[ILUG] Spamassassin questions

[Des Keane]

----- Original Message -----
From: Rick Moen <rick at linuxmafia.com>
Date: Thursday, June 10, 2004 7:10 pm
Subject: Re: [ILUG] Spamassassin questions

> Well, one way is for the MTA in question to have not accepted
> forged-return-path spam in the first place.  If it does so, the admin
> has a problem.  How the admin deals with that problem (if he 
> allows it
> to arise) is his responsibility, not mine.

Does that not suppose that every MTA on the Internet is suddenly going to get upgraded and that forged-return-path spam/virii is already a solved problem?

> The way forward is to fix responsibility with the admins generating
> inappropriate mail.  Making my MTA issue a Delivery Status 
> Notification(DSN) stating that one has refused delivery is not 
> generating mail:

In the scenario you outlined, where you send a 550 response code your MTA doesn't generate any DSN, the sending/reporting MTA does surely. How do you propose to hold responsible "the admins generating inappropriate mail"? If my OS is infected with a virus and it is generating mails from somevaliduserinmyaddressbook at mydomain.ie and sending them through the registered MTA for mydomain.ie whilst dialed up on mydomain.ie ISP IP address range what is the mail admin for mydomain.ie supposed to be doing?

> Right now, all major mail-handling sites publish SPF records, and an
> increasing number of smaller ones.  There's no longer an excuse for
> accepting -- let alone attempting to forward -- joe-job mails

Is SPF not an experimental Internet draft currently (amongst other contenders)? Are there not questions still being considered about the implications for mail forwarding and mailing list expansion? I haven't followed very closely at all so I may well be mistaken. As you say this approach is primarily about reducing volume.

Apart from appearing to take responsibility for a problem not of your making (which I appreciate), if I dare ask, what is your objection to silently discarding a joe-job spam/virus? Is this not effectively what you want the previous hop MTA to do in any case?

BTW, cool links on your kb page, thanks!

Des