First-Order Probabilistic Inference package in Java from
this project that lets you specify and query probabilistic models
in a first-order language.
C++ Machine Learning Template Library:
a collection of templatized machine learning algorithms that can be used on any class providing
a minimum example interface
(which arrays and maps provide by default).
Algorithms are decision trees, backprop neural nets, winnow and perceptron.
Also contains utility classes for user-defined algorithms and data.
DetailLogger in Java:
extends the log4j package to produce hierarchical debugging logs,
which can be navigated with Emacs or Windows-based log visualizer Squirrel.
a Perl script generalizing the Unix mv command allowing for pattern- or code-based destination generation (that is, you can specify a pattern or code that generates the new name for each of the files you are moving).