> I'm  all  of  a  sudden  starting to get a lot of bounce messages on
> accounts  that  I'm  filtering for. It's the trick where the spammer
> co-opts  either  a  domain  name or a full address...

"Joe Job" is the common term for this.

> Do  I  target  all  bounces for deletion?

Not if you want to retain your customers.

> I'm wondering if this will create problems with my users not getting
> their  bounces when servers like AOL seem to accept the message only
> to  issue  their  own  response  instead of letting my server, or my
> customer's server handle the errors.

Of course it will create problems.

In  another  thread, you've argued (unconvincingly, to my mind, and in
the  face  of  best  practices)  against  having  MXs  reject  unknown
users--and  in  that  same  thread, you seemed proud of your policy of
swallowing   all  misaddressed  mail  at  the  MX.  You  offered  some
reasonable  defense  of  that policy, but now you're escalating to the
position   that  no  bounces  whatsoever  can  originate  from  non-MX
machines?  That  unfortunately suggests that you haven't made an fully
educated  decision in either case. In addition to rejections generated
at the MX, bounces can be generated for mailbox- and content-sensitive
reasons,  unrelated  to  spam  detection, that can only be detected on
mailbox  servers.  The  bounces  therefore  may be routed through MXs,
through  outgoing  gateways,  or  directly  from  the  mailbox servers
themselves, depending on site-specific factors.

IMO,  it's in everyone's interest to reject as much as possible at the
MX,  but  mega-providers--and  anyone  running  IMail  as mailbox-only
behind  an MX, for that matter--simply cannot reject everything at the
edge.  It's absurd to stipulate that only envelope rejection at the MX
is valid.

The  only  arguably  deceptive, though common and reasonable, "seem to
accept"  scenarios  are  when  mail  is  NULlified after acceptance by
anti-spam  software  such  as  Declude. Accepting mail at the MX, then
sending  a  notification of deeper failure, is not "seeming to accept"
anything--it's  common, reasonable and RFC practice--unless *you* then
in turn NULlify the notification.

> Personally,  I've  had  all  outgoing bounce messages turned off for
> about 6 months

You  mean  bounces for suspected spam, or all bounces entirely? If the
latter, you're certainly doing your customers a disservice!

> It  seems that between the spammers and the virus programmers, error
> reporting  has  been  ruined,

You're concluding this based on 10 per week? :)

> In my opinion, bounces should only be sent when the MAILFROM matches
> the  REVDNS  domain

Well,  sorry,  but...your  opinion does not jibe with the way SMTP has
been  deployed  all  over  the globe. See Kami's response for just one
example.

--Sandy


------------------------------------
Sanford Whiteman, Chief Technologist
Broadleaf Systems, a division of
Cypress Integrated Systems, Inc.
e-mail: [EMAIL PROTECTED]

SpamAssassin plugs into Declude!
    http://www.mailmage.com/download/software/freeutils/SPAMC32/Release/

---
[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.

Reply via email to