Status

Done:

• NAL 1-8 reasoning supported, and the corresponding Narsese.

Future TODO:

Software-side:

• Explore lockfree datastructures further to make multithreading more effective.

• Show demos in a more fancy way in a separate repository, like Christian Hahm did: https://github.com/ccrock4t/NARS-Pong but don't rely on external libraries to show basic system capabilities in the main repository.

• Robotic experiments: they are becoming a key now as we believe autonomous robotics is where a reasoning system which can learn in real-time can make a difference. We are collaborating with the Palermo team in this regard already.

AGI side

• Increase inductive abilities by incrementally removing restrictions, considering Relational Frame Theory etc. (also see https://groups.google.com/g/open-nars/c/Kn3BQoIe5Is )

• Make NAL-9 effective, add reliable self-control aspects, which includes the learning of mental behaviors. (this will be for v0.9.x)

• Add more sensory channels and in a way that it will become easy for people to hook up ONA to a camera/microphone, depth camera, GPS and other sensors.

• More NAL features will be added incrementally in the following releases when they turn out to be application-critical.

Documentation side

• Improve and document the code with a literate programming tool.

• Remove all code or feature which isn't absolutely necessary, KISS is key.

• Create a Github pages site showing the system capabilities nicely presented, also summarizing results from the aforementioned example repositories.

• Update Wiki according to the system architecture publication in case there is an inconsistency.