[Mimedefang] On pinheaded ISP's (sort of OT)

[Kevin A. McGrail]

>   Here is an example of why that does not give enough information.  AOL
> has a service where I (as email administrator for our domain) can get
> what's called a "Feedback Loop" which basically means that I get email
> whenever an AOL user clicks "This is spam" on email that originates from
> our campus.  Problem is, they strip out all identifying information about
> the subscriber that made the complaint.  Now I can *somtimes* figure out
> enough information from my sendmail logs to see what the problem is, but
> here is something that happens all the time:  Mailing list has 100
> subscribers, 23 of which are AOL addresses.  AOL subscriber decides he
> doesn't want to be on the list any more so rather that unsubscribing, he
> just clicks "This is spam".  I get the Feedback Loop email from AOL.  I
> can see who sent the original email, and the contents of the message, but
> sendmail just shows me that it went to 23 aol.com addresses and I have no
> idea which one complained.  So how do I fix that?  I thought the Feedback
> Loop was a good feature, but all it is is an annoyance because I don't
> have enough information to remedy the complaint.

We see the exact same issue with 100% legitimate mailing lists as well.

I know exactly what you mean with the redacted email issue.  Right now, we 
use mailman so we'll see something like rly83=aol or something similar in a 
header.

But if that doesn't work, use MD and add a header do some ludicrously simple 
encryption like changing a's to 1, b's to 2, etc. for the original recipient 
on the outbound so you can back-calculate the original recipient.

My latest annoyance though is with the procedures from 
http://www.borderware.com/ and, for example, 
http://intercept.borderware.com/lookup?ip=63.216.184.10

We, of course, have decent procedures in place to extend the edge of our 
networks using fine programs like MIMEDefang.  But we are also the backup 
mail server for hundreds of domains.  However, they count NDRs as SPAM even 
though we only generate an NDR when it is impossible to otherwise and we 
have no reason to believe our problem is anything other than NDRs.  (see 
http://bsn.borderware.com/faq.php?ip=63.216.184.10&d=pccc.com&u=)

Is anyone else running into issues with their intercept system?

Regards,
KAM