenarios, wireframes, & executive summaries to wiki.

User-Centered Design Artifacts

Phase I: Analyzing Users, Competitors, and Initial Designs

Executive Summary

  • Competitive Analysis revealed key weakness of other accessibility tagging options:
    • Too expensive: many existing options are prohibitively expensive for the average user.
    • Vendor lock-in: some options are too tightly coupled with preexisting software that the user might not want.
    • Bad auto-tagging: many existing options use error-prone auto-taggers that require time-consuming manual fixes.
  • Heuristic Evaluation showed us how to improve upon PAVE:
    • Inconsistent user interaction: PAVE suffers from inconsistent interaction design, confusing users.
    • PAVE's issue detection quickly shows the user exactly what they need to start with when remediating.
  • Personas and Scenarios expand our insight into user needs:
    • Fast and easy tagging is a must!
    • Users must have a consistent and easily learnable workflow to follow when tagging documents.
  • Sketches consolidated ideas for initial concepts:
    • Multiple divergent sketches allowed us to narrow down our best ideas for design and workflows.
    • Better error reporting will prevent user frustration and allow them to recover easily.
    • A well-defined workflow will enable users to easily learn a system for quickly remediating documents.

Full phase I report

Phase II: Refining interaction and designing wireframes

Executive Summary

Improving user workflows became a primary goal after conducting cognitive walkthroughs. More specifically, it became clear that the application would benefit from better feedback on loading or error states. Reviewers also noted that the wireframes have minor inconsistencies in navigation. Feedback from a product demonstration revealed a need for OCR support in addition to a dedicated mode for image tagging. This image tagging mode will enable a list view for users to quickly browse images and add alt-text, with some display of progress remaining. Feedback also highlighted a need for improved tool layouts. Findings pointed out a general need for a more streamlined path for user editing. Wireframes were updated accordingly, adding both detail to tool designs and additional information pages. Research methods were limited due to a small number of evaluators, a lack of hands-on product testing, and inexperienced UX evaluators. Future testing will include a hands-on usability test of a working prototype.

Full phase II report

Phase III: Prototypes and User Testing

Executive Summary

  • This phase focused on designing and evaluating a Figma prototype which addressed issues found in the cognitive walkthroughs. After designing the prototype, we conducted user tests to determine whether the prototype met basic requirements for satisfaction, efficiency, and effectiveness.
  • Prototype design expanded and clarified user workflow to better communicate the state of the application and task progression.
  • User tests allowed us to gain insights into important aspects of the usability of our prototype.
    • Tests were conducted with n=5 participants, moderated by members of our team.
    • Participants were given a set of core tasks: PDF upload, accessibility issue identification, alt text editing, and PDF export.
    • Think-aloud style tests and follow-up questions were used to collect qualitative feedback. Timing of user tasks was used to gather quantitative data for efficiency analysis.
  • Key findings:
    • Most participants completed all tasks, indicating that the workflow was understandable and effective.
    • On average, tasks were completed quickly, which suggests high efficiency.
    • Users without accessibility experience struggled with accessibility-related terminology.
    • Other users found that certain screens were too cluttered.
  • Conclusions:
    • The prototype enables efficient and effective completion of tasks, but overly specific terminology can hinder satisfaction.
    • Future improvements should simplify the language to reduce cognitive load.
  • Caveats:
    • User tests were conducted with participants coming from a single university course, which may not reflect real users.
    • Figma prototyping constraints limited realism of the prototype for text input during image tagging.

Full phase III report

Clone this wiki locally