WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: NoCAPTCHA reCAPTCHA accessibility testing updates?


From: deborah.kaplan@suberic.net
Date: Aug 27, 2015 1:30PM

Google's NoCAPTCHA is basically, for all practical purposes, inaccessible.

* Many times when I have tested by screen reader, and about 95% of the time when I test with voice or keyboard, it triggers the image selection alternative. (This is not what Derek Featherstone's tests found, but this is what I have been finding more and more lately; I have been assuming that Google has been fine-tuning the robot detector and now keyboard/voice use trigger it. There could be something else going on, but it is definitely a change for the worse, across multiple platforms.)

* The image selection alternative is not keyboard or voice navigable. At all.

* The audio captcha is just as bad as any audio captchas ever are. I've never been able to solve it, anyway.

For what it's worth, this is actively taking away my ability to use important parts of the web. For example, something about my search engine frequently triggers Stack Overflow's "verify you are not a robot" page, and that uses NoCAPTCHA. Since I literally cannot use NoCAPTCHA by voice or keyboard, I have had real trouble getting to Stack Overflow since they made the switch -- which as you can imagine is actually making me worse at my job.

> It would read it if I dragged my finger over it, but one would have to know that something is there for it to be effective. Is this issue a result of Google's use of Javascript for NoCAPTCHA?

JavaScript can be perfectly accessible. This is a result of the NoCAPTCHA team not writing accessible JavaScript.

> I've read both positive and critical reviews of the NoCAPTCHA widget online. What's the general consensus on using this CAPTCHA solution as an accessible option?


File bugs with Google, but for now, don't use it. Until they fix this, it is a big flight of stairs into way too many parts of the web.


Deborah Kaplan