On Thu, Sep 26, 2019 at 6:39 AM Dave Crocker <dhc(_at_)dcrocker(_dot_)net>
wrote:
On 9/25/2019 11:57 PM, David MacQuigg wrote:
The big problem I see with new solutions to the spam problem is - there
is no longer a problem. Most of the big email services are doing such a
good job of filtering the spam that there is really no incentive to
deploy a whole new system.
This is a false sense of security.
90-95% of the email traffic across the open Internet is spam. Worse,
the bad actors are intelligent, aggressive and adaptable. The current
situation is a constant arms race. That defines a fundamentally
unstable situation, no matter how well the defenders are doing at any
given moment.
The volume of spam is not a problem for a properly implemented Receiver,
like I had at box67.com http://open-mail.org/BorderPatrol.html As for the
super-human intelligence of spammers, I have not seen any scheme that can
defeat a well-implemented whitelist.
So, yeah, end users typically see only a tiny fraction of spam, but a)
that requires massive amounts of continuing effort by those running
filtering agents, and b) enough still gets through to cause real-world
problems for end users.
I was surprised by how little effort it took to whitelist a few dozen of
the biggest senders, which accounted for maybe 95% of our legitimate
incoming mail. That's a 20X improvement over sending it all through a
statistical filter. Yes, there is still some doubt whether we could have
"scaled up" to a system the size of any of the big services, but the
possibility of scaling up was always in the design. Before I lost
interest, and moved on to other projects, we had plans to use our Receiver
at the University of Arizona with some 74,000 recipients. This would have
required 2 or 3 students working part-time to maintain our whitelist. I am
confident it would have worked. We just couldn't get any funding. Also, U
of A decided to let Google handle all their mail.
As for the continuing problems of real-world users, I suggest that they
switch to a service like gmail or yahoo.
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp