Another way?

SpamButcher has a huge internal database of phrases and their relation to spam email. I would estimate that the database has increased in size by about 50% over the last year. We've definitely added more code to the product in the last 12 months than in any time since SpamButcher's original release.

We have often touted that the use of fuzzy logic allows us to go longer between filter updates, but this isn't nearly as true as it used to be. We now update the filter weekly or more often to keep it in tip-top shape. This takes a lot of work.

So, I've been wondering lately if there are ways to block junk mail that could result in a similar level of effectiveness using other less update-dependant methods. There are two technologies that SpamButcher uses, but doesn't fully utilize which seem to offer promise for further development.

While SpamButcher doesn't by default use any DNS blacklists, it does have an extensive internal list of problem networks. Users can opt to use a blacklist, but we leave it off unless they specifically enable it. In early versions, SpamButcher actually used multiple blacklists by default, but we hit a serious false-positive some problem with one of them. Since we got burned, we've been hesitant to turn it back on unless requested by the user.

There's no question blacklists work. Specifically, the CBL is capable of blocking a lot of spam. I'm not sure of the exact percentage - but I suspect it's close to 75%.

SpamButcher is also aware that certain countries send huge amounts of spam (Korea, China - I'm talking about you!) - and blocks most email from them. I feel this is safe to do - as we don't sell SpamButcher to these regions.

Further, a lot of people don't exchange unplanned emails with people outside their own country. For instance, someone in the U.S. might block most of Europe. Vice-versa, someone in Europe might block the U.S. and Canada. Addresses could of course be whitelisted to allow for exceptions.

I doubt these two combined approaches would get close to SpamButcher's roughly 98% kill-rate - but it may be good enough to exterminate 90% of spam or so. That may be good enough to "productize."

Despite the heavy workload it requires to maintain, there's no question that SpamButcher works incredibly well. We'll continue to make SpamButcher the best spam killer possible, but maybe there's the opportunity to develop a secondary product.

Back