reCAPTCHA it!

Gluff, the DHCP lease logger, has attracted some unexpected attention by anonymous commenters and since I never thought anyone would find it unless I advertised it, I never bothered to set up the commenting system properly. The results are that I don't know who any of these people are so I can't tell them that a new release of Gluff is now out, with a patch for ISC DHCP 4.1.0, and also that the page has been overflowing with ridiculous SPAM comments in a neverending stream.

I hope things will get better since I have now added the option of posting contact info with comments, as well as added a captcha for the comment form. Sorry about that, folks, but it really is the easiest and cleverest way to get rid of form spammers...

While looking at the various captcha modules for Drupal, I found the reCAPTCHA module, and went on to the reCAPTCHA web site to read up on it. And this is truly clever stuff, I have to say!

reCAPTCHA is currently helping The Internet Archive and The New York Times to scan books and old newspapers. What they do is they get hold of especially tricky words from the OCR system, pair each with a word that has already been parsed with confidence, and then use the two words together for a captcha. If the user gives the correct answer for the known word, they just assume that the other is correct too, and store the interpretation with a confidence value. Then they do it several more times and increase the confidence value accordingly.

In other words, they are doing what Folding@home and SETI@home, but using people's brains instead of their computers - the world's first large-scale human cluster?

Now all we have to do is wait in horror for the organic botnets that may follow... :-)