Virtually anyone with an email address knows what Spam is, and
has, perhaps, considered giving up the speed, convenience, and
simplicity of email because of it. Those who have their own
websites are more vulnerable than the average person with a
single work or home email address from their company or Internet
Service Provider. Email addresses visible on a website can
quickly become Spam magnets, as automated programs, similar in
form to search engine spiders, roam the web, looking for
addresses to which new broadsides of Spam may be fired. Website
hosting companies generally provide their clients email accounts
for use with their domain, but are you with a host that provides
those email accounts with Spam and virus protection?
Local Blacklist Filters
Webmasters shouldn't need to
seek out local filters for their site's contact email addresses.
There are a variety of server level solutions a hosting company
can offer to protect their users from unwanted emails. A most
basic step is provision of a very rudimentary "blacklist"
functionality to their users, allowing them to prevent future
Spam emails from arriving from the same address. This type of
filter is virtually worthless in today's Spam environment,
though, as it is quite rare to see unsophisticated email
arriving from the same address multiple times. Spammers have
grown far more sophisticated than that. Blacklisting
functionality is only really useful in avoiding email from other
real people you don't particularly wish to hear from anymore.
Keywords and Regular Expressions
More advanced server
level Spam filters are available. A small advance is
accomplished using keyword filters. Keyword filters merely check
for instances of a certain string of characters and deny the
message if that string if found. The core problem with
keyword-only filters is they can "over filter". Someone who puts
"sex" on their keyword filter will find receiving local news and
event announcements difficult if they live in a town named
"Essex". Some filters attempt to address this deficiency by
using "regular expressions" in order to build a sophisticated
rule set to prevent Spam from reaching your inbox. Briefly,
regular expressions are syntax rules used to identify certain
strings of text or numbers. These rules can be set up to
identify text patterns that are commonly used in Spam. They can
become quite complex, but, as with most any filtering method,
are not 100% bullet proof. Some filters that use regular
expressions come with a basic set that can be appended by the
user. Obviously this kind of feature is of little use to someone
not familiar with regular expressions.
Bayesian Filters
Currently the most sophisticated
filtering methods use Bayesian inferences. Bayesian filters take
a large data set and determine the probability a message is Spam
based on its similarity to previous Spam messages. The more
emails that are processed and flagged theoretically make the
filter more accurate. Services that provide filtering on an ISP
or host level, like Postini's "SpamAway", filter billions of
emails and provide the highest level of success and fewest
"false positives". SpamAway is already highly intelligent about
identifying Spam and doesn't require any "learning" commands or
examples be provided. The online, browser based interface keeps
flagged messages in an easily accessible "quarantine" and allows
the user to check for any false positives. White list
functionality is provided to aide in the prevention of future
false positives. A hosting company offering such an advanced
service takes Spam and virus filtering for their customers
seriously.
|
