I'm sure you've seen those "capcha" images on various sites -- you know, those twisted, distorted images of a few letters and numbers which you have to try to read and input to prove you're human. These are a big PITA, but they're one of the easier way to defeat spammers, yet some still get past. This shows how.

I generally hate the capchas because some are nearly impossible to read, yet I also realize that they're (sadly) almost required to prevent blog spam. Yet, some spammers can clearly bypass these tests, thus allowing them to use programs which post hundreds of spam messages to hundreds of blogs per day.

Anyway, I've always suspected that they had a programmatic method to decode the capcha images, but since I'm not a spammer myself, I've never actually been in the right place to actually see what these tools look like. Well, tonight, I have seen one example of a tool which can be used to pass the capcha check.

One programmer decided, for various reasons, to write a .NET program using Visual Studio which would grab a capcha image from a blog and then decode it into the required text, thus emulating the functionality of being human, but with the efficiency of a program. I just thought it was kinda neat to see the process and image manipulation that was needed to decipher these particular images and thought some of my readers might enjoy it also.

Check-out the full article on breaking the capcha if you're interested.