Dnia 15.04.2022 o godz. 19:20:28 Laura Atkins via mailop pisze:
> 
> Would you really hold it against that company, given the data they have,
> if they blocked all mail with that tld in it? Given that it’s 90+%
> guaranteed that tld is spam? What if 90+% of the mail in the .eu.org
> <http://eu.org/> tld is also spam? Does it make more sense to block mail
> containing that domain? Or are we just refusing to consider any domain
> based blocking at all?

I think *domain* based blocking is OK, if the domain is confirmed to send
spam, but this applies to end-user domains, not to TLDs. TLD-based blocking
is definitely not.

Similarly IP-based blocking is OK, if the IP is confirmed to send spam,
netblock-based blocking is not.

Even if 95% of mail from some TLD (or netblock) is spam, that remaining 5%
could contain something very important to someone.

Of course I'm talking about general email providers. If someone runs a
private mail server and does not expect that he/she ever receives any
legitimate mail - for example - from Brazil, it is understandable that he/she
blocks the entire .br domain or Brazillian networks (a LOT of spam coming
from there!). But for a general email provider, blocking off Brazil would be
wrong, even considering the huge amount of spam coming from there.

There was a previous discussion in this thread regarding spam filtering more
concentrated on "badput" or "goodput". One can either have a goal to minimize
the amount of spam passed through, caring less about some legitimate
messages being filtered out, or one can have a goal to minimize false
positives, living with the fact that a few spams will end up in the inbox.
My opinion is that you cannot have both. Either you minimize the amount of
spam - sacrificing some legitimate email in process - or you maximize the
amount of legitimate email, allowing some spam to pass through.

I obviously prefer the second approach, while it looks that Google is taking
the first one. For recipient being an experienced user, the message getting
into spam folder is no big deal, because such users will check their spam
folders anyway. But majority of Google users have little experience :). They
just believe that things in spam folder are actual spam and they have no
reason to look there. My recipients for example, even after multiple
cases of my messages ending up in their spam folder, seem to still not
develop the habit of checking that folder too, nor clicking "this is not
spam" on messages that mistakenly landed in the spam folder.

Therefore I think a few spams landing in the inbox do less harm to an
average user than even one legitimate mail filed to spam folder, because
even an inexperienced user is able to note the obvious signs of spam and
delete the message manually (at least that's what I believe, maybe I'm
wrong?). And if we're talking about targeted, sophisticated phishing, even a
very strict spam filter will be usually unable to catch it anyway. Just
today at my work email I received a phishing message that no spam filter
would catch (unless the URL of the link that I was supposed to click would
be known to the spam filter - but for targeted phishing they usually use
fresh domains, not used previously). It was just a regular message that an
average person would send, only the content was suspicious because it
claimed things that were unprobable in relationship to my actual life.
But it wasn't any well-known scam like Nigerian fraud, so no filter could
know that, only a human can :).
-- 
Regards,
   Jaroslaw Rafa
   [email protected]
--
"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
[email protected]
https://list.mailop.org/listinfo/mailop

Reply via email to