katso is a library to score binary-categorized data: [low for anti-conformist data], [mean for conformist data], and [high for eclectic data: data that is quite normal (has a lot in common with other data but has a few unusual attributes)]. Data looks like (0, 0, 1, 0, 0, 1, 1) or like (1, 0, 0, 0, 0, 1, 0), all of the same size and corresponding to predefined criteria.
Whiteshoe is a text-based shooter game in which you explore a changing maze and complete your goals before the other players complete theirs or find you. It is based on the old game hunt, but was totally rewritten from scratch to use TCP packets and be more modular and expandable in design.
Syncless is an experimental, lightweight, non-blocking (asynchronous) client and server socket network communication library implemented in Stackless Python 2.6. It contains an asynchronous DNS resolver (using dnspython) and an HTTP server capable of serving WSGI applications. It aims to be a coroutine-based alternative of event-driven networking engines (such as Twisted and FriendFeed's Tornado). It is already about that fast, but it has fewer features and is less stable now.
MCM is a set of tools that ease the management of multiple servers. It's aimed at network or system administrators who need to connect to different servers by different means every day. It can be used via an ncurses interface without requiring an X server, and via a GNOME-based GUI.
MALODOS helps you to scan, store, and easily retrieve all your personal documents. Its storage format is open and documented, so your document archive can remain accessible even without MALODOS. The documents themselves are stored as standard PDF files, while their metadata (such as title, tags, and description) are stored into a separate SQLite database in an open format. With MALODOS, you can also manage existing files in PDF, JPEG, TIFF, and other formats, so you can still use the documents that you've already scanned. You can connect to any external OCR program to give access to a fulltext search feature.
pycdep is a tool for analyzing and visualizing C/C++ header file dependencies. A script extracts information about your code and dumps it in a Prolog database, which is used to query the dependency graph. It comes with a large number of predefined queries for finding out which files include or depend on which other files, how files depend on another file, circular dependencies, transitively implied dependencies, and design rule checks (limitations on which project can include from which other project). Visualization is supported by dumping graphs to .dot files. It also has the beginnings of a natural language interface.