Are you tired of SPAM? Well, there is an application to stop it all. It's SpamSieve
from C-Command. SpamSieve is a plug-in SPAM filter for popular Mac OS X mail applications.
Instead of just marking what it thinks is junk based on pre-set variables, it uses
a technique known as Bayesian filtering to learn with your mail and continually improve
its understanding of what is and is not SPAM. Simply put, this junk filter is smart,
and great for anybody who is tired of SPAM - who isn't?
SpamSieve is Universal and runs on Mac OS 10.2 and above though at least 10.3.5 is
recommended. The only other requirement is that you use one of the following supported
- Apple's Mail,
- Entourage (v.X through
- Eudora 5.2 or later
(Sponsored or Paid),
- Mailsmith (Bundled
with Mailsmith 2.1),
- Outlook Express 5,
Once downloaded, you can simply drag the application into the Application folder
on your computer. Launch SpamSieve to install the correct script for your mail program.
Afterwards, you must set up your mail program to recognize the script and send all
incoming mail through the script to test it for SPAM. Precise directions for all
of the supported mail programs can be found on the C-command.com website or here.
Since Apple's Mail is standard on all Macs running OS X, I will go through the steps
for that program.
First, launch SpamSieve and under its menu, choose to install Mail's plug-in. Then,
launch Mail. Create a new mailbox called "SPAM" - make sure it is named
SPAM and not Junk or anything else, otherwise SpamSieve will not work. Now, go to
Mail's preferences and create a new rule. (The description of the rule must start
with SpamSieve, but you can add additional text after that, if you want.) Change
the From menu to say Every Message. Then, next to Move Message, select your new SPAM
mailbox. Click ok, and, if asked, choose not to apply the rule to your open mailboxes.
In summary, make sure the rule looks like this:
Now, because SpamSieve
uses Bayesian filtering, you must show the program what is SPAM and what isn't. The
company suggests that, rather than training SpamSieve with all of your mail, you
use 65% SPAM and the rest good mail. This proportion allows SpamSieve to make the
best progress in improving its filtering. To train messages as SPAM or good, simply
select one or more messages and then select one of the newly added options "SpamSieve
- Train as SPAM" or "SpamSieve - Train as Good" respectively from
the Mail Message menu; there are also keyboard shortcuts for these commands. To tell
how many messages you have marked as each, go into the SpamSieve application and
choose to see statistics, or keyboard shortcut Command-2. Half way down the window,
you can see the percentages, as shown below
So in review, you set up your rule to send all incoming mail through the SPAM filter.
Then, train the filter with your own mail using the menu commands. Once you have
about a 65/35 ratio of SPAM to good, you are all set up.
Now that your filter has been set up, the in-use part is easy: send and receive mail
and never worry about dealing with SPAM again. If one of those sneaky SPAM messages
does get past the Bayesian filter, all you have to do is click on the message and
"Train as SPAM". SpamSieve will learn what makes it SPAM and then move
it into the SPAM mailbox. This also goes for the other way around. From time to time,
look in the SPAM mailbox to check for good messages and delete all of the SPAM. If
a good message accidentally got marked as SPAM, simply select it and "Train
as Good". SpamSieve will learn what makes it a good message and move it into
I have used SpamSieve for about a month now. One of its cool features is the ability
to see how well the filter is doing via the Statistics window. There is nothing better
than seeing that the SPAM filter is actually catching SPAM. Remember, we used this
window to see how we trained the program for initial setup? In SpamSieve, simply
go up to Filter>Show Statistics or command-2 to see the window, shown below:
SpamSieve Filter Results
Here you can see how
many messages SpamSieve has filtered and your daily amount of SPAM messages in the
first section. The second section shows how well it has done, listing its false positives
(how many good messages it marked SPAM,) its false negatives (how many SPAM messages
it marked good) and what percent of messages it correctly marks. It also shows how
many messages you have used for training, both good and bad, as well as how many
words are being used to identify SPAM messages. The next section shows how many e-mail
addresses it blocks and how many it always marks good. The final section shows you
the date statistics are being shown from; you can configure the date to see just
the past day or week, or the entire time you have had the program. For the month
I have used the program I have had, according to SpamSieve's stats, a 95.3% correct
percentage. This has saved me much time in reading e-mails, allowing me to get more
I encountered one problem; after setting up the program, I got a SPAM message that
was marked good. However, when I tried to mark it as SPAM, SpamSieve wouldn't let
me. The C-command website had some suggestions for information to send in e-mail
when asking for help, such as the log and screenshots of my set-up. I contacted the
author, Michael Tsai, who quickly got back to me on some steps to take, like restarting
the mail application. He also sent me a script along with some instructions. After
following the directions, I e-mailed the results back to Mr. Tsai who then explained
what had been done and what it meant. Although we are not sure why, the problem was
fixed by running the script and restarting my mail application. Overall, the customer
service experience was great, he was very prompt in his responses as well as very
kind and informative, and the problem was fixed. Other then that, I have not had
As far as preferences go, there is as much customization as you could want, including
everything from how hard SpamSieve is on messages to your dock icon. Under SpamSieve's
preferences, there are 4 tabs:
Under the Filters tab, you have the option to include and not include certain filters
such as using Address Book's entries, using the Habeas safe list (a pre-set list
of safe addresses) and enabling or disabling SpamSieve's main Bayesian classifier.
Under the Notification tab, you can set notifcations for the arrival of new good
mail such as playing a sound and bouncing the dock icon. You can also configure Growl,
a free third-party utility which gives you popup notifications of SpamSieve's activities.
Next, under Training, there are options such as whether or not to allow duplicates
in the database and training different filters such as the white list and block list
or the Bayesian filter.
Finally, under the Advanced tab, you can set how hard you want SpamSieve to be on
messages - I left it in the middle as suggested - as well as options like using a
photo realistic dock icon or quitting SpamSieve when your mail application quits.
No matter how you like your SPAM, SpamSieve can handle it.
SpamSieve Advanced Preferences
Now, I know that I went kind of fast over that stuff, but don't worry, because one
of the best features of this application is its manual. The manual for SpamSieve
covers everything in the application in great detail, and if there are any questions,
Michael Tsai is very easy to contact for help with anything you need!
SpamSieve is a SPAM filter using Bayesian filter techniques, allowing it to learn
and grow with your e-mail messages. Its set-up is a little complex, but the accompanied
manual walks through the process step by step with lots of pictures. After set-up,
you can sit back and not worry about SPAM again. This is definitely a very useful
application. It has saved me much time and effort in filtering my e-mail. Anybody
who is tired of dealing with SPAM needs SpamSieve; it is smart, easy, reliable and
a great piece of software to have.
- Easy to use
- Learns with your
- Very reliable after
- Easy to see statistics
- Great documentation
- A bit complex to
- Did have one problem
4 1/2 out of 5 Mice