I tend to only see them on a wiki I frequent, and occasionally edit. I can give you a masterclass on some aspects of (my instance of) their expectations. Except whether "Traffic Lights" should be just the lights, the entire housing or also the supporting poles/cables.
My 'favourite' is the (usually, and certainly no less than) three tiles in the grid that matches the desired description. If you're asked for Tractors and you've selected the actual two tractors, you do have to decide whether to also click the non-Tractor (e.g. a road-grader) that the system clearly thinks you should think is also a tractor.
The one with individual images where clicked things get replaced by further tiles (which may or may not also be worthy of clicking) quickly devolve into keeping an eye on the tile you just tapped and seeing if you need to tap it again, what with those that never were then not being made to change.
But Traffic Lights (or perhaps "Motorcyles" - which are often mopeds, really) in a full-CAPTCHA scene as a multi-tile single source image can be more troublesome, IME.
(I do occasionally get an "I am not a robot" tickbox on a Google Search, apparently my IP (i.e via my ISP's chosen shared gateway, I presume) has been over-used for suspicious searches. But never more complicated than that, plus whatever residual meta-activity data it harvests to try and determine I'm a person browsing.)
Author
Topic: The Generic Computer Advice Thread (Read 574554 times)