What

Here you will find a repository of scientific libraries and programs written primarily in Java covering a wide berth of topics. In making these tools available to the public, the author hopes that they will save you time from having to write your own.

Currently, the following projects are included:

• The Shared Scientific Toolbox, a library for scientific computing. It features a multidimensional arrays package, a high-performance networking package, an advanced class loading framework, a dataflow engine for multi-core, and JNI bindings to the highly-acclaimed FFTW library.
• Dapper, a large-scale, lightweight distributed computing environment built as an alternative to systems like Apache's Hadoop and Microsoft's Dryad. It provides configuration-free setup, rich and robust semantics for dynamic modification of the dataflow graph at runtime, and a user interface for dataflow visualization and management.
• Microcrystals, a machine learning system for annotating images of crystallographic trials. It uses the boosting algorithm to combine hundreds of marginally discriminative Gabor wavelet-based features into a single, accurate classifier. The system achieves a low false negative rate when ranking images based on their likelihood of containing crystalline material. A user interface and database schema accompany the machine learning brains.
• Miscellaneous utilities like SVNSync (not to be confused with the official Subversion tool of the same name), which generates the necessary "svn remove" and "svn add" commands on a destination working copy so that it remains synchronized with some source working copy.

Why

High quality scientific code is important in that it can lead to extensive reuse, while at the same time serving as a vessel for conveying research ideas.

How

Feel free to peruse the Subversion repository here. Alternatively, you can follow the links to the right to visit constituent project pages.