I think it's getting to the point where it affects the availability of a company's mail system. It also makes any sort of backup and recovery plan harder. I've just installed bogofilter and used spamassassin to teach it, along with an archive of 30,000 pieces of spam. I know bogofilter does analysis on html tags, because a red font in the html is considered a better gauge of spamicity than most words. Cheers --Mike ----- Original Message ----- From: "Crispin Cowan" <crispin@private> To: "Shaun Savage" <savages@private> Cc: "CRIME" <crime@private> Sent: Monday, June 02, 2003 11:35 AM Subject: Re: CRIME Interesting way around spam filter > Shaun Savage wrote: > > > I don't look at mush spam, but this caught my eye. Not the spam but > > the why it is hidden inside HTML. > > Yeah, they've been doing that for a couple of months now. It is clearly > aimed at frustrating word-based spam recognition. > > In a similar time span, my Mozilla Bayesian spam filter's effectiveness > has dropped like a rock, now recognizing only about 50% of spam. > > Question: is it the case that Mozilla Bayesian spam filtering is done > only on the raw message text? Or do they do word analysis on rendered > HTML as well? For that matter, where is the documentation on what it is > analyzing? I can't find it. > > More broadly: does the community believe that spam has become so bad > that it can be considered a security problem? > > Thanks, > Crispin
This archive was generated by hypermail 2b30 : Mon Jun 02 2003 - 15:52:59 PDT