That Fleming Gent | INSERT COIN FOR (1) PLAYER GAME

Marketers: READ THIS IF YOU'RE IN AUSTRALIA OR YOUR "TARGET MARKET" IS HERE. OR YOU WILL BE SORRY!

Spam Act 2003: General Information
Spam Act 2003: Guide For Business
Spam is about consent, not content. Religious and political spam, even charity spam, gets the same treatment as Viagra and porn. If I've not requested it, I don't want it, it's junk.
Communications over a private network is a privilege, not a right. I own my hardware and bandwidth, therefor I make the policy - I'm not bound by "free speech" laws or other legal scripture many bulk emailers love to quote. My Network, My Rules!

For postmasters and mail server admins:

DO: A strong anti-spam policy is a must in these times.
DO: Let your users know your mail policies, especially when it comes to filtering. If it's a good one, they'll be happy (and so will the community at large). Good policy means happy users
DO: Respect the mail standards, even if others do not. Well behaved and configured software is a boon to an admin and to the network at large
DON'T: When dealing with spam and virii, you have many choices. You can bounce, discard or tag+deliver. For the good of the network - please DO NOT reject virii or even spam especially after the DATA phase - both frequently use forged SMTP addresses and result in backscatter - some poor bugger's gonna have a full inbox! Either tag it so your users can filter it themselves or discard it at the server level - the latter is my preferred method for virii.
DON'T let your users send using invalid or non-resolving domains - these are bad even as a general rule as any bounces will queue on yours or someone else's server for the queue length time at that host, never capable of delivery - if your mailserver software is capable of rejecting unresolvable, invalid or nonexistent domains, use it!

What do my domains use?

I do not allow free relay, via the standard Postfix access controls for relay prevention
I run a local address / domain blacklist which is queryable via the previously mentioned check form.
I implement greylisting [What is greylisting?] via the excellent "policyd" policy daemon for postfix
I use the below linked and recommended DNS blacklists - which are quite effective.
I have some spamtrap addresses for filling the local blacklists with idiots who send to them (no telling what those spamtraps are though!)
Naturally I have antivirus and spam filtering (bogofilter) once the above has been passed

Blacklists I use. These are checked before any mail reaches me or my local domain users:

Many domains and service providers use the SORBS blacklists run by Matthew Sullivan. I don't and neither should you. Listing criteria are quite elastic in practice and once listed, delisting is questionably inefficient (I'm being kind here) if it works at all - you are asked to pay $50 to charity and notify SORBS, but I have seen occasions where this has been done and the listing is still present.
(..and what sort of public list demands money to be removed, even if it is for a good cause??)

Worst of all, the evidence presented is often minimal and useful - I can think of one occasion (via a work investigation) where an IP range listing's evidence file is a single Recieved: line. Not good Matthew...

Network Abuse Clearing House
Mail Abuse Protection System
Coalition Against Unsolicited Bulk Email (Australia)
Sender Policy Framework home page - help stop domain forgery and backscatter
Greylisting.org - explanations and examples of greylisting; keeping spamming hosts delayed and killing single-shot spambots dead