The anti-spam vendor was truly positive that we must have gotten something wrong in assessing the rate of false positives generated by its product in the most recent Network World "Spam in the Wild" test.

If you missed that test report amid the holiday hubbub and no longer have the Dec. 20 issue hanging about, I strongly recommend you give it a read. Network World Lab Alliance member Joel Snyder, a senior partner at Opus One, conducted the test - and fielded the inevitable complaints.

"Of course, many of the vendors who had poor false-positive rates in the anti-spam piece insist that this is impossible and that they never lose mail," says Snyder, who offers as evidence the following exchange of e-mail with the absolutely positive party above (left nameless here out of a combination of compassion and our inability to be absolutely positive that the vendor's product, as opposed to some other glitch, caused the ensuing problem).

"One, in particular, wanted the log files off the test server so that they could analyze what on Earth might have happened," Snyder says. "So, I sent the log files. But they reported not receiving them! So, I sent the e-mail log showing that they received them, the date, time, message ID and IP address of their mail server, and then they say 'please, please, we can't find it, please send it again.'"

"So, I sent it again. And they still said they didn't get it. So, I sent another e-mail log showing how the message was sent yesterday to two people at the company and received by their mail server, with all the relevant details. . . . I'm wondering just how long it will take before they're too embarrassed about their own false-positive rate to keep asking for the log files."

Personally, I've grown tired of rolling my eyes at anti-spam vendors that say their products absolutely, positively never spit legit e-mail into the spam bucket. But the good news, according to Snyder, is that the better vendors are coming much closer to delivering on that promise: In our September 2003 test, even the top scorers misplaced up to 3% of legitimate e-mail; this time there was "significant improvement, like a factor of three."

Please, sir, send me a bill

You may have read recently that The New York Times is considering charging a fee for what is now free access to the newspaper's Web site, which by any measure has to rate among the greatest bargains available on the Internet.