ietf-smtp
[Top] [All Lists]

Re: [ietf-smtp] A Zero Spam Mail System - What works?

2019-09-29 21:55:46
On 9/29/2019 7:59 PM, Brandon Long wrote:

As for size, I'm sure that there are a bunch of benefits of scale, but
I'd like to think we had better antispam even when we were much
smaller than the competition.

Thank you for saying this. From a dimensionless analysis design and engineering standpoint, the only difference is higher levels of data collection of pretty much the same data we have all seen at any scale. More here, less there, since 2003, I have automatic daily statistics collected for various "anti-spam" methods:

http://www.winserver.com/public/spamstats.wct

It has been pretty consistent and persistent. IMV, any level of operation who follow the same deterministic protocols like SPF, should, in theory, see the same basic percentages I have seem since MARID. I have strongly focused on deterministic filtering methods. IMV, heuristics are not good for reliable rejections. I fundamentally don't feel we (as mail transport systems) should be in the "context-analysis" filtering business. Our principle responsibility has been to move mail and bounce it when required.

Non-deterministic heuristics protocols and methods are still too high on false positives. Its good for always accepting with classification but I also see many false positives. Gmail appears to have a problem with confusing support list systems as "bulk mailers."

My point, Google has not solved the spam problem even with big data, and I believe you conveyed that point.

--
HLS


_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp

<Prev in Thread] Current Thread [Next in Thread>