Yesterday, the web magazine A List Apart posted an article from Gian Sampson-Wild, “Testability Costs Too Much,” where she makes the claim that the requirement of having every success criteria within the W3C’s Web Content Accessibility Guidelines 2.0 be testable is too steep of a requirement.
I completely disagree. Success criteria that can’t be tested shouldn’t be included in a document that is supposed to give guidance.
I was going to use the metaphor that Jeffrey Zeldman jumped on. If someone says don’t speed in your car because it will hurt people, thats fine but how do I know what speeding is. It’s a toothless and unenforceable law. But if you say that I can’t go above 65 mph or 100 kph, that is a testable and enforceable law.
I can’t tell a developer to do something unless I know specifically what I’m asking of them. Just giving some one general advice isn’t going to work. It is going to be interpreted a variety of ways. This leads to fragmentation of guidance and inconsistent implementations which don’t help anyone.
If the principle of testability of the success critieria is inconsistently applied within the document, I think thats a legitimate concern. Commenting on the latest WCAG 2.0 Working Draft closes on Friday.
Doesn’t taking out testability dilute the guidance that we’re really want and asked for? Am I missing something?