In the process of making the Python bindings for DungeonMaker, I came across various shortcomings of the original code. There were behaviours I wanted to limit or inhibit, and various features were missing. I went about implementing some of these, and code started accumulating. I eventually made the decision to fork the original source tree, and thus was born DungeonSpawn: a version of DungeonMaker with extensions.
RSAR treats datasets by removing attributes that are unnecessary for a classiﬁcation task. It performs greedy feature selection using various versions of the QuickReduct algorithm. It is useful in reducing redundancies in nominally-valued (i.e. discrete) datasets for exploration or as a preprocessing step to training machine learning algorithms on the data.
DungeonMaker makes complex random dungeons according to the designer's speciﬁcations. They can be as random or as deterministic as you want. The complex, organic look of the generated dungeons comes from the use of Artiﬁcial Life techniques: tunnels and walls are built by a-life creatures moving around the map. This is a modernised build of the library, packaged for modern distributions of modern operating systems.