W3C Web Accessibility Initiative

Open issues sorted by topic

General issues

5. Important


10 August 2000 - William Loughborough. I am very leery of leaving the decision about importance entirely up to the author.


8. Making checkpoints clear


11 August 2000 - Gregg Vanderheiden. We should not imply that someone needs to follow all of the related checkpoints to satisfy a guideline. If the checkpoints give alternatives, it needs to be clear that implementing one of the two alternatives satisfies the checkpoint.

9. Audio descriptions


11 August 2000 - Gregg Vanderheiden. Is the requirement to provide an auditory description of a visual track redundant with the requirement to provide a text description of a visual track? Are both required? An auditory description (synchronized) is prefered, except for people who are deaf and blind.

Checking with Gregg to determine if it is closed. Working group believes it is (as of 23 August 2001 telecon, issue #41)

10. Priorities and conformance


11. Bidirectional language issues


14 September 2000 - Lisa Seeman. In Visual Hebrew you have to use absolute width with page layout tables to create the appropriate breaks or the text becomes unreadable. This conflicts with current guidelines.

14. Usability testing of WCAG 2.0

Kynn Bartlett - 19 Sept 2000. We need to test our documents for usability.


17. Technology-specific normative info?

Some people feel that there should be normative technology-specific information that an author can claim conformance to. Should we do this? If so, how and what do we call them? People have argued that "technology-specific checkpoints" is confusing because then we will have used the word "checkpoint" in two different layers.

Refer to:

22. Public submission of techniques

At the5 October F2F, Tom suggested, "In publishing techniques, encourage people to submit techniques without strictly conforming to the structure of the document." Other suggestions about techniques:


23. New requirement: notification and alternatives

The discussion on "cyber-ghettos" and generating views of content at the 5 October 2000 F2F, produced the following 2 proposals for a new requirement:

Other comments and related issues:

26. Levels of abstraction

From the October 2000 F2F: IJ suggested that there are more or less abstract requirements in three different WAI guidelines. Perhaps the WG should try to establish criteria early on for determining whether a checkpoint is sufficiently or overly abstract. (WC it is not clear to me what is required by the WCAG WG. I have requested that IJ clarify this comment).

Proposed resolution

We have made the top layer (Guidelines and Checkpoints) abstract in the sense that there is no technology-specific information and that we push most of the details to technology-specific technique documents which will contain technology-specific checkpoints (or whatever they are finally called). Feel that there is consistency in the abstractness between Guidelines and between Checkpoints. Do we need to be consistent with UAAG and ATAG as well? No, since we feel our audience of the Guideline/Checkpoint layer is not as technical as the audience of ATAG, UAAG, and our technology-specific documents. Need confirmation from Ian that this is reasonable.

30. Content not intended for users

At the October 2000 F2F (during reports from subgroups), the server-side techniques group asked: What if you have content available but not necessarily for display by the content provider (e.g., in a database)? We can't ignore this because even if it is not intended for users, if it is intended for machines that in some way will help users....(what was the example?) Proposals?? Further questions or comments?

33. Guideline on testability

34. Mapping between 2.0 and other guidelines

At the 4 January 2001 telecon it was suggested that we not only provide a mapping between WCAG 1.0 and WCAG 2.0, but between WCAG 2.0 and other Web guidelines as well. This will help others understand our work as well as help us collect feedback and ideas.

38. Ordering of checkpoints

At the 11 January 2001 telecon we resolved to include a checkpoint on graceful transformation in guideline 1. Some people felt it should be the first others felt it should go last (refer to the minutes from the 11 January 2001call for further discussion). We expect that as we add priorities we will reorder the checkpoints but resolved to make it an open issue so as not to forget these concerns.

52. Examples - good and bad?

All of the examples are of things that will pass the success criteria. Do we want examples of things that won't pass? Might be a lot of examples...

53. Optimizing for disability types

Paul proposed a section of the Introduction called "Optimizing content for Specific Disability Types." Should we include it in the intro? Or is it part of the server-side scripting techniques and the "appendix" that paul mentioned at the june F2F (refer to WC's question to him on 23 July 2001).

