Tinhat Level A+

Through reading the DRC report on Web sites and personally reviewing a small selection of UK sites, it's become apparent to me that we need a formal accessibility level pitched somewhere between WAI level A and WAI level AA. In practice, many sites are aiming for this middle ground, and the DRC research appears to show that the vast majority of accessibility problems can be removed from a site without reaching full level AA conformance.

I'm not the only person to think this, and the draft version of the new Web Content Accessibility Guidelines 2.0 contains a similar proposal, although it also contains many reservations about whether the proposal will be practicable.

For convenience, I'll call my new level Tinhat Level A+. And here is the specification:

Level A, plus the following level AA checkpoints:

3.2 Valid HTML (or XHTML)
2.2 Good foreground and background colour contrast (partly an AAA requirement)
7.3 No movement in pages
10.1 No pop-ups and other spawned windows
12.3 Large blocks of information always split into manageable parts
13.1 Clear identification of link targets

And additionally, outside WAI guidelines:

A formal (written) accessibility policy
No overlapping text when using magnification technology

I'd fully recommend this new accessibility level to any Web developer. It takes account of the major issues identified in the DRC report, and shows a thoughtful approach, so in theory it should allow a site to conform to the UK Disabilities Discrimination Act (DDA) without the expense of reaching WAI level AA (and all its coding complications of relative units versus absolute units - ouch!).

Whether this kind of level is legally acceptable in practice, we'll only know when there's some applicable case law. But for now it should certainly keep you off the list of possible DDA prosecutions.

For larger sites, I'd also recommend:

User testing with disabled users
Standard user testing to avoid confusing layout and navigation

It's difficult to make these items mandatory, because for small sites they will simply be too expensive. The testing could cost more than the rest of the development combined. On the other hand, user testing does give a good return on investment, especially for e-commerce sites.

As rough guideline, I'd suggest that if your site development budget is over £20,000, you should at least get some quotes for user testing from usability companies. If your budget is over £100,000, you should definitely be testing, and return on investment should cover the cost.

Any suggestions for improving this Tinhat Level A+, please mail me at the following address, which is shown as a graphic to discourage spam address harvesting:

This Tinhat page is valid XHTML to WAI Double-A standard