Well I wrote a very basic system yesterday evening. But as long as I don't have a good (gameplay) idea, I probably won't continue it so I don't waste too much time on it.
Although I like to experiment with game(play) prototypes, I often think it's a better idea to think about it and visualize it in your head first. I don't say that you can know what is fun before you've actually programmed it, but I'm sure you can say what's not fun to play when you only hear or read the idea/ concept.