distributed.net is a loosely knit group of computer users from all of the world that is taking up challenges requiring lots of computing power (most notably the RC5, DES, and OGR cracking contests). It is simple to participate in the challenges by downloading and running their client software (which uses idle CPU time to complete its tasks).
Botan is a crypto library written in C++. It provides a variety of cryptographic algorithms, including common ones such as AES, MD5, SHA, HMAC, RSA, Diffie-Hellman, DSA, and ECDSA, as well as many others that are more obscure or specialized. It also offers SSL/TLS (client and server), X.509v3 certificates and CRLs, and PKCS #10 certificate requests. A message processing system that uses a filter/pipeline metaphor allows for many common cryptographic tasks to be completed with just a few lines of code. Assembly and SIMD optimizations for common CPUs offers speedups for critical algorithms like AES and SHA-1.
Torch is a machine learning library written in C++ that works on most Unix/Linux platforms. It can be used to train MLPs, RBFs, HMMs, Gaussian Mixtures, Kmeans, Mixtures of experts, Parzen Windows, KNN, and can be easily extended so that you can add your own machine learning algorithms.
Calcoo is a scientific calculator designed to provide maximum usability. Its features bitmapped button labels and display digits to improve readability, no double-function buttons, undo/redo buttons, copy/paste interaction with the clipboard, both RPN and algebraic modes, two memory registers with displays, displays for Y, Z, and T registers, and tick marks to separate thousands. Calcoo is written in C using the GTK+ widget library (v.2.2+).
GLgraph is an interactive OpenGL based function grapher for Linux written in Perl. It visualizes any mathematical function in 1, 2, or 3 unkowns (x,z,t) in a 2D, 3D, or 4D wireframe or solid surface. It creates an animation after one time period. GLgraph has a command line interface to input a function, to specify the minimum and maximum plotting bound, and more. It can be interactively controlled with the keyboard.
The Maximum Entropy Toolkit provides a set of tools and library for constructing maximum entropy (maxent) models in either Python or C++. It features conditional maximum entropy models, L-BFGS and GIS parameter estimation, Gaussian Prior smoothing, a C++ API, a Python extension module, a command line utility, and good documentation.