Comment by dantodor

Honestly, I don't know enough Clojure to understand the proposal, and all the explanations are centered around it. My current understanding is that what you propose is an ascribed way of doing things by mixing several pre-existing concepts, like mono-repo, convention over configuration, and some generic dev principles that are universally desirable, like components "ensure encapsulation and composability but at the same time they are simple, easy to reason about, and flexible" - nothing wrong in that, all good principles, but I fail to understand how you enforce it. Because usually every monolith or micro-service out there starts with that in mind, and ends with spaghetti, because dev happens. At least that's my experience with Elixir/Phoenix umbrella apps, which, as per my current understanding, is the closest thing I see to Polylith. However, I would be very curious about your traffic lights evaluation. Generally, I would consider a "large" app something like 200+ microservices (yes, I'm biased towards that). In your FAQ, you describe and app that needs 4 services. Sorry, that doesn't qualify as a large distributed system. First, I would really doubt you can find a monorepo which is one language only (sorry, I didn't understood how Polylith would work if I have a codebase that's Scala, Elixir, Rust and Python, by example). Second, I don't understand how you handle builds. Third, I don't understand how you handle deployments and other operational concerns.