54. Executive Summary

55. Limitations/scope - usability/accessibility

Paul proposed a "Limitations" section that discusses the usability/accessibility conendrum which we have not yet resolved. It sets expectations that we discuss some usability issues but will not attempt to cover them all.

65. Meta data

Charles McCN's comments about meta data (as success criteria for 2.1) are very important but since they deal with machine-generated/readable searches, is that success criteria for 2.1 or 2.7? or does it deal with 1.3 - using markup and database to provide structure...and something more. semantics. This is the other issue Jason has brought up before that we have not captured well.

66. Examples - diversity in interfaces

Most examples are GUI-oriented. We need more diverse examples, for example speech interfaces. Otherwise we will fall into the WCAG 1.0 trap of being general on the surface but assuming GUI agent/HTML-dominated Web underneath. Jason White.

77. Documenting assumptions

This combines open issues: 1, 6, 15, 28, 40.

Assumptions that need documenting:

Conformance issues

29. Need for invalid or deprecated markup

At the October 2000 F2F (during reports from subgroups)the server-side techniques group noted: There are some exceptions to the WCAG 1.0 proper markup requirement. To support older browsers, you may need to send invalid or deprecated markup. Currently, this would not conform to WCAG 1.0. How will we handle in WCAG 2.0?

31. UA-specific conformance claims?

At the October 2000 F2F (during reports from subgroups), the server-side techniques group noted that it's not clear when you claim conformance that you are claiming conformance for certain user agents. Proposals?? Further questions or comments?

36. Hardships and conformance

Proposal by Len Kasday - 30 December 2000. Proposes that we have the following 4 ways to deal with people raising objections over conformance to a checkpoint:

  1. Throw out this Guideline/checkpoint because there's a consideration X objection to it in some circumstances - which implies throwing everything out since there have been objections to the most basic checkpoints.
  2. Just go on like we're doing and arguing consideration X objections to each checkpoint as it comes along, and throw out the checkpoint if the objection is--what--serious enough? affects too many webmasters? What's the criterion?
  3. Focus now only on accessibility and return to the consideration X problem in a comphrehensive, consistent way later. For example,
    • Just say we only defining accessibility, and not considering considerations X (like WCAG 1.0 seems to say)
    • Make a blanket policy allowing violation of checkpoints in cases where there are legitimate consideration X concerns
    • Define "qualified" compliance that refers to considerations X
    • Make a detailed catalog of all the hardships that each consideration may entail.
    • -- your suggestion here --

Gregory Rosmaita and Gregg Vanderheiden propose that for each checkpoint we offer an option that says, "I did not use any of these proposed techniques. Instead I did it like this and this is how it is accessible."

76. Modularizing the guidelines

Charles Munat proposed on 20 August 2001 to split the guidelines into 3 separate specifications.

Techniques issues

27. Linking between techniques documents

At the Octoboer 2000 F2F (during reports from subgroups)this came up specifically in regards to SVG linking to other techniques documents (like CSS). How should the various techniques documents link to each other?

32. Tool-specific requirements

Raised at the October 2000 F2F. For some proprietary formats, such as PDF, what the author can do is dependent on the authoring tool. Is it helpful to classify those? should this info be attached to the technique somehow?

58. Skip navigation link

The following phrase has been deleted:

a link that jumps over navigation links and positions the user at the beginning of the primary content on the page

We seem to agree that this is a technique, but under which checkpoint? Which technologies should we discuss it in?

Interface and presentation issues

35. Help feature

At the 4 January 2001 telecon it was suggested that we create a help feature similar to Microsoft help to assist people in finding the information they need. Basically, providing a variety of views to accomodate the variety of users of our materials.

