Catdoc is a MS Word file decoding tool that doesn't attempt to analyze file formatting (it just extracts readable text), but is able to handle all versions of Word and convert character encodings. A Tcl/Tk graphical viewer is also included. It can also read RTF files and convert Excel and PowerPoint files.
radare2 aims to create a complete, portable, multi-architecture, Unix-like toolchain for reverse engineering. It is composed of a hexadecimal editor (radare) with a wrapped I/O layer supporting multiple backends for local/remote files, debugger (OS X, BSD, Linux, W32), stream analyzer, assembler/disassembler (rasm) for x86, ARM, PPC, m68k, Java, MSIL, and SPARC, code analysis modules, and scripting facilities. It also has a bindiffer named radiff, base converter (rax), a shellcode development helper (rasc), a binary information extractor supporting PE, Mach0, ELF, class, etc. named rabin, and a block-based hash utility called rahash. Radare was rewritten as radare2, and the old version is only maintained for bugfixes.
mDNS Responder with Unicast runs on a server anywhere in a network and responds to mDNS queries across the network by performing a unicast DNS lookup and returning the result. Many networks use a .local top-level domain for their intranet. This has proven to be quite problematic as operating systems such as OS X have begun handling .local domains differently (in particular, prioritizing multicast DNS over conventional unicast DNS). Ideally, you'd have no problems moving away from a .local domain name, or simply not using one in the first place. That's not always possible, though. So with this tool, a lookup for "host.local" will be resolved via your existing unicast DNS servers, even if "host" does not have its own mDNS responder. Of course, the tool isn't limited to just .local domains, and can realistically work on any network where mDNS is in use.
BNNS is a research tool for interactive training of artificial neural networks based on the Response Function Plots visualization method. It enables users to simulate, visualize, and interact in the learning process of a Multi-Layer Perceptron (MLP) on tasks that have a 2D character. Tasks include the famous two-spirals task or classification of satellite image data.