does Polylith have a recipe/opinion for communication between modules? I have the vague impression that Poly is agnostic wrt this topic, which while makes sense, might also be surprising when scaling out a modular solution. In specific terms, if for the first couple years of your app components talked to each other synchronously via function calls, then introducing an async and/or IO-heavy means of communication (Kafka, w/e) could be a tough challenge, while one expected the architecture to precisely be eventually scalable without a huge refactoring effort. Probably there's an overlap between 'modular architectures' (Poly, https://github.com/puppetlabs/trapperkeeper also comes to mind) and 'functional architectures' that seek to make async, io, timeouts/failures more structured. So I was wondering if people are trying to compose various ideas into one holistic solution, what are they using, etc.