My rant about spam last week drew several interesting responses: The CEO of MailFoundry offered to send me an email filtering appliance, a 1U server that goes in front of the email server and claims to have no false positives A guy hight Ronald said that he had successfully built a front-end filter that rejects emails with the network signatures characteristic of botnets before they ever get to the mai My rant about spam last week drew several interesting responses: The CEO of MailFoundry offered to send me an email filtering appliance, a 1U server that goes in front of the email server and claims to have no false positives A guy hight Ronald said that he had successfully built a front-end filter that rejects emails with the network signatures characteristic of botnets before they ever get to the mail server, leaving him with only 1-2 spams a month Randy Bruckhardt, the CEO of RR Software, said he’d built a spam filter application for his company’s email server that combines a number of techniques, and as a result lets only 3 spams a week through out of about 200,000 attempts, albeit with about 40 false positives a week.None of these do me any good with my current setup, but they all have something in common: they filter the bulk of the email as early as possible at the receiving server. This should be more widely available.I’d like to see an open source project to build a reliable email pre-filter along the lines of Ronald’s. Ideally, it could either run on the same box as an existing email server, or on a separate box, depending on the volume of spam being filtered. Ideally, it would have Linux, Windows, and Mac versions, and would be easily monitored, updated and tuned. I’d also like to see an open source filter project along the lines of Randy’s. Maybe it could be combined with an existing filter like SpamAssassin on Linux, but I’d also like to see Windows and Mac versions.Still, that isn’t enough. The spam problem should really but cut off at the source, to avoid tying up so much Internet bandwidth and so much in the way of computing resources for filtering. I’d really like to see a credible effort by ISPs to block botnets from sending spam. It shouldn’t be hard: shut down the outgoing mail port on any connection that hasn’t demonstrated that it has a legitimate, properly secured mail server. For good measure, shut down all incoming IRC ports unless the owner of the account asks for them to be opened (ask by voice, not email) and demonstrates the presence of a working two-way firewall: that will block the herders from sending commands to their bots.Is that so far-fetched? My final solution probably is far-fetched: find all the spammers and put them out of business. I know that a few of the more notorious US-based spammers have been shut down, but spam has gone up since then. Many of the remaining spammers are out of reach of US authorities. What would it take to get to them, I wonder?Do you have anything to add? Software Development