I’m sorry if this is just me, but last time I checked I had pretty good color vision, I wear glasses but otherwise do occasional advanced visual work, and so far I guess aging hasn’t had the biggest impact on my sight yet:
But yet, I needed around five tries to get through the leminal space account signup captcha. Not only can I barely make out anything in many of the images, some I didn’t even attempt to solve, but for those that I do it still failed me. I have a feeling it likes to use letters where the capitalization is ambiguous, but then requires strict capitalization, or letter o vs number 0 or something like that.
This seems like a barrier that might be a problem for some people, given I assume a huge amount of the population has significantly worse eyesight than me, even though mine isn’t perfect.
Or it’s just me! But I wanted to bring it up in case it’s not.
Yes, you’re probably right. The captcha difficulty was changed to ‘hard’ maybe a year ago due to a lot of spam bot activity. I have set it back to medium for now and will see how we fare. It should provide a lot more consistent contrast in the challenge, but might not help much with ambiguity of letters (the captcha library might not be the best implementation).
Perhaps some answer and response captcha as an additional test might work, if lemmy supports that? Something asking a community-specific question, like asking for the name of the meta community that’s probably easy to look up for a human that is actually interested in how this place works, but which a bot can’t easily get from a web search.
I don’t believe there’s any support for that currently. There’s just the one captcha system that’s tied in with the backend.