The next draft of the Web Content Accessibility Guidelines 2.0 probably will not require HTML content that conforms at Level A to be fully valid. The following is my rationale for agreeing to this.
(Disclaimer: I am not speaking for my present employer — neither W3C nor WAI. Nor was I present at the face-to-face meeting where this was decided. I have been a participant in good standing of the WCAG Working Group for almost five years.)
A little background: the W3C is a technical standards body. It is not a policy body. It can make validity a technical constraint — and did, with XHTML and the application/xhtml+xml MIME type. But it cannot make validity the law of the land.
The Web Accessibility Initiative, however, is not so cut and dried. WCAG, the crucible in question, is the basis for actual government policy in a large number of countries and smaller municipalities. It is also a measure of accessibility among developers who know nothing about accessibility other than it is required of them, as well as those who use it as a goodwill gesture toward users with disabilities. It’s all things to all people, and that’s the very problem we face.
Those who have seen fit to malign the W3C for stopping short of validity need to dial down the rhetoric a few notches. To hear some of the commentary, this is a travesty, the greatest step back for the Web since
blink. How distasteful, they say, for the WCAG Working Group to roll over for their corporate masters so shamelessly.
Would that it were that easy. Few, if any, of those who argued against validity as a mandatory element did so out of self-interest. Rather, the concern was that it limits the applicability of the spec across the range of content already available, and that jumping through this particular hoop is not strictly necessary. It’s not, as has been (ex)claimed, a vote for tag soup. It’s a pragmatic decision — and pragmatism is valued over idealism where policy is concerned.
This debate is happening on a fault line between technology and policy. This is not liberal vs. conservative, even in the Jon Postel sense. Those who are arguing for validity in this point seem not to realize that this is a vote for authoritarianism. What you do for users with disabilities, this argument goes, is not as important as passing this binary test. Who would be champing at the bit for their Java code to be evaluated by a government body? Then why is this desirable for HTML?
Think about the ramifications of making valid code a legal requirement. Say you’re a Web designer. You leave out a semicolon in a Â on your home page.
You move a fragment of code from a place where it is valid to where it is not.
You redesign your CMS templates. The CMS happily churns out whatever fragments are in its database, blithely unaware of whether it is valid in context.
You use a feature that requires an additional attribute not found in HTML 4.01, which improves the experience on some browsers and does no damage on others.
You use the default code to embed Flash content.
Do you really want to be liable for these errors, even if they don’t actually damage overall accessibility?
WCAG has to apply to all content found on the Web, constrain itself to accessibility flaws at the document level, and be compatible with policy implementations. We can’t specify something that’s overly broad, lest people get the wrong message (for example, that validity equals accessibility, full stop — still a popular myth). All we can do for this document is to ensure that its guidance is sufficient to increase accessibility without overreaching, and hope that it maintains hull integrity when organs of policy evaluate its feasibility.
I have said it a million times now, and I’ll say it a million more: accessibility is process, not product. If this were the Web Development Process Accessibility Guidelines, and validity were the question, I would sign off on it in a heartbeat. In fact, I requested the addition of validity to the Authoring Tool Accessibility Guidelines 2.0, which I co-edit. And that’s where it belongs: not at every author’s doorstep, but in the products they purchase and use.
As process, validity is a no-brainer. It’s one of the first things I teach. I am a staunch advocate of standards-based design. Accessibility in seven words: “Write valid code, and use metadata well.”
But process is much harder to legislate, and blanket guidelines are hard to keep clean. If it can’t be proven that validity itself improves the accessibility of all Web content, then the case for WCAG’s promotion as a policy vehicle is called into question. WAI is not a Trojan horse for other interests; it is focused on increasing access to the full breadth of the Web by users with disabilities, and nothing more.
So, then. That’s how a standardista comes out against validation. I’m not, as I have been accused of, playing devil’s advocate here. I genuinely believe the lowest conformance level of WCAG 2 is the wrong place to require validity.
Here’s the comment box. Flames against me will be deleted and users will be banned without warning. People on my buddy list are not immune. Play nice.