Comment Spam Blocker

Concept

Provide a service that matches IP address against comment spammers, and indicate known spammers to web forum software.

This service relies on a honeypot, or network of honeypots, to collect the spammer data. It turns out to be pretty easy to set up a honey pot, just by turning off any kind of controls on received messages.

The benefit to users is that the spam blocker requires no intervention on the part of operators to build the data set, unlike Mollum, which requires submission of spam to the Mollom service for processing and heuristic development.

The downside is that the comment spam blocker will probably not be as effective as mollum or other heuristic processing software.

One possible implementation is to make the list of offending IPs a downloadable list that can be downloaded every day (or maybe even every hour). A blacklist like that is easy to integrate into almost any config.

Minimally Marketable Features

  1. A working honeypot
  2. A data harvester
  3. Server to validate against
  4. A minimalist API for queries
  5. Client modules for at least Drupal.

Collection Methods

  1. For drupal, set up comments, only visible to logged in users, and no anti-spam controls.
  2. For dokuwiki, leave registration open and allow uploads to all logged in users.