Thanks Ludum Dare and 1GAM, for encouraging me to put out more games … even if they are shamefully dull and unplayable. I had been working on a game about operating an underwater base but it was taking me too long to figure out what the game is actually about. I mean, okay, yes, it’s about an underwater base, but the game I want to make which is part economic sim and part micromanaging the staff was again reaching beyond my February grasp. Instead I cranked out a quick idea I had about manipulating Conway’s Game of Life. The game, Death to Conway, is a turn-based simulation of the Game of Life, and you’re given an opportunity to kill an extra cell in between generations. Excitement? Thrills? Who needs em! We’ve got cells, lots of terrible cells, and they must all be killed!
Play Now (in browser)
This was another HaxePunk joint. HaxePunk is sort of in flux now as the developer is making some additional changes to better support hardware acceleration on native targets, so I rolled back to HaxePunk 1.72a to build this one.
I don’t know why I have to code everything twice. I think it’s a failure to personally commit to my own requirements. I knew I wanted to have the ability to run the simulation separate from the display, but since I didn’t commit to it at the start, I tightly coupled the simulation with the rendering system. Then of course I get the game working and realize no one is going to have an idea if they did well or not. The game showed you how many turns you took, but this doesn’t factor in how hard the level is. So then I decoupled the simulation from the rendering. Now I can quietly put it through 50 generations or so to determine how long it runs before it becomes stable, and then use this number of steps and number of remaining live cells to come up with some estimated “par” value for the player to play against. If I had simply committed to adding this feature in at the start, I could have saved myself some rework.