Sidney Markowitz wrote: SM> [moved from Razor-users mailing list]: SM> On Mon, 2002-04-15 at 11:36, Craig R Hughes wrote: SM> > Is the problem that you're seeing "real" email generating scores in SM> > excess of 30 in SA, or that people are ignoring the warning and SM> > reducing the threshold substantially below 30? SM> SM> Craig - SM> SM> As someone said later in the thread, is there ever any justification for SM> automatically reporting to Razor what anyone who chooses to run SA will SM> detect as spam anyway? If the people who run Razor servers do not want SM> such mail, perhaps there is not a good reason to have the autoreporting SM> feature at all. Or at least not until the Razor server code is released SM> and the feature can be combined with the ability to specify private SM> Razor server addresses to report to.
Yes, I think it's probably best to more or less disable it for now. I think I'll do it by making the default be non-submission, and removing the documentation sections on the autosubmission options. I tend to agree with Charlie and others on the razor list who think there should be only two sources for razor submissions: 1. Manually verified spam, and 2. Spamtraps. Auto submission of mixed inboxes seems of dubious value. SM> I do see how one might want the ability to autoreport to one's own local SM> Razor server and then use that to prefilter duplicate spam before SM> running it through the full SA rule set. I think even with ones own server, it's of dubious value. If you were running SA to id the spam in the first place on your local system, then why submit it to your personal razor DB? Just pass all mail through SA and it'll all get scored equally high. Passing mail through the full SA process doesn't generate that much more load than just razor (which currently has no razorc/razord type thing), is it? C _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
