On Tue, Jan 20, 2004 at 02:44:32PM -0600, Terry Shows wrote: > Maybe it is good for -16, but in every case I looked at that passed > thought with habeas set, none of them set the violator, and every > single one was flagrantly spam.
+16 not -16 just so people reading the archives don't get the wrong idea. I know you knew what you meant. > I have several email servers I manage, and will set some to zero and leave > some alone and see how things change. I have a few myself. I use the network checks because they take the accuracy from about 80% to about 97%. My local changes and the efforts of Chris Santerre and Jennifer Wheeler help to whitle down what is left. My complaint regarding the overreaction of people here has to do degree. I have personally lowered the HABEAS_SWE setting. Lowering the HABEAS_SWE to +/-0.001 is, effectively, the same as turning it off. No? (Well, unless you want to save the CPU cycles.) I left it - in the spirit of the intent of the rule. The other part of the overreaction is people's general expectation that a generic solution should be perfect in every situation. You, this is the third part non-specific you, can't expect anyone other than yourself to make the proper choices for everything in your situation. You will always have to tweak things to match your need. So tweak it and quit belly-aching to the list with things like "How could the SpamAssassin developers have missed the fact that this rule would be exploited when they looked into their crystal ball?" > these folks had a great idea, but fell short. To have made it work > right, habeas.com should have set up a server that, when queried with > the right authority, released a code that could be included in the > outgoing email. It should be random code, that can only be recognized > by the habeas server when queried by the receiving server. There is no flawless design for this kind of thing. If there were, we wouldn't have a spam problem. Habeas has theory on one way to curb spam. Rhyolite has another; Vipul another. Bayes is another. Are any of the perfect? No. Are they making a good faith effort? In my opinion, yes. It is the combination of these efforts that makes SpamAssassin work so well. SpamAssassin is still learning from its experience. It's easy to bash something when after it's been shown to have weaknesses. Your idea turns the HABEAS_SWE test into a network test. That may not be a bad thing, but it's a consideration which seems to be relevant to you individualy from your other message. > The way it is now, it is just another header that can be added by > a spammer, and as long as nobody turns him in to habeas.net, he is > guaranteed an easier path for his junk. (if I am missing something, > please feel free to educate me. Just be nice when you do it) My comments were never specifically directed at you. Yours was just the straw that broke my will to resist jumping in on this bikeshed discussion. Also, my tone was not meant to be inflamatory or mean. E-mail doesn't have sufficient bandwidth to carry the full nuances of the speakers intent. I try to assume the speaker was trying to be helpful until the ALL CAPS and %#&&$^ come out to play. Sometimes I'm wrong but I don't get ulcers worrying about it. Bikeshed -- See : http://www.freebsd.org/doc/en_US.ISO8859-1/books/faq/misc.html#BIKESHED-PAINTING -- Scott Lambert KC5MLE Unix SysAdmin [EMAIL PROTECTED] ------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
