Know When to Stop

I get myself into these situations once in a while where I get really excited about a particular problem and end up with a mess as the solution.  Here's my workflow.

1. Think about it.
2. Sketch out a solution.
3. Code towards my solution.
4. Refactor (go to step 1)

Somewhere in that loop I go from 4 to 3 to 4 to 3 a bunch of times and when I come up for air I have a complete mess.  Ideas pop into my head quickly and I seem to go down strange paths here and there. Before I know it my code is littered with random design thoughts that don't hang together in any way.

STOP!!!

... git here we come ...