It looks like a very nice and well thought beginning. Now we should wrap this into safer abstraction and maybe we should explore if it is worth to create a sup

It looks like a very nice and well thought beginning. Now we should wrap this into safer abstraction and maybe we should explore if it is worth to create a supervision mechanism and a communication model. Would love to explore this further, but my time is lacking, feel free to drop me an email for any kind of discussion :)