Easier Access with HTML5

Summary: Study how authors use HTML to help HTML5 be more accessible more easily.
Proposed Grantee: Ben Millard
Mentoring: Marco Zehe
Proposal started: 18th March 2008
Proposal revised: 21st May 2008
Other sponsorship: None
Estimated start date: 1st June 2008
Estimated end date: 1st December 2008
Requested amount: $8,000

Summary

Ben will collect numerous documents from the web and assess the authoring practices within them. He will compare this with the semantics and access features in HTML5. He will draw attention to any gaps and help develop realistic ways to fill them.

Problem

HTML is designed for access. Currently, few authors make the most of this. As the primary format for publishing web content, how authors use HTML affects access throughout the World Wide Web.

The elements and attributes of HTML give documents structure and semantics. Assistive devices use these to adapt documents to the needs of their users. But a legacy of presentational markup persists on the web. Elements are used for their appearance instead of their structure and semantics. When authors use incorrect elements and attributes, adaptation fails.

HTML elements are sometimes a near match for a specific use but the specification disallows it. For example, HTML4 defines <kbd> as “Indicates text to be entered by the user.” Keyboard shortcuts are not “entering text” so using <kbd> for Ctrl+T is disallowed. Presentational markup must be used for keyboard shortcuts to conform. HTML5 could redefine elements to allow such near matches, helping authors migrate from presentational to semantic markup.

Authors sometimes hack around bugs in specific products, including those which do adaptation. This is sometimes detrimental to better implemented products. Hacks with differing compromises counter previous hacks, adding to the legacy content problem.

HTML5 is an opportunity to make accessible documents easier to author.

Challenges

Websites, like all commercial products, are built to a budget and a schedule. Professional authors have no time to investigate the mountains of information about web access. Making HTML accessible with simpler markup would make the best techniques clearer.

Filling the gaps in HTML must strike a balance:

Similarly, smarter access features have competing considerations:

Feeding my research into HTMLWG will help the process of specifying HTML semantics and features.

Benefits

Smarter features in HTML5 will make better use of document semantics:

Examples in the specification will show scenarios relevant to what authors do. Simplified techniques will enable accessible documents to be written by people who aren’t specialists.

Work Plan and Deliverables

  1. Collect diverse web pages for each type of content by several methods:
  2. Qualitative analysis of each collected document:
    1. Determine what type of content is being marked up.
    2. Write brief, bullet-point analysis of what markup was used.
    3. Suggest what might be more appropriate.
  3. Quantitative analysis of collections:
    1. Popularity of techniques.
    2. Frequency of “near match” cases.
    3. Patterns in markup which could be used for smarter access features.
  4. Publish links to all documents with all analysis and statistics.
  5. Publish ‘early and often’ during the work for maximum openness and transparency.

Success Criteria

Experience