I’ve started seeing a trend in spam over the last few months increasingly using samples of Linux/Open Source software documentation for the subject line / random text to confuse Bayesian filters. Mostly it doesn’t help them get past SpamAssassin, but they’re trying to be clever and increase their chances. It does seem kind of futile though – technical users generally aren’t stupid enough to respond to spam.

From a few spams in my spam box right now:

• “The smartmontools package supports a number of different operating systems.”
• “Please report unusual or incorrect behavior to the smartmontools-support mailing list.”
• “Windows uses the same code for all instances of the application.”
• ” The disadvantage of using a larger number is that redrawing can become slow.”
• “IIS Double-Byte Code attempt.”
• “Used as an absolute reference for timestamp conversion.”

Some of them have what appears to be randomly generated “sentences” like:

Product years industry experience but doesnt fit out. Outlook
powershell security malware user groups webcasts. Xp xbox news posting is!
Route use moved new refuses reactivate? Jamesone filed under comments, oneil
excellent, post. Whether purchased retail, by an. Period before falling reduced
mode seems useful. Caveat has different versions sometimes skus stop stock,
keeping. Pick deployment published friday pm jamesone filed under comments.

I’m not the only one noticing this either:
http://radar.oreilly.com/archives/2007/02/is_tim_oreilly.html