WCAG 2.0: Normative checklists?

Background

Requirements for WCAG 2.0 states,

1. "WCAG 2.0 requirements should be expressed in generic terms so that they may apply to more than one markup language or content format."

   and

2. "Each requirement must be verifiable" and that checklists, techniques, and test suites will provide the necessary detail and tests to determine conformance.

Do general success criteria provide enough information to determine conformance? If not, is it possible to make them testable enough to determine conformance? If not, should we conclude that checklists must be normative? How would putting checklists on the Recommendation track effect our timeline? Can we satisfy the Requirements for WCAG 2.0 with informative checklists?

This document attempts to frame issues for WCAG WG discussion. The issues seem to divide into two themes: Interpretation and Timing.

Interpretation

Audience

• Who is the primary audience we expect will be claiming conformance to WCAG 2.0?
• Will it be people who are creating HTML pages to advertise their business? Or will it be the people who develop the authoring tool used to generate the HTML pages?
• Is our audience people who use XML and XSLT to generate a product catalog? Or the people who develop the policy about how to make the product catalog accessible?
• What are the benefits of normative checklists for each of these audiences?
• Are we sure we know the answers until we've polled them?

Standards harmonization

• If a WCAG 1.0 test suite for HTML had been available at the same time WCAG 1.0 was published, would there have been as many diverging policies and interpretations of WCAG 1.0?
• Would it have made a difference if that test suite was normative?
• What is the impact of the informative test suites for SVG and CSS1?
• What is the impact of the normative test cases for OWL?
• What is the impact of the Markup Validator?
• Which of these approaches (informative test suite, normative test suite, validation tool) has the greatest impact?
• Will normative checklists increase the likelihood that WCAG 2.0 will become the convergence target for international standards harmonization?
• See also: Updates, Timing: Effect on standards harmonization.

Conformance claims

• If checklists are normative, will people claim conformance to the checklist?
• Will we need separate logos for each checklist?
• Will normative checklists effect how many people read the guidelines? (i.e., if checklists are normative, will people be less likely to read the general principles?)

Which technologies?

• Which technologies do we create checklists for? How do we decide?
• If checklists are normative, don't we need a static "view" that can ascend through the W3C Recommendation track? Would this static view include all technologies? Would we have one checklist for each technology? What about technologies like CSS and scripting that rely on a "host" technology?
• One reason to make WCAG 2.0 non-technology-specific is that people could use non-W3C technologies and conform to WCAG 2.0.
  • If checklists are required for conformance, how can someone make a conformance claim using a non-W3C technology? (To remain vendor-neutral the W3C will not publish documents about PDF or Flash. ECMAScript is being considered since it was developed by ECMA and accepted by ISO/IEC and has no licensing terms.)
  • Could PDF, Flash or other tehcs be checked against the guidelines themselves, with the aid of any non-normative checklists?
  • Is it possible to define a conformance scheme for checklists, i.e., a checklist meets the guidelines by applying them to a technology? Content wouldn't conform to the guidelines directly but the checklist would.

Updates

• User agents and assistive technologies change. New technologies are adopted. New techniques are discovered. Checklists will need to be updated to reflect changes in the world.
• Potential opportunities
  • Greater interest and necessity to update checklists.
• Potential barriers
  • Only documents that do not introduce new "features" can be published as Edited Recommendations; it is likely that to update normative checklists it would require taking the document through the entire W3C Recommendation process.
  • If each publication of a checklist is normative, which one does an organization adopt as policy? Already, organizations that have adopted WCAG 1.0 are concerned about what to do when we publish WCAG 2.0.

