The purpose of this page is to capture recent discussion about the relationship between accessibility and validity, to give a brief history of how we got to where we are today, and to outline the variety of issues related to the discussion in hopes of furthering the discussion and finding compromise and consensus.
Prepared for the WCAG WG as a companion to the 30 June 2005 WCAG 2.0 Working Draft.
If commenting on this issue in non-W3C space (other mailing list, blogs, etc), please link to this summary.
People agree that validity is a good first step towards accessibility and that validity does not guarantee accessibility.
It seems that the variety of concerns raised about the issue boil down to two primary perspectives on how validity should be prioritized in WCAG 2.0:
Since we are all trying to make the Web accessible, what is the middle ground?
Since the earliest drafts of WCAG 2.0 (as early as 2002), there has been a success criterion that requires validation but allows exceptions for backwards compatibility. More recently, the WCAG WG has discussed questions about the exception such as, "do we need an exception?" and "should we broaden the exception to invalid content that is invalid because of a new accessibility feature?"
Guideline 4.1 Level 2 SC1: "Except where the author has documented that a specification was violated for user agent compatibility (including compatibility with assistive technology), the content has:
There was consensus of those attending the face-to-face to adopt the proposal with an action item for a sub-group to investigate a level 1 criterion to address well-formedness - although HTML is a SGML application and SGML does not have the concept of well-formedness, it does seem like a reasonable approach to say "make sure the DOM can be consistently formed" and refer to well-formedness in XML. The sub-group should also investigate PDF, Flash, and other formats and propose a parallel criterion.
Not all arguments summarized here, but tried to get those that seemed to appear most frequently.
Many organizations have been promoting W3C specifications and adherence to those specs. In particular, groups such as the Web Standards Project are looking to make things easier for authors so that content can be written to specification rather than being tweaked for each browser. There is a sense that if WCAG 2.0 does not include validation to specifications at Level 1, that we are taking a step backwards and opening up the possibility of "tag soup" which everyone wants to leave to the late 1990s.
Some argue that WCAG 2.0 is a technical specification and no more. It is produced by the W3C who is not a policy body but a technical standards body, therefore we should not be concerned about policy or legal issues. On the other hand, WCAG 1.0 has been referenced or adopted as policy in many countries and organizations. Several of those policies say that when WCAG 2.0 is available, WCAG 2.0 becomes the effective policy.
The mission of the WAI is to support the W3C commitment "to lead the Web to its full potential [including] promoting a high degree of usability for people with disabilities." The WAI accomplishes this through a variety of activities. The WAI Technical Activity, "promotes implementation of accessibility improvements in Web technologies through development of a set of three WAI guidelines as Recommendations: Web Content Accessibility Guidelines; User Agent Accessibility Guidelines; and Authoring Tool Accessibility Guidelines. These guidelines describe accessibility features needed to achieve different levels of accessibility, and include reference checklists as well as implementation techniques."
Therefore, some people feel that the purpose of WCAG 2.0 is to categorize those things at Level 1 that are essential to accessibility and Level 2 is more broad. Some have argued that validation is essential for accessibility others have argued that it is not.
Looking at W3C/WAI precedent: WCAG 1.0 Checkpoint 3.2 "Create documents that validate to published formal grammars" is a Priority 2. User Agent Accessibility Guidelines 1.0, published in December 2002, says it is a Priority 1 to implement accessibility features of a specification (8.1) and a Priority 2 to conform otherwise (8.2). The UAWG's rationale was that accessibility is the primary goal. Conformance should help achieve accessibility, but that conformance is not an absolute prerequisite to accessibility.
8.1 Implement accessibility features (P1)
Implement the accessibility features of specifications (e.g., markup languages, style sheet languages, metadata languages, and graphics formats).8.2 Conform to specifications (P2)
Use and conform to either
- W3C Recommendations when they are available and appropriate for a task, or
- non-W3C specifications that enable the creation of content that conforms at level A or better to the Web Content Accessibility Guidelines 1.0 [WCAG10].
Some have argued that in 1999 it made sense for validity to be Priority 2, but looking at the progress made in adopting standards and looking forward five years it's time to move validity to Level 1.
