You know what a captcha is, right? Sure you do. You know -- those annoying little "tests" on websites where you to type the wacky image text into a form box to prove that you're actually a human being and not a computer.*
Yeah, they bug the hell out of me too.
Well, the folks over at Carnegie Mellon University have done a very smart thing. See, they've tied the text used in those captcha images with a large database of digitized, but not transcribed, books. When a book is electronically scanned, the character recognition software is usually enough to transcribe 90% of what's there into the correct computer text. The other 10%, though, has to be done by humans, who are smart enough to parse the visual part of the text, even when it's slightly blurry, or has lines drawn through it, or whatever.
With the marriage of this slightly illegible text and the constant need for new captchas for website verifications, the Carnegie Mellon group have released "reCAPTCHA" as a web service so that websites can take advantage of the idea (for free!) and all of us poor human users who have to use captchas to log on or sign up will actually be contributing to the digitization of human knowledge via this effort.
Next time you're filling in one of those de-humanizing "human" tests on a website, make sure the site is using the reCAPTCHA system. If they are, you can feel good that once again, you're better than a machine.
* CAPTCHA is actually an acronym for "Completely Automatic Public Turing Test to Tell Computers and Humans Apart". See? The 3'oclock blog isn't just about fun, it's also about learning.