Peachpit Logo (bevel)

Peachpit -- a Tarpit for Censorware

Peachpit is an automated tarpit utility intended to get in the way of the spiders employed by censorware manufacturers.

"Tarpitting," briefly, is the tactic of slowing down a TCP connection when a hostile party's autonomous agent is on the other end -- an approach originally conceived to deal with spammers, though in principle it should work well with most any known-hostile web spider, to some degree.

Censorware is a (harsh) term for the content-based web filtering mechanisms which attempt to filter out objectionable content, most particularly of a sexual nature or espousing unpopular political views. Unfortunately, the censorware manufacturers have, despite all claims to the contrary, relied extensively upon the services of automated spiders to find potentially objectionable material, which they then add, without verification, to block lists. Blocked material generally becomes inaccessible at all facilities who have purchased the software -- including public institutions (schools, libraries, etc.) where such software is in violation of the First Amendment rights of citizens, students, etc.

A more detailed discussion of the software and ethical issues may be found in the README. Further discussion on censorware, including useful background, can be found at, among other places. Before using peachpit it's important to consider the ramifications thereof, and reach an informed conclusion. My own discussion of these issues is in the README.

Tarpitting can cause problems for legitimate spiders, so peachpit provides flexible pattern-based spider recognition to help ensure that only targeted censorware spiders are affected, without adversely affecting other spiders.

Peachpit is an adaptation of a Sugarplum, an automated spam poisoner utility with similar features. Peachpit and Sugarlum are both written in Perl, and released under terms of the GPL.


Monday, 3/27/2000: First public release. While functional, this is intended as an initial solicitation of improvements, especially to the (currently thin) censorware recognition code -- currently the provided rules recognize only the BirdDog spider used by BESS, and X-Stop's Mudcrawler. Further work is also needed to isolate hostnames/netblocks from which the spiders originate.

Devin Carraway <peachpitA@TdevinDO.Tcom> (RSA 6E479901, DSA E9ABFCD2)