Re: [exim] Spam prevention methodics

On 12/29/05 6:08 AM, "kig@???" <kig@???> wrote:

> I can be wrong to ask it here, but
>
> i heard there a lot of custom spam prevention methodics used in MTA during
> SMTP session time
> currently i'm interested in a one when mail is prohibited in case of "too many
> not found recipients in TO header."
> Is it reasonable to use it in production and does not violate netiquette? Or
> there are some more methodics (where to read?).

It can be useful to do that, but there are pitfalls. At the time the ACL
for RCPT TO: is running, there are ways to count total and bad recipients
seen so far (see the spec, or search this list's archives).

One example happened in the community were I live a few years ago. A new
person took over an emergency notification list (list in the Outlook Express
group sense, I think) which had been moribund for some time. Naturally
there were bad addresses in it (the more so because the community had
recently gained cable-based internet access, and many residents had fled
their dialup accounts). The messages directed here were received or bounced
appropriately...the messages directed to some other local ISPs bounced
because of too many unknown recipients (in some cases, with no indication as
to which were unknown).

An alternative, if there is a mix of known and unknown recipients, is to
make the message seem "more spammish"--however you do that in your
installation. But if you quarantine some spam silently, that cure can be
worse than the disease.

--John