50. Visual presentation

63. Style of normative vs non-normative info

Most W3C specs seem to only identify that which is informative and assume the rest normative. There usually isn't much that is informative. Each checkpoint text and success criteria is marked with div class="normative" so that we could apply a style sheet to them later. In the 14 August 2001 draft, marked definitions, examples, and benefits as informative, e.g., "Definitions (informative)".

Glossary issues

37. Glossary

Once a glossary has been endorsed by the various WAI groups, WCAG 2.0 should use terms as defined in the glossary. Refer to the minutes from the 11 January 2001 telecon for discussion.

How we are using the word "content" was raised by Ian Jacobs on 16 January 2001. At the 18 January 2001 teleconthe group decided not to change anything until the glossary was solidified.

Non-text content and presentation are also issues raised with the 14 August 2001 working draft.

New proposals

Checkpoint 1.1 Provide a text equivalent for all non-text content.

44. Revisting Checkpoint 1.1

21 May 2001 - William Loughborough uses XMLGL as a model for repurposing all content, including text. In other words, providing content to illustrate text. Something along the lines of, "provide alternative content for all content and pieces of content"

Checkpoint 1.2 Provide synchronized media equivalents for time-dependent presentations.

41. Synchronized equivalents for audio-only timed interactions

Raised at the 18 January 2001 telecon. WCAG 2.0 does not address the case of an audio-only presentation that requires a timed response. GV proposes to modify checkpoint 1.2 to read, "Synchronize text equivalents with multimedia and time-based interactive presentations" Others feel that this might be addressed by a combination of 1.1, 1.2, and 2.4.

67. Tolerance of synchronizing captions and audio descriptions

Checkpoint 1.2 in the 14 August 2001 drafthas the following success criterion and note:

3. descriptions and captions are synchronized with the events they represent to within a tolerance of X. [Note: We need to research the tolerance. Any information on this is appreciated.]

68. Real-time broadcasts

Checkpoint 1.2 in the 14 August 2001 draft has the following success criterion:

if the Web content is a real-time broadcast, it is possible to provide real-time commentary (as with a sporting event) and real-time captioning. If this creates an undue burden, the presentation is available after-the-fact with captions and audio descriptions

There are issues with how it is worded (an if followed by an if), but it also seems to be a statement of fact. (Jo Miller - 14 August 2001)

Jo Clark (16 August 2001) warns that we should not advance play-by-play as something similar to audio description.

Checkpoint 1.3 Use markup or a data model to provide the logical structure of content.

57. Data model

Isn't a data model inherit in the language that you choose? for example, the RDF model vs. the XML model? Therefore, will an author directly insert content into a data model? e.g. refer to the XPath and XQuery data model spec, it's a processor that inserts data into the data model. Is that within our scope? WC proposes we delete the words "data model" from Checkpoint 1.3 Use markup or a data model to provide the logical structure of content.

Checkpoint 1.4 Identify the primary natural language of text and text equivalents and all changes in natural language.

75. Broaden "identifying languages" to include other semantic distinctions

Jason White raised this in an editors meeting. That we broaden checkpoint 1.4 of the 14 August 2001 draft to encompass other semantic distinctions. This would make it parallel with 1.3. Or we ought to make it a special case of 1.3.

Checkpoint 1.5 Separate content and structure from presentation.

4. Semantics in style


8 August 2000 - William Loughborough. If I use style to convey semantics, how do I pass along the semantics to someone who does not use style sheets? The class name is not something that gets shown to the user. There must be some way to convey intended semantics presented via presentation mode.

Checking with william to determine if it is closed. Working group believes it is (as of 23 August 2001 telecon, issue #4)

Checkpoint 2.1 Provide multiple site navigation mechanisms.

Checkpoint 2.2 Provide consistent and predictable responses to user actions.

60. Consistent and predictable responses to user actions

Checkpoint 2.5 Use device-independent event handlers.

70. Device-independent event handlers

Checkpoint 2.5 of the 14 August 2001 draftincludes the following note after the success criteria:

Note to reviewers: In criterion 2, one of the device-specific event handlers must be encoding (e.g., keyboard) although it has been noted that encoding device handlers do not appear to be usable in lieu of direct manipulation (e.g., mouse) in those cases where the action or the result can not be expressed in words. Our challenge is to express this simply. Feedback is encouraged.

