Wednesday, 23 April 2008

ReCAPTCHA

Every so often while surfing the web I'll run across something that makes me think "what a great idea". ReCAPTCHA is one of those things.

The aim of a CAPTCHA (yes, its an acronym) is to differentiate between humans and automated computer programs such as spam robots. You'll meet one if you try and comment anonymously on this blog and lots of others; they're on sign-up forms for forums; basically anywhere where spam is a problem.

The great thing about ReCAPTCHA is that it not only performs this function (of a spam prevention system), it is also a proof-reading service. It supplies its users with words from books that its creators are trying to digitise but the OCR system was unable to read. The humans who complete the CAPTCHA form do the digitiser's work for them.

Of course, you might be wondering at this stage as to how the ReCAPTCHA system knows what the right answer is. Its creators got around this problem by serving users with two words - one of which is known, the other is not. The known one is the actual CAPTCHA of course, however the system does not indicate which is which, reducing the likelihood of users deliberately giving the wrong word. All in all, this is a really impressive, innovative approach to a problem.

No comments: