Block by referrers project

I am getting way too many hits from bloody casino, poker, and drug (the pharmacy type) sites.

Typically, they come in, look for the comment forms, and try to submit comments to the site. So I – and anyone else who has a blog and doesn’t want it clogged with comment spam – have to disallow posting except by members, automatically kill any comments that contain words like poker etc., and still have occasional hassles with garbage posts.

Not to mention the bandwidth and server time that these requests burn.

Somehow, there’s got to be a nice clean way of blocking unwanted visitors. Here’s a typical example of my referrers for the first few days of July 2005. I seriously doubt that any decently updated blog will be any different.

Naturally, I’m aware that you can block people or agents coming to your site by IP (like this guy), but that’s laborious. I’d have to check each of these referrers, run a traceroute, and enter the IP in my block file.

That’s too much work, and it’s too draconian. What if 2 months later the person or entity with that IP address is kosher? I’m very unlikely to follow up and check my blocked IPs months and years later.

We, the web, and bloggers in particular, need an open and collaborative system of identifying comment spammers in real-time and pushing a common list of them to all the nodes (blogs) on the network. A P2P model such as BitTorrent could be a workable archictecture, with no one central node from which the entire system can be controlled.

Each node would need to run a local copy of the software which would analyze incoming requests, flag suspicious ones, and report the list to other nodes at regular times. The software should be able to be run in PHP and Perl, and interface with common blogging software such as WordPress, Moveable Type, pMachine, Textpattern, etc.

Anybody willing to take this up as an open-source project?