Dispatches from the war on spam

Those little images with words on them that you read and type in when you sign up for a hotmail account or buy tickets at TicketMaster™ are going to get trickier. According to a Scientific American article, the spam bots are learning to read the current images, so the image makers are fighting back. (via Sci-Tech Daily).

These image-reading tests are called CAPTCHAs, which stands for the mouthful of "Completely Automated Public Turing Test to Tell Computers and Humans Apart". According to captcha.net, researchers have achieved anywhere from 78% to 93% success rates in breaking the old images.

I dug around, and found some of the new CAPTCHAs over at PARC (no longer Xerox PARC). This online quiz will test your acumen. When I took it, some images with 2 or even 3 letters with a big fake ink blotch or blotches on top baffled me. This might not be a problem if they were real words, but they have switched to made-up but pronounceable words like "orderapy".

And, of course, any style of CAPTCHAs, old school or new, are completely inaccessible to blind or dyslexic people. (via Anitra Pavka).

Posted by Chad Lundgren on Saturday, November 22, 2003 (Link)