[Mimedefang] Reject without PTR record

[Kevin A. McGrail]

> What do you think folks, is that a good idea to reject emails from relays
with
> no PTR record?


Dmitry,

<SOAPBOX>
I would say absolutely not.  While it certainly might be RFC mandated or
recommended, the fact of the matter is that most ISPs treat reverse DNS as a
non-necessity and coordinating changes for legitimate customers can
sometimes be like pulling teeth and in some cases ISPs simply ignore/refuse
to do reverse DNS changes or implementation.

Additionally, because of the growing prevalence of
firewalls/proxies/natting, the amount of people that actually care about
reverse DNS has dropped dramatically.

I've had conversations with high-end Admins who still don't even know what
reverse DNS is or why it exists.  It takes me 20 minutes to explain why they
can't use FTP or IRC or explain the 30 second lag on connecting to XYZ
service :-(

Because of these reasons and probably quite a few more, forward/reverse DNS
verification and PTR existence verification is simply too tough to implement
and leads to excessive failures for legitimate users.  I am not a person who
is out to change the world unless I see it as a final-end solution and this
is not one.

This is the same argument I make about compliant-standards Nazi (Ximian,
RFCIgnorant.org and UW) who will not implement simple fixes to work-around
broken/non-compliant (read Microsoft) products.  Do I think the companies
care? no.  Do I think they will change? maybe but unlikely.  Do I need the
solution to work regardless? absolutely.  I work in reality not in academia
or the ilk.  A reasonable margin for error is critical to allowing crucial
services such as email to work.  A college student who loses a few emails,
boohoo.  A company that loses a few customers because email delivery was
screwed up and I'm out of a job.

This is the same reason I support SpamAssassins ability to TAG emails and do
not implement deletion or bouncing of SPAM no matter what the score from
SpamAssassin.
</SOAPBOX>

Regards,
KAM