One thing that has helped me also as I'm figuring my way through this is to leave time to just "play". But the play has to be related to the problem. In particu

One thing that has helped me also as I'm figuring my way through this is to leave time to just "play". But the play has to be related to the problem. In particular, this is suitable for ill-defined research problems (the type I am frequently working on these days, vs. more standard engineering-based implementations/refactor of yore). And then as you play with the problem a direction kind of magically opens itself up and then you want to explore further.