HTM.Pony

Status: a small experiment, work in progress
based off the Go implementation: https://github.com/htm-community/htm (MIT License)
no direct dependency on other implementations

Rough Plan

Start with porting the Go version to Pony using classes at first
Once an opportunity arises to parallelize, start experimenting with actors & benchmark
When the Temporal Pooler is ported, align the implementation with the latest NuPIC/htm.java details
No attempt at proactive refactoring is made at the moment. When a rough working version is made, it might make sense to reimplement whole classes/actors
Error handling is ad hoc currently, and will need to be designed once there is a high level executable that uses the library. Trade-offs: errors, logging & partial functions vs explicit union return types
continue

Open or take an issue for the next small problem to solve, and open a PR if there's some code to be merged
If there are tests and these pass in CI, the maintainers should merge the PR into master
After a successful PR, the contributor may become a maintainer if so desired
Should merge conflicts arise, the master will be protected, all work will be done via PRs
Long-lived branches are to be avoided. All progress should go into the master branch as soon as the tests are green
If there's significant traction, we'll switch to the ZeroMQ process C4 and a protected master

Sparse Binary Matrix, Dense Binary Matrix: Go tests translated & running. Untested Go code left for later
Initial implementation ready: Scalar Encoder (#1)
Next: Date Encoder (#2), any other chunk with minimal dependencies