1 reply to this topic

[Darius]

Just recently came across this and thought it was a pretty cool project along the same lines as distributed computing, but human-based.

ReCaptcha is an update to the CAPTCHA tool for websites. Instead of a mixed up sequence of letters/numbers it uses two words. One of the words is the true "security" word. The other is a word from digitized archive projects that the OCR could not recognize properly. These projects have to spend individual man-hours going through these unrecognized words and manually correct them.

ReCAPTCHA adds these words and Internet users will enter them. The information is sent back to the ReCAPTCHA folks and if three or more individuals provide the same result, they consider the word to be recognized. According to he site approximately 150,000 work hours are carried out per day with about 60 million words per day recognized.

ReCAPTCHA Web Site

[GroovyDude]

NPR's "All Things Considered" did a story on this last Friday. Supposedly they're burning through volumes of material that would otherwise take a dedicated teams a lifetime to go through. Very Cool, but I still hate Captcha.