The Assimilation Monitoring Project is a highly scalable discovery-driven monitoring system. It integrates continuous discovery of servers, services, service dependencies, switch connections, and lots of other things into the monitoring process. The discovery is "stealthy" and will never set off any network security alarms. Adding servers doesn't measurably increase monitoring load, and the system is expected to easily scale into the 100K server range. The discovery work is distributed among all the nanoprobes (agents), which run scripts that spit out JSON. The central system (CMA) stores these strings and runs optional plugins to create graph nodes.
pycms is a minimalist component-based Web Content Management System based on CherryPy. It has most of the goodies of Plone/Zope: Python instead of a PHP security mess, no MySQL can of worms, clean separation of code and content, and the ability to run nicely behind a Web server/cache/load balancer. In addition, it features plain text storage of all site contents and settings in JSON files, effortless theming, and a small memory footprint.
planes is a hierarchical surface framework for PyGame. It introduces a "Plane" class which extends PyGame's "Surface" functionality, offering a hierarchy of mouse-sensitive, draggable Surfaces that can be used as sprites, windows, icons, or other interactive elements. In addition, the planes.gui module uses planes to provide basic GUI elements like container, label, text box, option list, and button.
Speedpad is a small and portable ncurses-powered tool to test, train, and increase typing speed on arbitrary text input. It is designed for intermediate-to-advanced level typists and assumes that you have already learned how to touch type. It does not use lessons, single words, or other synthetic stuff. It supports tab expansion, auto indentation, and syntax to train on code. It features a reference speed robot and supports CPS, CPM, WPM, PPM, and CPH/KPH metrics. It shows detailed statistics about speed and helps find and eliminate frequent typos. Stats are dumped to standard output in a machine-readable format after completion, and can be piped into gnuplot.
Backshift is a deduplicating (variable-sized, content-based blocks), compressing (xz or bz2) backup program. Full saves and incrementals are pretty indistinct other than the amount of data transmitted, somewhat like with "rsync --link-dest" but without the huge number of hardlinks. It also de-duplicates large file content at a granularity of about 2 megabytes on average; there tends to be a unique copy of each file with size less than around 2 megabytes on average.