Mon Nov 11 20:55:30 EST 2019
a state machine language
Two important conclusions:
1. Clocked logic is NOT good as a base substrate. It needs at least
one higher abstraction, which is events.
2. This means the implementation language that sits below the
functional specification layer can be anything. Which also means Seq
is going to be good enough, and it can grow and change as long as the
interface is kept.
So how to move this forward?
Create a mapper. I have a real-world state machine that I can use to
test this, the DHT11 driver.
I can't "design" this. So I'll have to evolve it. Start with writing
out the state machine in the most straightforward Haskell form. Then
rethink from that vantage point.