Functional programming in Rust

I think that imperative code is easier to understand at an intuitive level, since, in real life, we don't deal much with recursion and immutable objects. However, it is very difficult to reason about formally and prove properties like correctness, complexity easier. On the formal reasoning front, functional wins most of the time.

So, all other things equal, I like a functional approach for formal problems like when building algorithms (most of the time). On the other hand, I find it much easier to use an imperative approach when trying to implement business logic and constraints as they pertain to users and the real world.

I agree mostly, but used in imperative code-bases, it can be a nightmare when it comes to debugging - at least, I find it harder to understand the division of responsibility. Yes, not having to define new classes saves time, but having fewer "categories" or "definitions" also means less structure - all functions blur into one.

Isn't writing functions that depend on other functions as input essentially the same as dependency injection, except with functions as opposed to objects? It seems to me like this would complicate things if anything.