The World Wide Web Consortium (W3C) develops interoperable technologies (specifications, guidelines, software, and tools) to lead the Web to its full potential.
from W3C Home page
Some people feel that for WCAG 2.0 to be "interoperable" with other W3C specifications WCAG 2.0 must make validation a Level 1 requirement for conformance. One of the advancements in XML is that once a fatal error is detected, "the processor MUST NOT continue normal processing" meaning that some documents or data may not be displayed. Fatal errors include "violations of well-formedness constraints." Does this then mean that WCAG 2.0 should reinforce or enforce that, and if we do not, are we undermining other specifications? Is this then the mission of every other W3C specification as well? Do these issues affect people with disabilities any differently than any other users? In other words, is validity essential for accessibility?
Examples have been given to support both sides of the argument: there are examples of invalid content that is accessible as well as valid content that is inaccessible.
In response to the proposal to move validity to Level 2, some of the comments made include:
Some have argued that producing valid code can be difficult: with legacy CMS, content contributed by many users, aggregation sites, or sites with a mass of legacy content. Others argue that to fix the issue, ATAG and WCAG both need to require validity to push developers to fix the situation. However, is that the role of WCAG? This relates to the question, "is validity essential for accessibility?" and WAI's role within W3C.
Well-formedness is a new concept introduced by [XML]. Essentially this means that all elements must either have closing tags or be written in a special form (as described below), and that all the elements must nest properly.
Although overlapping is illegal in SGML, it is widely tolerated in existing browsers.
The underlying issue of this whole discussion is determining what is essential for assistive technologies. We all agree that screen readers need to correctly parse the document tree. More research is needed for other types of assistive technologies such as speech input, magnification software, and scanning software.
At the face-to-face, the WCAG WG thought we might be able to put well-formedness (or some assertion of correct parsing of an application/delivery unit) at Level 1 and validity at Level 2. For XHTML and other XML applications, this is possible. However, well-formedness does not apply to non-XML applications including binary formats (Flash and PDF) nor to SGML applications such as HTML. In both cases, validity is a reasonable concept; "valid SWF file" or "a file that plays without error" are reasonable tests of valid Flash files. Some argue that the only test for HTML is validity, since parsing and extracting information relies on validity. Others argue that instead of using invalid HTML, authors who wish to extend XHTML with new accessibility features should move to XHTML or modularity of XHTML to define new attributes. However, the majority of Web content today is HTML and an HTML solution is needed today. Another consideration is that there is a difference between conformance and validity. "Normative Definition of XHTML 1.0" particularly "Document Conformance" and "Element Prohibitions" is another set of constraints that are not defined in the DTD.
One proposed compromise is to put some types of validity errors at level 1 and others at level 2. Others feel that other WCAG 2.0 success criteria may address particular validation issues or we could create new ones to fill in the gap. The concern with that approach is, "what if we miss one?" and that we would be too HTML-specific. The XHTML specification (quote above) says "overlapping is illegal in SGML" that seems to be a precursor to "well-formedness" - could that be addressed at Level 1? Another suggestion is to acknowledge the importance of validity through other means such as a white paper from W3C on the value of validity. People could point to that as a clear statement from W3C and remove the burden from WCAG 2.0 to promote validity.
From WaSP:
In the end, it's about balancing flexibility and freedom with order and control. I don't yet know where the right balance lies with respect to validity and accessibility, but I do know I'm glad we're having this debate.
If commenting on this issue in non-W3C space (other mailing list, blogs, etc), please link to this summary.
The WCAG WG published a Working Draft of WCAG 2.0 on 30 June 2005. Please send comments by 2 August 2005 to public-comments-wcag20@w3.org. The WCAG WG will review comments and attempt to address them in the next Working Draft (tentatively scheduled for publication in September 2005). We will continue to listen to comments about this issue and engage in conversation to find a way forward that has consensus among the WCAG WG and the Web accessibility community.
$Date: 2005/07/01 22:38:12 $ Wendy Chisholm. Prepared for the WCAG WG as a companion to the 30 June 2005 WCAG 2.0 Working Draft.