I have recently been hit with a whole bunch of online poker comment SPAM. My SPAM filters have been happily blocking it however. While doing some lookups on the hosts sending the information I was worried to see how many appeared to be trojaned legitimate webservers.

One of the methods I use to prevent comment SPAM are Completely Automated Public Turing tests to tell Computers and Humans Apart (CAPTCHAs). These have long been in use and it seems some research is being done to try and defeat them (one, two, three)

Possible solutions to the problem involve using audio files, logic problems (e.g. 5 + 2 = ), and Anti-OCR fonts. So far I have found a simple word filter to be the most effective, particularly with SPAMmers using the same words all the time.