Arian said:
How can I block offensive e-mails? I am no prude but I am sick and
tired of
having to filter through endless e-mails relating to sex, the size
of a
penis, or cheap pills or replica watches etc....etc....before I get
to the
real e-mails. It all takes up too much valuable time which I do not
have, I
am trying to run a business. I do have mailwashers and junk filters
but they
do not block enough. All suggestions gratefully recieved.
Mailwasher (and not "mailwashers") is worthless in its free version.
Far too much is crippled in the free version to make the product
nearly worthless. You didn't say whether you had the paid or free
version; see
http://www.mailwasher.net/. The "learning filter" is
their way of hiding that they use a Bayesian-type filter. The DNS
blacklists (of known spam sources) and their global spam database
(another blacklist but their own and not a public one) is not
available in the free version. And NEVER use the bounceback feature
of MailWasher since that can get you blacklisted because of sending
your unsolicited bogus bounces to innocents who never sent you the
spam and will end up reporting you for that "bounce spam".
Backscatter, like bogus bounces, is reportable to the blacklists.
SpamPal is free. It has a Bayesian plug-in to learn what is spam and
what is ham. It uses blacklists but you can configure which ones to
use depending on how aggressive you want to be. NEVER use the SPEWS
(now UCE-Protect) blacklist as it is inappropriate for identifying
spam mail itself. It rates how spammy is a domain, not if a
particular e-mail is spam. I use the following blacklists: SpamHaus
SBL+XBL (includes SBL, XBL, Composite BlackList, and blitzed.org
lists), ORDB, SpamCop, SpamBag, and NJABL. A link to each blacklist
is provided so you can see how they detect spam or their particular
orientation to spam. I don't know which blacklists are used by
Mailwasher and if you even get a choice as to which ones to use. My
trial of Mailwasher ended immediately because the free version only
supports a single e-mail account and I have several (plus it was
overly crippled). I was only interested in finding other FREE
alternatives to SpamPal, and MailWasher Free was definitely not up to
the task.
I also use the MXBlocking plug-in which detects if the spam originated
from a dynamically IP addressed host that is typical of some user's
infected host running a mailer trojan. Real mail servers use static
IP addresses. Infected hosts and some hobbyists run their own mail
servers. However, this is an old plug-in and requires modification
that is beyond the typical user (you have to edit its config file to
point at an active list since the one it points at has been
discontinued).
SpamPal supports POP3 and IMAP accounts. However, SSL is not
supported so, for example, you cannot use it with Gmail which demands
that you use SSL to connect to them but you could add the sTunnel
proxy to give you SSL support with SpamPal. It would be far more
handy if SpamPal had included SSL support instead of making users
install another local proxy (SpamPal itself runs as a proxy unlike
MailWasher that runs as a seperate e-mail monitor). Unlike
MailWasher, SpamPal does absolutely nothing to your e-mails, like
deleting them, and instead merely tags them. You decide (using rules
in your e-mail client) what you want to do with the tagged e-mails.
You could simply delete all spam-tagged e-mails, or move them to a
[junk] folder, or even differentiate between those that got tagged by
a blacklist and those tagged by the Bayes filter. SpamPal can be used
as a install-and-go anti-spam solution but you'll probably find that
you want to tweak it to make it better at spam detection.
Obviously a server-side anti-spam solution is better than a
client-side anti-spam solution. The server-side solution is running
all the time and filters out the spam before it gets into your mailbox
so you don't have to expend your resources doing the same. If your
e-mail provider has an anti-spam option, make sure it is enabled -
unless they are very poor at detecting spam or generate lots of false
positives. For example, I enable the anti-spam filter in my Comcast
e-mail accounts because they are loose enough not to generate false
positives (good mails tagged as spam) but that does mean more false
negatives (spam that leaks through). AOL sucks at spam filtering and
I disabled it. Way too many false positives so I was losing good
e-mails. Users are far more sensitive to losing good e-mails than
getting a few spams leaking through per week.
