(Please read through the whole message before flaming.) :) On Thu, 24 Oct 2002, 4:50pm (-0700), John J. Stimson-III wrote:
> > Why do I do it this way? Two reasons. One, it works. Two, it's a _lot_ > > less work for me. Much better to revoke a false positive every couple of > > weeks or month, and manually report a false negative once a week or so, > > than to have to manually report 20-30 spams a day, and manually revoke a > > false positive every day or two. > > That's great. Laziness on your part makes more work for everyone else > who uses razor. "A truly great computer programmer is lazy, impatient and full of hubris" -- Larry Wall :) > Actually, I don't understand why, if your automated mechanisms have > already caught the spam, you feel the need to report them to Razor. > Is it to help *me*, a third party who is not running the same > automated filters as you? Please don't! I have the same access to > those other filter programs, and if I want to run them, I will. In > addition, I would want to configure them to my own preferences for > false negatives/positives, not yours. Razor is one of many methods But what if, hypothetically, my filters are never wrong? Hypothetically, I've saved you the need to use your CPU cycles to run the filter yourself. So it is to help; not only you, but every user of razor, including those who are not experienced, smart, or savvy enough to know or install those other filter programs. > for eliminating spam. Please do not contaminate the Razor database > with the output of other methods, thereby inflicting upon me *your* > preferences for spam detection. Somebody else has already pointed out that by using Razor, you're already allowing others to inflict their preferences on you. I can't really improve upon that point. Just because it's a human making the decision rather than an algorithm, doesn't necessarily make it any more reliable for you. > If you want to create a system whereby many automated anti-spam > systems can share data to catch more spam, feel free to do so. > However, please don't kluge one together by assimilating Razor into > your hodgepodge. If you're going to use Razor at all, please maintain > a separate database. "The avalanche has already started. It is too late for the pebbles to vote." -- JMS, via Kosh > Thinking back to the "laziness" excuse, if you want to bulk-report the > messages that have been caught by your automated systems, why can't > you do so by inspecting the spam-mailbox for false positives, deleting > them, then dumping the whole mailbox into razor-report? That would > seem to require *less* work than automatically reporting messages as > they are flagged, then inspecting the spam-mailbox and individually > revoking the false positives. IMHO, it would still be more work for me. As I see it, either way, I have to inspect every message. My way, if there are no false positives, that's it, I'm done. And false positives are so rare (for me) that the couple of minutes I need to take for each one are almost negligable. Now, your way, I still have to inspect every message, but I also have to do the work to report once for every arbitrary unit of time. If I make that unit of time too short (say, an hour, or even every time a new spam message comes in), then I'm going to be spending a lot of time reporting spam. If I make that unit of time too long (say, a week), then most of the data I'm reporting is probably going to be too old to be any good, plus it's going to take significantly more time to do the actual verification. > Regardless of any philosophical arguments, the documentation says not > to do it. Therefore, you are misusing the system, which is a resource > shared by many, many others. Please don't. RTFM. The whole _point_ of my philosophical argument is that users don't read documentation. And many of those who do are going to try to force the software to do what they want anyway. You can yell and scream as much as you want in the docs, in the FAQ, and on mailing lists, but most users are not going to see it (and I bet there are a _lot_ of lurkers on this thread right now thinking, "to hell with this guy, I'm going to autoreport anyway.") The plain and simple fact is that if the _software_ doesn't stop the user from doing something, they're probably going to do it. And the funniest part about this whole argument is that the trust system is essentially _already_ stopping the user from doing the "wrong thing". Ironically, despite all these points that I have just made, I'm converted. As soon as I get the chance, I'm going to turn off autoreporting and dump all "candidates" into a mailbox that I can review and report once or twice a day. Now I just have to figure out a method so that I don't see that mailbox every time I get "candidates", but so that I don't neglect to check it for weeks at a time. (I use pine.) :) -- Edward Hennis ___ [EMAIL PROTECTED] ___ http://www.vaxer.net/~eah |\ /| | O | There are two ways to write error-free programs. |/_\| Only the third one works. ------------------------------------------------------- This sf.net email is sponsored by: Influence the future of Java(TM) technology. Join the Java Community Process(SM) (JCP(SM)) program now. http://ads.sourceforge.net/cgi-bin/redirect.pl?sunm0004en _______________________________________________ Razor-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/razor-users
