Killing spam, and more

Like many of us, I have multiple email accounts. Some of them get large amounts of spam and while the spam killers on those servers kill most the spam, they don’t kill all of it.

Plus, it’s a pain to read some accounts in Thunderbird (my mail client) and my online email, like Google Gmail and Yahoo Mail, elsewhere. So, I’ve combined all of them into one and killed virtually all the spam in the process.

Here’s what I did.

I now forward mail from all my other email accounts to my Google Gmail account. You can easily set the forwarding in the configuration for those accounts.

When the forwarded mail hits Gmail, it is gets filtered. The filter for my polizeros email, for example, says take any mail addressed to me at polizeros, label it as polizeros, and archive it immediately. A label is like a folder. Thus, any mail from polizeros immediately is put into a GMail folder named polizeros. Plus, Gmail has a great spam killer. So not only can I read email from all my accounts in one place, there’s practically no spam in any of them because they’ve been spam checked twice – once by the originating email account and then again by Gmail.

You can respond from Gmail to any of those accounts using that email address, not your Gmail address, a useful feature indeed. Also, I used to read some of these email accounts via the Thunderbird mail client on my home computer. But that mail was difficult to get to if away from home or traveling. Now, all that email is being forwarded to Gmail, so I can access from any location with a net connection.

What do all of you do to kill spam and make email easier?


  1. Unless I’m mistaken, you can’t forward hotmail email. Please someone let me know if that’s not true and you really can; that would certainly be a useful thing to be able to do!

  2. I’ve been running my own email, one way or another, since 1984. In the beginning, there wasn’t anybody else to do it. Now, nobody else *can*. I’ve had the same email address since ’91 and it’s widely published. It gets hundreds or thousands of spam attempts per day. I don’t want content filtering, haven’t had to go there yet, but the day is coming.

    So my spam defense is different from that of people who do not control their own email server.

    I begin by rejecting “malformed” messages and messages from network addresses with no name. These are mistakes most spamware still makes, but few legitimate senders do. It stops about three quarters of the crap, without even looking at it.

    Then I (actually my email server Postfix) compare the sender’s network address against a map of “spammy places” I’ve been collecting since I started running Internet servers in ’98. Email from spammy places is returned with a terse explanation, “see,” and we “hang up the phone” before they waste our network bandwidth and CPU cycles with their stupid message body. South Korea is a spammy place. Russia and China and Brazil are pretty spammy. Turkey is really bad. So are all those trojan-infested Microsoft PC’s connected to cable modems worldwide. Then I consult a few public maps of similar information, known as DNSBLs. If you run an email server, you should subscribe it to the combined block list. Spamhaus is the best. is good, too. It’s Not Just Another Bogus List.

    Then Postfix looks for about a hundred “spamware artifacts” I’ve noticed over the years. Stupid little mistakes spam software makes, that real mail software doesn’t.

    We have now rejected about 95% of the incoming crap, without even inspecting it, and with fewer false positives than you get with content filtering. At this point, Postfix forwards it to my workstation, Truffula then inspects the messages with a program called Spambouncer written by Catherine Hampton. Spambouncer sorts them into three piles (spam, not spam, and maybe), and it’s been about 90% accurate. Spambouncer is too compute-intensive to run on an email server. Sometimes the messages can take half a second or more in there.

    (Sometime soon I will add a special content filter. There are accurate realtime block lists now of Domain Names Seen in Spam. I’m also looking at Vernon Schryver’s Distributed Checksum Clearinghouse. These are basically ways for groups of mail servers to share information about spam runs in progress.)

    Finally, I run through the spam pile and add new spam sources to my map, and post samples in where Google lets you search them. And I report the sources to the relevant ISPs. If they say too bad, eat your spam, I add the whole ISP to the map. If you don’t play by the (IETF) rules, you’re not on my Internet.

    I *read* email with the same email reading program (mail user agent) I started using in 1985, Berkeley Mail. It doesn’t know about attachments or fonts or mailbox files stored remotely. I’ve got other programs for that. It doesn’t know anything about filtering spam. Doesn’t have to. It’s really fast.

    Sometimes I *test* things with Mozilla Thunderbird or Kmail, but the Graphical User Interface really gets in the way.
    One of these days I’ll switch to a more powerful mail user agent, probably Mutt or mh or something. What I really want for christmas is for Berkeley Mail to learn about the mouse the way Vi did about ten years ago. That would be cool. Maybe that’s a reason to learn Mutt.

Comments are closed.