One of the biggest problems I'm finding as a developer so far, is that darned deadline!
Let me start at the very beginning, if I may. Every single one of the games I make starts off with a prototype. I'll spend a few hours building a simple version of the game, just to see if it works and (most importantly) if it's any fun. If it's not, I'll file it away (perhaps never to be looked at again). If it has potential, but I'm not quite sure how to proceed with it, I'll put it on the backburner. That's what happened with StarLight.