> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Vincent Zweije > Sent: Monday, October 03, 2005 9:40 AM > To: [email protected] > Subject: Re: [Spambayes] Inspecting images (was: SpamBayes to > HandleEmbeddedImages) > > > On Mon, Oct 03, 2005 at 09:31:35AM -0400, Jesse Pelton wrote: > > || Of course, whatever the approach, the images must be fetched, > || which often means letting the spammer know that you've received > || (and seemingly viewed) their message, which encourages more > || messages. If such a feature were to be added, it should be > || disabled by default. > > Bah. Good point, given that, as I suppose, most images are > not in the spam itself. I never bother to check, as I don't > view images in mutt.
But not retrieving the image may not do any harm. Recognizing spamlike patterns in an encoded image is a hard problem, while classifying the image's URL as spamlike is easy. (If the spammer wants you to see the image, he can't disguise the URL very much.) I'd guess that working with URLs only would give better results. Bob _______________________________________________ [email protected] http://mail.python.org/mailman/listinfo/spambayes Check the FAQ before asking: http://spambayes.sf.net/faq.html
