Interesting. It's like the Refactoring phase of TDD. So basically during the day you experience new observations ("tests" if you will), you address/process them however you can on the spot ("green bar"), then during sleep you take stock and generalize so you can reuse the experiences of the day for the future ("refactor").

I've been seeing this TDD process in a lot of things lately. Scientific progress seems to follow the same process as well (at least that's one way to capture what Popper or Lakatos have been telling us, not that I've actually read them yet mind you).