New Downloads page established for my three projects: JDNN, Neural.NET and Slick.
New version of the Slick library posted, along with an example that demonstrates IPv4, IPv6 and Unix sockets networking.
SMF has been installed and setup to help facilitate communication. Please feel free to sign up and start posting your questions, comments and observations.
|
Jo Dene Neural Networks |
Neural.NET | Slick |
The TCP and UDP classes now work with Unix, IPv4 and IPv6 address families. The directory and buffer classes has been tested. The file class is being re-written and is currently broken.
My first shot at a neural network library was a .NET library I named Neural.NET. Based on the limitations of that first stab, I created the Jo Dene Neural Network Library (named after my wife). It is a stable, and somewhat decent neural network library written in Java. After that, I ported the Java library back to C#, but adding certain .NET features. The net result is JDNN and Neural.NET are about the same. I'm tired of Java, so I prefer to work with Neural.NET, but the choice is yours.
C is the lowest common denominator for writing Unix applications. Whether the device is a small, embedded PC with a hyper-minimalist kernel or a massive server, C APIs are always available. Even so, working with these APIs can be a little tedious at times. For example, writing network code requires a fair amount of nearly boilerplate C code to do simple things like manage inbound requests on a server socket. C++ is almost as popular as C, with the advantage of hiding this boilerplate code behind class interfaces. Hence, the Slick Application Library was born. It attempts to make multi-threaded, network-aware applications easier.