State system works well for a small amount of reducers but needs more thinking

[Tsarpf]

Notes:

• In the current system, all state reducers have to be run every time a state change occurs
  • if we knew which parts of the state a reducer is interested in, we could run them only when that part changes.
  • since the reducers are "pure" (for the lack of a better word), we know beforehand which parts of the state they're interested in --> listening to just those parts should be viable.

This is starting to sound a lot like https://github.com/Day8/re-frame