I got involved in the IEEE Micromouse project a few years ago at the University at Buffalo. Back then, we didn't have good tools for testing our maze-solving code, so I decided to build my own. The project started out super simple - draw a box that moves around - but quickly grew complex. Especially early on, it was tremendously difficult to stop adding features. After coding myself into a hole, I realized that the only way to finish the project was to rip out most everything I had added over the years. Nowadays, it does a lot less, but it's polished, and I'm comfortable calling it "done." Let me know what you think!
