Dnia 7.10.2019 o godz. 10:19:57 Brielle via mailop pisze: > > Hate to break it to you, but its always been "my network, my rules". > When people started being shitty neighbors, those people who were > once understanding started protecting themselves better. > > I've been a network admin my whole professional life. If you don't > play by my rules, you don't get to talk to my equipment.
Me too, not only a long time admin but also developer of some anti-spam solutions myself. Some of them I still use on my server :) But deliverability was always my priority and was priority of all people I cooperated with. If you had to shut off a spammer, you did shut off a spammer, and not the whole netblock. Protecting from someone nasty doesn't mean harming the innocents. There's no such thing as "collateral damage" in network administration. At least it shouldn't be. I was always taught that the main goal in spam filtering is of course filter as much spam as possible, *but* at the same time keeping the number of the errors of second kind (ie. non-spam message being mis-classified as spam) at "virtual zero" level. Errors of second kind are much bigger problem than errors of first kind (ie. not catching an actual spam), because if user accidentally gets one or two spam messages in his/her inbox, he/she can easily delete them; but if a legitimate message is accidentally put into spam folder, the user will probably never see it, as most users don't look into their spam folders at all. And if the user doesn't look there and doesn't correct the mis-classification, the algorithm will still "think" the message is really a spam. Thus, next messages from the same sender have higher chance of being wrongly marked as spam as well. This, as a final effect, makes e-mail unreliable as a means of communications. You could never know if the message you sent reaches the recipient or not. And when you learn that it hasn't it may be too late, because eg. the event you were asking about has already happened. Also, if we look at this closer, the very meaning of the word "spam" refers to *content* of the message, not to the way the sending mailserver is configured or where it is located. It doesn't matter where the message is coming from and how it was sent, the content - and content only - decides if it is spam or not spam. So, if an email that has perfectly legitimate content and not any signs of spam, ends up in spam folder because of the "reputation" of *a whole IP block* (and not even that particular IP address), it means that the configuration of the spam filter on the receiving end is obviously *wrong*. Of course, the spam filters can - and do - use things like "IP reputation" as *hints* if the message can potentially be spam, but they should be hints only - in the end the content analysis should have the highest weight in spam-score for the message. If the spam filters pay more attention to "IP reputation" than to the content, then something's certainly wrong with those spam filters (of course, I exclude here the spam-dedicated mailservers, that send virtually no legitimate messages, like those listed on Spamhaus' SBL, but they should be *confirmed* as such before we start treating mail from them automatically as spam). Don't you agree? -- Regards, Jaroslaw Rafa r...@rafa.eu.org -- "In a million years, when kids go to school, they're gonna know: once there was a Hushpuppy, and she lived with her daddy in the Bathtub." _______________________________________________ mailop mailing list mailop@mailop.org https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
