Filed Under (Geekspeak, Work) by Justin on 2007-05-19

Tony has a post about anti-spam, specifically what the three worst cases for filitering spam.

The third worst thing: having spam get through to users. Of course, this is what we’re trying to fight, but having spam get through, although bad, isn’t the absolutely worse thing that can happen.

The second worst thing that can happen is blocking a good email. False positives are the challenge in anti-spam solutions. I would much rather have hundreds of spam messages get through than to block a single good message. But blocking a good message isn’t really the absolutely worst thing that can happen.

The worst thing: The absolutely worst thing that can happen in your anti-spam solution is to block a good email and not let anybody know about it!

I’ve been getting more and more complaints from my end-users at work in regards to junk in their inbox, and regardless of how they managed to subscribe themselves for it, it is still a problem. In my almost three years there, our policy has always been “Let the end-user be their own filter” rather than us filtering it on the server end of things. I’ve been pondering lately turning on our spam filtering service within Ipswitch, simply for logging purposes for a few weeks so I can gather some stats and not change a thing on the user end of things for now. After collecting data, I could make a judgment on whether or not something needed to be done. I probably don’t need to get too cozy with it though, as we’re planning to move our mail to a Linux box later in the year and that would just be another feature to have to implement.

(4) Comments   


Tony Dye on 20 May, 2007 at 8:58 pm #

Have you considered a solution like spamassassin? You could implement that on a low-cost linux box as one step in your email pipeline. Even if you don’t spend much time tweaking it, you could easily take out a significant portion of the worst spam. Of course, something like Postini or Katharian would be a great no-hardware solution.

Peter Brockmann on 21 May, 2007 at 7:56 am #

Frank Urro on 21 May, 2007 at 6:06 pm #

The truth is – spamassassin will make a mistake now and again, as will Postini, or Katharian. And yes, even Challenge-Response will sometimes fail as the Challange message itself could be blocked by the recipients anti-spam technology.

The truth is, I have become more annoyed by my outbound messages not be delivered due to the recipient’s A/S.

Additionally, I can not say that my product, even though it has been awarded PC Mag Editors choice is immune from false positives. Sounds a bit strange that I would be sharing this with you, but the truth is, there is no technology that can guarantee 100% accuracy.

So, that spammers are dedicated to confusing filtering and that they are not going to give up anytime time soon, we need to make a bold move to address the root of the problem.

Now, the question is – what is the root, and what is the bold move that needs to take place to eliminate disrespectful undesirable personal interruptions?

This is what I explore on my blog –


Justin on 21 May, 2007 at 9:52 pm #

Peter: I despise Challenge-Response spam systems. I never respond to them, therefore I can’t expect folks that my company corresponds with to. As a matter of fact, I believe that Challenge-Response anti-spam systems fall under Tony’s second and third points.

Frank: I actually disagree with you to some extent. If someone is trying to stop all spam, then yes, they will surely set the filters to be too aggressive and it will result in false-positives. For my situation, I would actually be pretty happy with eliminating even half of the crap.

Post a Comment