We need to make sure that all interaction modes are accounted for. More than one may be provided, but if is not one that I can use, it's no good.

Checkpoint 2.6 Avoid causing the screen to flicker.

61. Examples of and tests for flicker or non-flicker

Checkpoint 2.6 does not currently have examples. How do you test for flicker? Doesn't it depend on the hardware? Isn't it different for each monitor? Is this an authoring issue?

Checkpoint 2.7 Handle input errors, such as misspellings.

71. Handling input errors

Checkpoint 2.7 of the 14 August 2001 draftincludes the following note which outlines the issues with this checkpoint:

Note to reviewers: This is a new checkpoint that is being explored. It does not have full support of the working group. We know that spelling mistakes are a serious issue for people with writing disabilities and dyslexia. We have generalized this checkpoint to include all input errors but highlight spelling mistakes. Are there other input errors we should highlight? Are there other success criteria we can define? If not, then we will not use the general case and make this specific to spelling errors.

Checkpoint 3.1 Use consistent presentation.

62. "Use consistent presentation"

Checkpoint 3.4 Supplement text with non-text content.

46. Illustrating text

Checkpoint 3.5 Annotate complex, abbreviated, or unfamiliar information with summaries and definitions.

39. Abbreviations

There used to be a note under "define key terms, abbreviations, acronyms, etc" that read:

only the first occurrence of an abbreviation or acronym occurring in a document need be expanded (@@some would argue against this. particularly those advocating needs of people with reading or cognitive disabilities). Expansion dictionaries, for instance in metadata, may be provided as an alternative to an expansion in the text of a document

However, what does "first" mean. If I can enter a document at any point due to a link to a subheading, then I will miss the "first" occurence of the abbreviation or acronym. At the 11 January 2001 teleconferencethere was discussion about pointing to external glossaries and ensuring that the appropriate context was made clear. For example a glossary could define several meanings for one abbreviation and the author has to make clear which one they intend to be used.

72. Annotating with summaries and definitions

The second success criterion of checkpoint 3.5 of the 14 August 2001 draft is table specific. can we generalize it? Is this an example rather than success criterion? It currently reads:

a summary is provided for

In general, these success criteria need work.

Checkpoint 4.1 Choose technologies that support the use of these guidelines.

43. Content modes for Checkpoint 4.1

73. Choosing technologies

The success criteria for checkpoint 4.1 in the 14 August 2001 draftneed work. Is this an exhaustive list? Is it in agreement with the most recent XMLGL?

Checkpoint 4.2 Use technologies according to specification.

47. Success criteria for protocols

As of the 14 August 2001 WD, checkpoint 4.2 contains the following note:

Please help us identify accessibility issues with transport protocols so that we can properly address them in this checkpoint. Our current thinking is that these issues might be server-side issues rather than issues with the content.

What are the accessibility issues with transport protocols such as HTTP? SOAP? XMLP?

Checkpoint 4.3 Design user interfaces compatible with assistive technology.

74. Requiring usability testing

Checkpoint 4.3 of the 14 August 2001 draft contains the following success criterion with the associated note:

4. the interface has been tested using a variety of assistive technologies and preferably real people with disabilities who use assistive technologies to determine that assistive technologies can access all information on the page or hidden within the page. Note to reviewers: there is active discussion on the requirement of user testing as success criterion.

Some have said that this is a method and not a success criterion.

Checkpoint 4.4 Ensure that content remains usable when technologies that modify default user agent processing or behavior are turned off or not supported.

64. Use of the word "default"

In the 14 August 2001 draft, Checkpoint 4.4 reads: Ensure that content remains usable when technologies that modify default user agent processing or behavior are turned off or not supported.

Cynthia cautioned against using "default presentation". However, the following article supports this use of default. From XML content to HTML display. We should also be in agreement with UAAG.