> It's also important to realize the purpose of the HELOBOGUS test. It
> isn't designed primarily to catch spammers. It's designed to help
> detect poorly administered mailservers -- ones that are likely to be
> abused by spammers.
True, but if you're using HELOBOGUS for anything other than
advertising to your clients' clients--which Declude is definitely good
for :)--you're giving it a weight, so you are using it not only as
community outreach, but as a spam test.
> And those Fortune 500 companies that have their mailserver advertise
> itself with a name other than what it really is, well, they are
> running mailservers that are poorly administered.
I have zero respect for people who think they're too big to change:
CitiGroup actually has a stated policy that they "do not make changes
for outside companies" or suchlike, which they use to avoid fixing
problems they don't really understand. But we can't have
zero-tolerance for HELOBOGUS in practical terms, since we risk losing
clients by losing their clients, and the more hoops it takes to get to
an IT group, the more annoyed everyone becomes (even if their own
bureaucracy is at fault).
> But if you don't penalize them, they will definitely continue
> bending the rules too far, which helps increase spam.
Yes, something must actually break, even if it just means that they
consistently trip the weekly ALERT threshold. But again, speaking from
a combo of experience and my own grudges, a dead HELO of
'www03.example.com' is a lot less likely to get fixed than a dead HELO
of just 'mail.' Even the stupid mail admin can see and fix some
problems with the latter, while the former will likely involve
contacting the much-feared DNS group, blah blah blah. And when people
do ask us how to fix pass a "looser" test, we will of course continue
to say that a published FQHN is required, still spreading the
"tighter" word to those admins.
We're pretty strict on our own. SPAManager, for example, was not our
idea. But clients dictate varying tolerances. Something that has
surprised me is how likely difficult internal users are to have
irascible, irrational external contacts/friends--self-evident, I
suppose, but the parity is just uncanny sometimes! At any rate, a
looser HELOBOGUS option (maybe a separate test completely, now that I
think about it, to enable varying weights) would make HELOBOGUS less
of a liability for us.
>>But I WOULD use a negative test in the style of IPNOTINMX,
>>"rewarding" a site slightly for having the ability, experience, and
>>control to match the two and hopefully combatting some FPs.
> Aha -- like the IPNOTINMX test. That's a good idea.
Glad you agree there! I think the two tests (exact match and
parent/grandparent domain match) would be perfect.
-Sandy
---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.
