Tame OS X Mail Spam with SpamSieve
After going through a privacy / paranoia phase concerning the use of Gmail as my primary account, I eventually made the decision to go with a hosted email solution [Dreamhost] that would provide a little more control and peace of mind. Although Gmail does offer POP access, I yearned for the convenience of IMAP. Having access to both email protocols allowed me to access my mail from a mobile handset via IMAP, while continuing where I may have left off via POP on the MacBook.
Although I enjoy the flexibility of both IMAP / POP, I found myself questioning the move due to the increasing Spam arriving on a daily basis. Despite Apple’s integrated Spam filter, and my regularly updated filtering rules, the Spam continued to find its way into my safe folders. Rather than tweaking the built-in filtering rules to coexist with Apple’s joke of a Spam filter, I opted for a small Spam filtering utility by the name of SpamSieve.
It is important to understand that SpamSieve itself is an application that must continually run in the background while your default email client is running. If you’re not too keen on having the dock icon setting up permanent shop within your dock, check out Dockless which provides a simple GUI for hiding specific application icons.
Integration with your default mail client – Apple Mail, Emailer, Entourage v.X through 11.x (2004), Eudora 5.2 – is a simple one-click install from the SpamSieve menu. By design, SpamSieve was created to replace your existing mail applications Spam filter with its own Bayesian filter. During the initial install, users train begin to train SpamSieve by selecting “safe” messages and marking them as good (Ctrl+CMD+G). Items which should be marked as Spam (Ctrl+CMD+S) will be added to the SpamSieve blocklist for future reference.
To date, SpamSieve has amassed the following statistics concerning Spam mail to four separate email addresses:
- Filtered Mail: 976 Good Messages, 986 Spam Messages (50%), 32 Spam Messages Per Day
- SpamSieve Accuracy: 24 False Positives, 23 False Negatives (49%), 97.6% Correct
With continual training, I anticipate a 99% accuracy report. At the moment, emails which I consider Spam that manage to slip through SpamSieve’s filters are messages composed in different languages.
If the performance of your desktop email clients Spam filter is beginning to disappoint, check out the 30 day demo of SpamSieve and decide if the solution fits the bill.