Trivial to guess CAPTCHA text automatically

The captchas generated by this module are trivial to guess automatically using simple image processing and widely-available OCR libraries. They would not be effective at stopping targeted spam bots.

I've attached an incredibly naive Python script that gets the captcha right around ~50% of the time (I'm sure you could achieve higher accuracy with just a little more effort).

It simply splits the generated captchas into 5 characters using white-space, then runs the open-source OCR engine tesseract on each character.