Timing

Coordination

• Because we are developing materials about other W3C technologies, we will need review by and coordination with the Working Groups who have developed the technology.
• If HTML Checklist for WCAG 2.0 is normative and restricts the use of HTML 4.01 or XHTML 1.1 (existing W3C Recommendations) the process to move the HTML Checklist through the W3C Recommendation track will require a higher level of coordination with and sign-off by the HTML WG. (higher level of coordination than if the document were informative).
• Potential opportunities:
  • Closer coordination with the HTML WG would give them a better understanding of accessibility issues.
  • The HTML WG could incorporate suggestions into XHTML 2.0.
• Potential barriers:
  • It is difficult to estimate how much time this coordination would take nor how it would effect our timeline.
  • We would have to approach them delicately to ensure that we are not "stepping on their turf."
  • (this amount of effort) * (the number of technologies we address) = (time needed to complete WCAG 2.0)

Effect on current timeline

The Milestones and Deliverables sections in our draft charter are based on the presumption that development of checklists and techniques is staggered across the next several years. The idea was that at Last Call we could demonstrate applicability of WCAG 2.0 to a variety of technologies with fairly exhaustive coverage of general, html, css, and scripting/web applications techniques and checklists while checklists and techniques for rdf, smil, svg, rdf, and voicexml would be less complete. The milestones show that all of the techniques and checklists would continue to evolve over the course of the WCAG WG charter (proposed to end in 1Q 2006).

If checklists become normative, it is difficult to predict the effect that publishing multiple Recommendation would have on our current timeline. The existing timeline already quite aggressive.

Effect on standards harmonization

The longer we take to develop WCAG 2.0 the more likely other organizations will publish hybrid documents based on WCAG 1.0 and extended to meet the organization's circumstances.

Other options

Ideas to encourage creative thinking.

• Divide success criteria/checklists into two categories (1. those that rely on user agent/assistive technology implementation and 2. those that do not rely on user agent/assistive technology implementation). Group 1 could be normative since they would need more vetting and rely on user agent implemenation. Group 2 could be informative since they could be updated more frequently to immediately increase accessibility (since they do not rely on user agent implementation). Refer to Gregg's message from 12 May for more details.
• Publish Checklist for WCAG 2.0 (HTML,CSS, scripting, general) in 1st quarter 2005 followed by other checklists and WCAG 2.0 later in 2005.
• Publish WCAG 2.0 as a Working Group Note and the technology-specific checklists as Recommendations.
• Create conformance profiles that provide more detail but not checklist-level detail.
• Create conformance modules or in some way divide WCAG 2.0 so that we can more easily stagger suites of Recommendations (look to DOM and CSS3 for models on modules).

Diagram of the issues

To help collect issues related to the topic, I created a "map." It's just a brainstorming tool, so others might not find it useful, but it helped me collect the issues. jpg version of map.

Description: There are two primary concepts - Interpretation and Timing. Each concept is in a circle and hanging off of each circle are several concepts. The following is an outline of the concepts.

Interpretation

• Primary audience?
• Forces more frequent updates
• Readers use checklists instead of guidelines?
• By evaluation, repair, and authoring tool developers (related to WCAG 2.0 Test Suite)
• WCAG 2.0 Test Suite
  • Are normative checklists the only solution?
• By policy makers
  • Ask policy makers if normative tests are requirement to adoption.
  • Latest Access Board comments ("we could harmonize on direction you are taking with 11 March draft")
  • EU adopted WCAG 1.0 as is, writes own checklist
• More review? Effect on future W3C Recommendations? (related to "Managing conflicts with other W3C specifications")
• Harmonization of standards/policies? (related to "Timing")
  • WCAG 1.0 not convergence target because tests were not normative versus WCAG 1.0 not convergence target because did not contain detailed technology-specifics.
• Testing for conformance
• Non-W3C technologies?
• Taking to ISO?
• Which technologies? How decide?
• Claiming conformance to checklists
  • Separate set of conformance claims and logos? One for each technology?

Timing

• Related to current timeline and milestones (staggering the development of techniques)
• Updating normative documents
  • Entire W3C Recommendation track for each update
• Managing conflicts with other W3C specifications
  • Political issues related to turf?

$Date: 2004/06/09 03:24:33 $ Wendy Chisholm