How do we handle errors? Race conditions? Re-tries? Anything beyond the boundary of our executing process is veering into distributed systems. From what I can

How do we handle errors? Race conditions? Re-tries? Anything beyond the boundary of our executing process is veering into distributed systems. From what I can tell that's exactly what a 'module' is. By weakening the contracts between modules and turning synchronous execution into asynchronous calls, you've taken on creating a framework / language that takes the difficulties of distributed systems and made them local, so every bit of important logic is a distributed systems call.