Anne is a very limited procedural programming language which nevertheless has some interesting features like native and language-level parallelisation and asynchonous execution constructs. It was designed as a front-end language to be compiled into ANNs (see the Wikipedia article).
ANNs have some interesting properties, and there are two main reasons for why it could be interesting to compile some code into a neural network performing "it":
I'd consider the project being in beta-state; it already works basically and as a proof-of-concept including some example projects to verify that evolution still works well for "compiled" ANNs and documentation about the found results, but there's still much work to be done to get some piece of software that can be released as a 1.0-version.
Originally this project was designed for a submission to the Jugendinformatikwettbewerb of the Austrian Computer Society, therefore the documentation was written in German; now I've started translating it to English, but it's still some part missing. Available documentation:
The current testing release is available for download at the Sourceforge project page, as well as more information and a description of anonymous read-only Subversion access to the development source.
I'll love to hear your comments, suggestions and ideas for improvements and whatever else as well! Feel free to contact me: Daniel Kraft