On Tue, 28 Jun 2022 00:21:00 -0700
Keith Lofstrom <[email protected]> wrote:
> I need a better spam filter process.

On Tue, Jun 28, 2022 at 03:45:06PM +0000, Cy wrote:
> /etc/postfix/main.cf. Then I can make a "from_someone: user" line any time I 
> have to
> provide an email address. When "someone" sells me out to spammers.

An OK answer, but not to the question I asked, or relevant
to my situation. 

I am asking for a multi-site collaborative spam filter
for postfix, not losing long-term email collaborations. 
My communities are not so ephemeral or trivial that I
can abandon my net presence and start another.

Explanation:

"[email protected]" and "[email protected]" have been "out
there" since the beginning of the internet domain system;
[email protected] before that, bang paths to ucbvax
before that, back to the late 1970s.  Tens of thousands of
legitimate email correspondents have used them and still
do, hundreds over the last few months, some after silent
intervals of two decades.  

I have a HUGE white-list with most of those correspondents.
Unfortunately, some are "domain homeless" and frequently
change their email providers and addresses.  They may have
abandoned their individuality to "google consumerism".

Over the decades, I have used my main email addresses in
many journal papers and publications.  Readers contact me,
sometimes decades after publication, leading to new
collaborations and friendships.  Hiding behind transient
aliases thwarts that precious but unpredictable
collaboration.

Some are unwelcome, but blacklists work, until the
unwelcome change addresses to pester me again.  Trolls
don't understand "go away", but postfix can.  

In "one time use" situations, a throwaway email address
like Cy suggests makes sense.  "two time use" email
addresses with time delays between uses -- ordering items
from rural China through Alibaba, with very long delivery
delays, for example -- is another maintenance issue for
one-time addresses.   Perhaps solvable using yet another
hypothetical postfix plug-in (suggestions?).

But spambots ALSO harvest my permanent email addresses
and spew crap at me.  Fortunately, they spew the same
or very similar crap to thousands of others. 
Networked spam detection could easily detect that. 

Indeed, gmail (with millions of users) gets thousands of
copies of the same spam, very close in time, allowing
Google to DETECT that spam-sign coincidence.  Google bots
may use such detection to filter crap feeding ALL of their
captive gmail users, so that Google doesn't waste disk
space or user connection bandwidth on spam. 

I imagine large corporate and academic email systems can
do something similar.  Many use open source software,
which implies the existence of open-source-compatible
tools for a broadcast-spam detection process.  If that
process is collaborative among those large organizations,
they can "divide and conquer" spammers before wasting
their own disk space and user connection bandwidth.

I ask to collaborate with that hypothetical shared
detection and filtering tool. 

Perhaps it does not yet exist, because most developers are
not innovators, merely cogs in someone else's innovation
system.  OTOH, perhaps some clever young lass or lad WILL
create such tools, earning a small fortune selling support
to corporations and universities, and maybe even me. 

----

So, howzabout it?  Does anybody on this list know of tools
like I am asking for, or dream of developing them?  Are
there any innovators or entrepreneurs on the PLUG list?

Keith

-- 
Keith Lofstrom          [email protected]

Reply via email to