What's the desired behaviour? Audio file to read the captcha?

Well, this would surely improve the experience for screen reader users.
I don't know if there are more subtle/sofisticated way to do this but I think an audio alternative would be quite good.

The important part is that it cannot be readable by a computer. :) It was to be visual (picture) and/or audio.

yes, you are of course right!
By the way, some of our other users, instead, have noticed that the captchas are not so challeging and a bit "recurring"... many people, many ideas...

Sure, but a simple captcha is usually enough, until someone decides to target your specific application for hacking. Then it gets hard.

https://forum.yiiframework.com/t/remove-ccaptcha-from-the-core/46332

Else in french : https://a11y-guidelines.orange.com/fr/articles/les-captchas-et-l-accessibilite/#honeypot-et-time-measuring-deux-techniques-simples-a-mettre-en-place-pour-identifier-les-bots

Captcha is good only for bots .
Then with https://bugs.limesurvey.org/view.php?id=17322

My opinion : a settings of 1 time with 1 second is good enough for bots.

I think I'd categorize this as a new feature more than a bug, though.

Good link, Denis.

@ritapas Did you consider just to disable captcha for accessibility?

@ollehar I see your point.
Unfortunately we have a LS CE installation on premises and we were requested by the susdamins not to leave "open" surveys.
Of course, if this is a completely new feature to add, and not a bug, we'll try to have them rethink about it.

@ritapas : in 3.7.0 : there are a protection from bot's (protection by IP).

You can use participants? Then it's not open, but also not using captcha.

Oups … right only with token/participant surveys

@ollehar sometimes we need our participants to self-register themselves

Alright. :)

@denischenu it's not easy for us to switch to a new version and make real tests with a look to the webserver access log... I'll wait until we will

where can I find a list of all pending feature requests? Just to avoid duplicates or wrong categorisations

nevermind, I've found it!

Since this is not a bug, I will close the report