Re: Let's add an approved date field to Failures and Techniques

> On May 3, 2016, at 3:37 AM, Adam Solomon <adam.solomon2@gmail.com> wrote:
> 
> if we generalize the test procedures for failures we would avoid this pitfall
> for instance for alternative text for images, instead of saying or alt or aria-label or ...
> we could just generalize and say if there is no programmatic determinable alternative then it fails (and then link to the how to docs to reference possible techniques)
> enumerating all the possible techniques in the failure test was the main problem with failures


Correct.  But be careful that we don’t just echo what is already there in the success criteria. When we first started writing failures, we had to toss a couple because they essentially just said:

Failure of X.X by not  <copy of text in SC X.X>

 that is, they basically just repeated what the SC already said.


 however, using part of the text of the success criteria verbatim in a failure is one of the best ways to avoid accidentally broadening or narrowing a success criterion.


But yes —