The idea :

A vertical SHMUP with procedurally-generated enemies undergoing natural selection.

What went right?

• Planning : I had a solid idea of where I wanted to go with the game from the start – I followed through with that idea to the end, only changing the cosmetics of it.
• Taking Risks : I’d never created a shmup, or implemented a genetic algorithm of any kind before. I didn’t fall back to my comfort zone when things got tough – I stuck with it and learned a lot in the process.
• Implementation : I worked out what kinds of classes would be necessary for the game at the beginning, and implemented/tested all of them before starting. This ground up approach helped me maintain a grasp on the scope.

What went wrong?

• Organization : It seems the deadline got the best of me, as towards the end I started throwing quick hacks into wherever I could to get things working. This resulted in a chaotic mess of unreadable code – code which probably could have saved me time working with if it had been more cleanly organized.
• Playtesting : I got so caught up in finishing the game that I forgot the most important part – making sure it’s fun. I played the game, but only for the sake of debugging – not in the mindset of a player. This resulted in very unbalanced gameplay, and a poorly-executed difficulty curve.

Also, I forgot about autotracker. Definitely would have helped, as I had “musicians-block” all weekend.

If you’re interested, you can play my game Modula here.