coding: November 2007 Archives

Like everyone else, I get a lot of spam. Google mail generally does a good job of filtering it out, but even so 2 or 3 items of spam get into my inbox each day, and on bad days I'll find 200 emails sat waiting for me.   I also keep getting grief about the amount of spam generated from websites I help manage, so when I heard about reCAPTCHA -  a system that's designed to reduce website spam and help digitise books at the same time - I was interested.

Since most spam is automated - spammers send millions of emails at the same time - a good strategy to avoid spam is to try to prove that the person sending it is not a computer.  CAPTCHAs (Completely Automated Public Turing Tests) are designed to do exactly this. You'll have seen CAPTCHAs all over the place - they're the warped, sometimes colourful text at the bottom of the page which you need to identify before you can sign up to the latest and greatest website or post comments on your favourite blog.  The idea is that the website offers you a word which is designed to be hard for a computer to read.  If you can see the word and type its letters into the box accurately, you're more likely to be a human than a computer.  Although spammers can occasionally beat CAPTCHAs (e.g. if the word is not warped or disguised in some way, a computer can use Optical Character Recognition to decode it), they're generally pretty effective at stopping spam.