Maybe it's just me, but I'm finding the implementation of "Captcha" on some sites to be increasingly painful.
In case you don't know, captchas are those little images you run in to that contain skewed text, the idea being that an automated system won't be able to identify all the characters correctly. In other words, you have to be a human to get the text right, so captchas help thwart spam-bots.
Here's one I hit the other day:
Look at this image. Sure, you can see GA92SU, but there's a "K" and an "r" and an "I" or two. The small grey text indicating that you only need to find 6 characters hardly leaps out at you.
Bah humbug. Grumble grumble.
I think Luis von Ahn (the inventor of Captcha) is a genius. Luis also invented the ESP Game and gives great Google Talks.
As an aside, he recently launched "Re-Captcha" which combines a traditional captcha with an unrecognized scanned word from a digital archive (currently they are working with the Internet Archive). When you complete a re-captcha, you are helping digitize previously un-readable fragments of documents. Current estimates are that re-captchas are digitizing over a MILLION words per day.
Whenever I see captcha examples Luis has created, they are eminently readable. I just wish folks wouldn't feel the need to "improve" on his system. It works. Leave it alone!
These days I tend to prefer math captchas just because you don't have problems like this. There are other problems (or as Al Bartlett ( http://www.hubbertpeak.com/bartlett/bio.htm ) says, "the most common source of problems is solutions!"), but I personally like the trade-off they offer.
Posted by: Tanner Lovelace | October 22, 2007 at 02:38 PM
kittenauth - started as a joke, seems a good idea
Posted by: Al | October 22, 2007 at 07:00 PM