The Bottom Line
SpamSlam implements effective challenge/response spam filtering and combines it with a Bayesian filter to avoid unnecessary challenges. SpamSlam doesn't weed out all the problems, though, and the statistics-based filter can't be trained. SpamSlam is no longer available.
Pros
- SpamSlam is effective against spam
- Works as a POP proxy, with any email program
- SpamSlam can use Bayesian filtering to avoid unnecessary challenges
Cons
- SpamSlam challenges can still be annoying
- Doesn't auto-whitelist recipients of sent mail, filter can't be trained
- Spam from whitelisted address gets through
Description
- SpamSlam is a spam filtering tool that uses a challenge/response system to detect spam.
- Working as a POP proxy between email client and mailbox, SpamSlam works with any email program.
- Only mail from whitelisted senders is allowed through to your Inbox by SpamSlam.
- New senders are sent a challenge (simple reply or a simple captcha), to which they must respond.
- Senders who do not respond within a certain time frame are automatically blacklisted by SpamSlam.
- Additionally, SpamSlam can employ Bayesian filtering to automatically black- or whitelist senders.
- Flexible rules let you black- or whitelist automatically using a number of criteria, too.
- Pending mail can be previewed and recovered in SpamSlam (doesn't whitelist the sender, though).
- SpamSlam can import addresses to the whitelist from text files.
- SpamSlam supports Windows 98/ME/2000/XP and Mac OS X 10.2.
Guide Review - SpamSlam 2.1.1 - Spam Filter
Good mail comes from good senders. All the other messages are junk. To prove they are good, new senders must respond to a challenge in SpamSlam. If they fail to do (spammers usually won't go that extra mile), they are blacklisted and all their mail is trashed.
SpamSlam is easy to use and, thanks its proxy nature, works with any email client. Unfortunately, SpamSlam can't automatically whitelist recipients of outgoing mail. What it intelligently does, though, is employing a Bayesian statistical filter before sending any challenges. Mail that's sure to be spam blocks the sender while senders of messages with a low spam probability are automatically whitelisted.
It's really a pity that you can't train the Bayesian filter. It also doesn't eliminate the burden placed on (legitimate) newsletter senders completely.
