Mandatory Spam Filter Tested

Roughly 61,000 e-mails containing everything from mortgage offers to penis enlargement proposals—along with a few real e-mails—were intercepted Wednesday under a new Faculty of Arts and Sciences-wide spam filter that began testing last Monday.

In a new effort to fight internet viruses and spam, FAS computing services is testing a mail filter that does not permit users to opt-out and that blocks certain file extensions for attachments on all fas.harvard.edu accounts.

“The viruses and spam take up people’s time and take up space on the system so [the new measures] are good for the clients and good for the system,” said Kevin S. Davis ’98, the director of residential computing.

The attachment extensions .bat, .cmd, .com, .cpl, .exe, .inf, .pif, .scr, .shs, .vb, .vbs, .wcs, .wsf and .wsh have all been blocked, and any e-mail with an attachment containing one of these file extensions will not be delivered, as these extensions have the highest risk of carrying viruses.

Computing services also began a two-week test of the “widely used and respected” anti-spam filter Spamhaus last week, said Davis, who is also a Crimson editor.

Though the new program is limited in scope—it does not protect against worms, which are technically not viruses—thus far response to the new measures has been mostly positive.

“We’ve been really thrilled with the results and reactions, we’ve had an enthusiastic response to the test,” Davis said.

Davis admitted that there have been some reports of “false positives” or e-mails that were blocked even though they were not spam.

“Any anti-spam system has false positives just as there are things that are spam that get through,” Davis said. “There is no optimal solution.”

Legitimate files which have the banned extensions may still be attached and sent as long as they are zipped or the file extension label is changed. Complete instructions on how this can be done are found at www.fas.harvard.edu/computing/block.

As the summer progresses the computing staff will be testing a number of other alternatives to the spam problem, Davis said, after which a final decision will be made on whether or not to run the program during term-time.

“At this point in time we’re still in the test phase and evaluating a number of different solutions,” Davis said. So far people are reporting receiving between 50 and 90 percent less spam, he added.

—Staff writer Joshua P. Rogers can be reached at jprogers@fas.harvard.edu.