Thursday, February 07, 2013

My web accessibility ReadMe file

My "start here" list for web accessibility is a short list, but the drilldown is deep.
I'm a big believer in testing everything. If you don't have a test that shows that it works, then you must assume it does not work. Even for stuff that Steve says works, or I do. Prove me wrong, make the Internet better.
I typically test on:
  • Keyboard-only in one Windows browser.
  • JAWS and NVDA
  • IE8 and latest Firefox. I call this my "bookend strategy." If something works on both a modern standards-compliant browser and on an older but still widely used browser, then I stand a good chance of working on the stuff in the middle.
  • Ignore Colors in one Windows browser. "Ignore Colors" is the IE term, but Firefox has pretty much the same thing.
  • Browser zoom to 200%
  • I should test more on VoiceOver, but I don't have good access to that right now.
Whenever I can, I test on more combinations. But I would usually only have time to go further if I were proving out a piece of code that is going to be widely deployed, without a lot of chance to go back and fix it. Again, with this list tell me where I'm wrong! There's always room for improving or complementing a solution.

I actually believe there is not just one single ideal list of target client technologies. But I do believe each developer needs to make a clear, rational decision about whom they're developing for. I start by trying to understand the range of disabilities my customers could have, and how those disabilities would affect their choice of technologies. Then I gather whatever data I can on disabilities and assistive technologies (ATs). Here are a couple of sources - I might or might not agree with their conclusions, but they give me their data and methodologies, so I can make my own judgments.
Finally, I devise a "graded support" approach for how to deal with the world of users and ATs, inspired by Yahoo's graded browser support model. Unlike Yahoo's model, targeting is not usually possible, since there is no reliable sniffing for ATs.

No comments: