2 Objectives

The objectives of aegee-milter are to:

• save time of humans, spent processing data, at the cost of either delayed mail delivery, double delivery or rejection messages for partially delivered emails and partially undelivered emails.
• solve the problem with legitimate emails being put in the Junk folder and are not read in time.
• have a mail server with good IP reputation, which is achieved by reducing the amount of sent bounces.
• use Sieve for deciding whether to accept or reject a message at SMTP level.

Sometimes mailing list subscribers post with email addresses, they are not subscribed to the list. The mailing list manager can defer the decision on whether to accept the email to the mailing list owner, or bounce the message back to the sender, telling her that she cannot write to that address. Bouncing messages has the risk of blacklisting the mail server in some backscatters DNSBLs. Delegating the decision to the mailing list owner steals her time, forcing her to handle also spam messages, and does not scale. aegee-milter offers to reject at SMTP level such messages and communicate in this way to the sender what is going on. With this approach the time of the mailing list owner is saved, as the work is shifted to the sender.

Anti-spam software classifies sometimes legitimate emails as spam. If this happens, the messages are dispositioned in the Junk folder and neither the sender nor the recipients are aware, that the mail is missing. aegee-milter offers to reject the messages, telling the sender why the message was evaluated as spam, and possibly inserting in the rejection message how the recipient can be contacted (snail mail, phone, fax etc). This saves the recipients the time usually spent digging in the Spam folder, and the concequences of reading a mail much later than expected. ¹ A further option is to deliver the message as Junk to the recipient, but use rejection at SMTP level, telling the sender that the mail was delivered as spam, why it was evaluated as spam, and what are the alternatives to contact the recipient.

Imperfect anti-spam software makes errors and senders can ask why their email was evaluated as spam. When using anti-spam software from the public domain, the outcome of the anti-spam evaluation is not a secret. Putting the output of the anti-spam software in the SMTP rejection reason, when the email is rejected due high spam probability, saves to the site administrator questions, why a specific email was evaluated as spam.

Sieve is a standard language for filtering emails RFC 5228. There is software that can upload and download Sieve scripts to a server using the ManageSieve protocol RFC 5804. With Sieve, the user can inspect the headers a message has, in particular the amount of scores assigned by the anti-spam software, and based on them to decide to accept or reject the message. When a message is rejected, the user can specify the text to be returned to the sender. This text can include the telephone or fax number of the recipient and its snail mail address. In this way, when the anti-spam software classifies a legitimate email as spam, the communication flow is not interrupted by silently delivering messages as Junk.