Abstract
The Seattle distributed system is a framework for running python programs across many computers on the Internet.
Introduction
I've been interested in distributed scripting since I was a graduate student in The University of Texas at San Antonio. I was pleasantly surprised to see that there is a new python implementation of distributed scripting.
Seattle is geared towards creation of distributed network programs, but uses a restricted python interpreter, you can write arbitrary programs as well.
Implementation
I haven't seen the source of entire system, but there are three main components. Users run a seattle server and donate compute resources to the Seattle community. The Seattle clearinghouse arbitrates access to the donated resources, and manages reservations on one or more nodes (called "vessels" in Seattle). The devkit (available for download on your clearinghouse profile page) provides tools to develop Seattle programs and run on the vessels reserved for you.
Users are identified by public/private keypairs, and have access to a restricted subset of python called Repy. Several example programs are included in the devkit, and more are available on the Seattle website.
Repy limits the features available to programmers, but you can develop a program in another language and call repy to perform distributed computing, or use Repy or SeattleLib features to implement your program. Developer specific information is available online.