[strictfp]

Well, sort of. I have established a habit of thinking about my problem during workdays and planning actions for my commute. During the commute I execute one or two planned actions.

This works really well for me, since I easily get overwhelmed if I just sit down without an action plan.

But occasionally I use longer sessions to cleanup, refactor and review my code and assess the current state of the project.

Sone days I just read up on some subject instead of coding. But there is always a risk in getting bogged down by resesearch until the point where you can think of 20 possible solutions. I then sit down and code a bit to help me decide which method would work best for my concrete case. You can always come back and reevaluate a decision if you discover problems further ahead...