Because it sets what is stored in `state` it does not have to be. It could be inside the if, but there two major reasons why it should not: 1) Testability. Wit

Because it sets what is stored in `state` it does not have to be. It could be inside the if, but there two major reasons why it should not: 1) Testability. With it separated, we can write automated tests for our state calculation logic that are hardware independent, like 'normal' code. We just need to create `currentState` data, pass it into our `calculate()` function, and verify the new `state` returned. This can run both on our host-system, and on-target. We can even build simulations if we want, http://www.jonnor.com/2017/03/host-based-simulation-for-embedded-systems/ http://www.jonnor.com/2017/03/host-based-simulation-for-embe... (shameless plug). 2) Deterministic execution time. Many embedded systems have soft or hard real-time constraints, action must always be performed within X microseconds. The more branches we have the harder it is to reason about whether we can keep this guarantee. By eliminating branches we are executing our 'worst case' always, giving it much more testing.