[afnog] AOL rejecting hosts with no rDNS?

[Brian Candler]

On Mon, Jun 28, 2004 at 07:21:55PM +0000, ALAIN PATRICK AINA wrote:
> Part of the community has understood that we first need to define" what the 
> problem of spam is", before we dig for solutions. 

Sure, although I think the problem of spam is pretty easy to define:
receiving unwanted mail from unscrupulous people who send out E-mails by the
million. And despite being annoying (most spam being untargeted, fraudulent,
obscene, or some combination of these), for me the main reason it's a
problem is because the *real, important* mails I want to receive can get
lost in a sea of spam.

In my opinion, the main problems with E-mail at the moment are:

(1) Spam
(2) Receiving bounces to mails I didn't send (with forged envelope-senders,
    aka "joe jobs")
(3) Viruses
(4) Forgery and general lack of trustworthiness

Of these, (2) has a reasonable technical solution - sender cookies - which
can protect yourself immediately without requiring the rest of the Internet
to change their practices. (4) isn't going to happen without some widespread
deployment of public-key crypto. (1) and (3) are real problems.

IMO the best way to identify spam is "it came from a spammer", and the only
realistic way of doing that is IP blacklists (of spammers' netblocks and of
open relays and machines they have hacked into). Trying to filter based on
any other characteristic of the mail - such as it being non-RFC-compliant in
one of many ways, or looking for invalid envelope-sender - is going to
falsely drop mail which is from misconfigured but otherwise well-meaning
people. And as soon as any such filtering is widely deployed, spammers just
change their spams so that they get around them anyway.

B.