From WCAG 1.0 to 2.0: Accessibility You Can Actually Test

The old web accessibility checklist was written in 1999, and it reads like it. Checkpoints about technologies the web has outgrown. Guidance assuming pages are documents rather than applications. Requirements phrased in ways two reasonable auditors could score differently. Accessible sites are not a legal expense. They are more audience for the same content.

For a decade, teams that genuinely wanted to do right by disabled users worked from a standard that made ‘done’ hard to define. Its successor, WCAG 2.0, a W3C Recommendation since December 2008, fixes exactly that. And 2010 is the year it stops being optional reading. Procurement language is beginning to cite it. The federal Section 508 refresh now underway is aligning with it. The legal system has started attaching dollar figures to the underlying obligation. Time for the practitioner’s translation.

The structural fix: principles over checkpoints

WCAG 2.0’s architecture is its improvement. Instead of technology-specific checkpoints, it defines four principles. Content must be perceivable, operable, understandable, and robust. The community says POUR. Under each sit specific, testable success criteria, deliberately written technology-neutral so they will survive the platform churn that dated their predecessor.

The translation of each principle, in working terms. Perceivable: can every user receive the content? Text alternatives for images, captions for media, color contrast sufficient for low vision, information never carried by color alone.

Operable: can every user work the interface? Everything reachable by keyboard, the assistive-technology common denominator. No traps, adequate time, navigation that can be bypassed and understood.

Understandable: does it behave predictably and read plainly? Consistent navigation, labeled forms with useful errors, language declared.

Robust: does the markup hold up under assistive technologies? For standards-based builders, that one is the freebie. Clean, semantic, standards-based structure is robustness, and the four-year argument for it just picked up more support.

‘Testable’ is the word doing the work. Criteria like contrast ratios have numbers. Keyboard operability has a pass-fail procedure. Two auditors now converge. That is what lets accessibility enter contracts, checklists, and definitions of done: the boring machinery by which values become practice.

Conformance levels, chosen sensibly

The standard defines three levels: A, AA, and AAA. The emerging consensus target is Level AA, for good reasons. A alone leaves gaps real users hit daily. Contrast, for one, sits at AA. Blanket AAA is impractical for most content and was never intended as a general requirement.

AA is where the Section 508 refresh is pointed, where the legal settlements are landing, and where our own project standards now sit. Treat A as the floor you never advertise. Treat AA as the commitment you write down. And treat specific AAA criteria as à la carte upgrades where your audience warrants, like higher contrast for aging donor bases. The standard is a menu, not a mountain.

The price signals, briefly and without scare tactics

The legal environment deserves plain statement. The Target settlement cost six million dollars to the class, plus fees and years of oversight, for an inaccessible retail site. It established that ADA claims against commercial websites can be expensive to defend and more expensive to lose. Universities operate under Section 504 and 508 obligations that the refresh will sharpen.

None of this should be the leading reason to do the work. Organizations that treat accessibility purely as legal cover reliably do it badly. But budget conversations run on price signals, and these are the current ones.

The fuller accounting: somewhere north of one in ten of your visitors has a disability that affects their web use. The overlap between accessibility technique and the SEO, mobile, and quality engineering we teach is enormous. Crawlers and screen readers want the same semantic structure. One investment, several returns. And institutions whose missions include words like ‘access’ and ‘service’ are, we would gently note, describing this work.

The economics: build it in, or pay the retrofit

The number that should change behavior, from remediation projects we would rather not repeat: retrofitting accessibility costs multiples of building it in. Pick a small integer. Retrofit means re-touching every template, re-cutting every color decision, re-shooting every uncaptioned video, and re-testing everything. Under deadline, usually with counsel cc’d.

Built in at discovery, accessibility is mostly decisions rather than labor. Contrast chosen when the palette is chosen. Structure planned when templates are planned. Keyboard paths designed when interactions are designed. Alt-text fields required when the CMS is configured.

Our project process now places WCAG 2.0 AA at discovery. It goes in the design brief, the CMS spec, and the QA checklist. It also goes in the training for content editors, whose alt text and heading habits are half of ongoing conformance. The launch is only ever half.

The line we use in kickoffs, offered for reuse: accessibility is not a feature of the site. It is a property of every feature. Priced that way, it is cheap. Priced the other way, it is a settlement.

What the criteria look like on an actual page

Abstraction being accessibility’s enemy too, a worked pass over the page type every institution has. The program or service page with a lead image, an introduction, a registration form, and a promotional video.

Perceivable: the image gets alt text that conveys its function, or empty alt if decorative, a distinction your CMS training should teach. The video needs captions, which is the item most budgets meet reluctantly and most audiences use more than anyone predicts. Captions serve noisy offices and non-native speakers long before they serve deafness. The palette’s link color must clear the contrast ratio against its background, checkable in seconds with the free tools.

Operable: the form must be completable entirely by keyboard, with visible focus so a keyboard user can see where they are. That is one of the most common failures in the wild, and usually one CSS rule away from fixed.

Understandable: field labels attached programmatically, so clicking the label focuses the field. That is the test anyone can run. Error messages that say what to fix rather than announcing failure.

Robust: all of it in semantic markup, which, if your site was built to standards, it already is.

Total incremental effort on a new build: minutes per template plus caption budgets. That is the whole trick of discovery-phase accessibility. Itemized, it is a list of small decencies. Deferred, it is a project with a lawyer attached.

Starting from here, wherever here is

For existing sites, the path is triage, not paralysis. An honest AA audit: automated tools catch perhaps a third, the machine-checkable third, and human review with keyboard and screen reader covers the rest. Findings ranked by user impact and traffic. Template-level fixes first, since they repair every page at once. And a schedule that shows sustained progress. Perfection is not the standard anywhere. Demonstrated, documented diligence is: legally, ethically, and practically.

For teams starting their first project under 2.0: welcome to the version of this discipline that finally has edges. The standard is readable in an afternoon and the criteria are checkable. And the sites that result are, this is the part the compliance framing always undersells, simply better sites. Clearer, faster, more usable by everyone, on more devices. That should sound familiar by now. Good structure keeps being the gift that gives to whoever shows up. That was always the point.