Alright, I know the thread topic is not extremely clear but what it refers to is when a form on a webpage has a series of numbers in image form usually over a multicolored background. And to test for spam bots or prevent them, it asks the user to type in the numbers in image form. The theory being that a program is unable to find the letters in the image. A relatively easy task for a human, in theory. With me so far?
In theory at least. I must be a spam bot because I have trouble doing it. I have a problem in matching the image numbers and typing them in. I don’t know if it is dyslexia, or what, but I just changed my email on ebay and it needed me to do the number thing, I did it twice and was told I screwed up twice. I ask my roommate to do it for me and he gets it right, so I knew it wasn’t a server problem.
The main problem I’ve found with those things (and we have one on our website too) is that there are many letters that are ambiguous, and an extra dash or squiggle caused by the background can turn a D into an O or a Q, or turn a P into an R.
So I guess it depends on where you’re tripping up.
Aside from accessibility issues, there’s another big issue: they don’t work anymore. Programs have been written that will correctly guess these with a good bit of accuracy (well over 80% for the usual presentation of distorted text and background).
Aside from accessibility issues, there’s another big issue: they don’t work anymore. Programs have been written that will correctly guess these with a good bit of accuracy (well over 80% for the usual presentation of distorted text and background).