It would seem that everyone on planet Earth has received spam. In fact, it is probably not unrealistic to think that spam has been somehow transmitted across the wireless waves in space to our yet undetected cosmic neighbors who can only conclude that we are obsessed with online meetings, free prescription refills, and miraculous growth enhancers. Indeed, the Internet is as much a place of dodging unwanted material as it is searching for the content of choice.
No network administrator is unfamiliar with the pains associated with spam filtering. Many enterprises turn to a third-party solution out on the perimeter, and yet others opt for a separate service provider who will accept and filter all their mail for them and then relay mail to their organization once it has been sanitized. Few organizations rely solely on client-side applications.
Often, a layered approach to battling spam is what is required, one that utilizes each of the avenues through which spam can make its way into the organizations e-mail system. As such, the client-side fight will always remain; spam or virus-generated mail could come from within the organization, or from a visiting consultant or a poorly protected laptop that is brought into the company. So, taking yet another step forward in improving the e-mail experience for former Outlook Express users, the Windows Mail design team has integrated a Junk E-mail filter into Windows Mail.
Contrary to popular belief, the Junk E-mail filter built into Windows Mail is not simply another Bayesian filter. The algorithms that compose the learning engine for this product are more sophisticated than that. In fact, this product is not a client-side application; what ships with Windows Mails Junk E-mail filter are filters derived from the Microsoft SmartScreen spam-filtering solution that has been refined over the past few years at the enterprise level.
First deployed to its Hotmail servers, the SmartScreen technology leveraged learning algorithms against large sets of data (hence Hotmail and eventually Windows Live Mail). The design team that is focused exclusively on the antispam effort at Microsoft found higher levels of learning with a broader user base and deeper concentration of mail. Based on user input, the algorithms adjust their detection settings and then generate filters accordingly. This approach has provided SmartScreen with such a high success rate at the identification of spam (95 percent, according to Microsoft) that the product was eventually incorporated into the Exchange 2003 and Outlook 2003 products, as well as their respective 2007 versions.
Because of the nature of the algorithms, the application does not learn at the client level, that is, no personal settings or adjustments are made to tailor the filtering to a users preference. To do so would be to take these finely tuned enterprise-level algorithms and point them at a ridiculously small subset of data, which would essentially stunt the learning process. However, like in Outlook, the Windows Mail Junk E-mail filter is set to pull filter updates from Microsoft via the Microsoft Update service. In this way, the filter settings in use for Exchange are at work on your local machine from the first use of Windows Mail.