Important branches

Up-to-date branches:

cpp: minor code modifications for the ability to compile ONA with C++20 compiler.

portability: removes POSIX dependency altogether, with scripts to compile with C99 or C++20 compiler, or to compile to web assembly for execution in browser or on Ipad shell app.

Forks:

"This is a fork of the ONA code with some small changes, a Python REPL and a Perl based HTTP daemon for querying ONA via REST/JSON." - Seamus:

https://github.com/Prediction-by-Invention/onars

Outdated branches:

MSC3: Minimal Sensorimotor Component (NAL-7/8 inference rules only, sufficient for procedure learning) (can be emulated by simply setting SEMANTIC_INFERENCE_NAL_LEVEL=0 in master)

QLearner: Q-Learning wrapped into NAR module for procedure learning comparison without need to alter examples. (This was for dissertation purposes, which compared ONA v0.8.7 with Q-Learning)

Numerics3: Numerical terms and inductive inference introducing plus/minus functions to describe numeric change, plus numeric argument babbling for operations (experimental)