Saturday, November 10, 2007


I just heard of reCAPTCHA.  This is a very interesting spin on CAPTCHAs.  Accoring to their web site, 60 million captchas are solved every day.  reCAPTCHA uses part of the effort to "scan" books.

 A reCAPTCHA captcha contains a word that OCR didn't find a match when scanning books from the Internet Archive.  In addition to the word that is not recognized, a successfully scanned word is displayed.  The captcha is considered successfully entered if the known word is entered, and then the answer for the unknown word is added to a database. If enough people enter an unknown word with the same answer, then the probability of the answer being correct goes up.

There are several plugins for blogging systems, as well as a PHP api