SpamNet

I am on a constant search for new and better products to help with the fight against spam. It's a war, and like all wars the combatants are each improving their skills and finding new ways to attack and get around defenses. This requires me to maintain my spam-fighting skills to keep my inbox clean. More importantly, I want to keep my wife's inbox clean - a happy wife, after all, means a happy life.

Recently I found a product called SpamNet which works extraordinarily well in the war. It has a unique but extremely precise algorithm for identifying spam.

SpamNet is appropriately named, in that the product ties into a vast network of people who also use the product. As each person identifies spam (by blocking it), the message id (in the header of the email itself) is transmitted to a central server. This server keeps track of what is reported as spam.

This database allows SpamNet to be very precise in it's identification of what is and is not a spam message. There is another twist - as you correctly identify spam the system begins to trust you more and more.  Thus, as you use the system (and accurately report spam), your word carries more weight.

I liked this program as soon as I read about it, as it uses an entirely different technique to figure out what is and isn't spam. Other methods use on blacklists of ALL email from an entire TCP/IP address (or even a range of addresses) or compare each message against patterns typical of spam. Both of these methods work (and they work well), but tend to generate lots of false positives (marking good messages as spam). While I generally don't care about the occasion false positive on marketing junk or newsletters, I do want to receive everything from my egroups and topica lists as well as my friends.

The interesting fact about this method is messages from friends will NEVER be marked as spam (unless the message gets sent to a lot of people), because the message will not be marked as spam by anyone. Messages from egroups or topica (or other similar mailing lists) will also not normally be marked as spam, since these do not get sent to more than a few dozen or perhaps hundred people. This seriously reduces the chance of false positives.

You download SpamNet and install it on your system. It gets installed as an add-on to Outlook 2000 or XP, and thus works very closely with those email clients. When you receive mail, each message ID is checked against the database of known spam messages. Any message with an ID known as spam gets moved to a special folder call, well, spam. You can review the messages in this folder at any time and "unblock" them as desired.

Once you have received mail into your inbox, you simply click the "block" button on any message which you feel is spam. The message ID gets set to the spam server and added to the database so others will not receive that message.

While the system does work well, it has some disadvantages. First, the software must communicate at least once per email message with a remote server. For slower connections to the internet, this could slow things down quite a bit (I don't notice as I've got a 6mbit/second link). 

Second, I've noticed that newsletters and such do occasionally get marked as spam even though they are not spam. For example, I receive newsletters from Internet.com. These messages are definitely not spam, yet since some people in the SpamNet system indicated they ARE spam, then they get treated as spam. However, it's not that much work to look at the spam folder once in a while and unblock those messages.

One feature that appears to be entirely lacking but is desperately needed is a whitelist. This would allow you to define email addresses which are never blocked. In fact, if this were incorporated into the product it could entirely eliminate the false positive problem.

Oh yes, I forgot the best thing of all. This program is free. This makes it one of the best bargains on the internet. Highly recommended.