Lots of tidying up done today, making the game look a bit better. Katia has started working on new snake images to fit the new board size, but they're not done yet.

I've added a new invulnerability powerup, which lets you slither through yourself, other snakes, or bones - but if you're still crossing them when you leave invulnerability, you die. You can still eat friendly powerups in the meantime. You flash faster and faster as your time runs out.

The client/server protocol is now fully implemented my end. It now sends "prepare" messages with the countdown if they're in the next game, and "play" and "gameover" messages to complete the set.

The game no longer starts immediately on state transition; there's now a six-second countdown with big flashing numbers to give people a few seconds to find their snake. And there's a few seconds to see the scoreboard at the end of the game, before it transitions back to the 'insert coin' state. So now the game can run fully standalone in a loop.

More of the client-server communication is now implemented. The client is now put into the queue (and sent a message) after they enter their name. The server now broadcasts a 'gameover' to all clients on startup, in case any clients are stuck in the game from a server crash.

The 'insert coin' state now lists players in the order that they joined the queue, and the 'play' state now adds players to the game in that same order.

Edward came over to the atelier and we've been working on adding networking support for the mobile phone clients. We're using his Python software that sits between the clients and the game servers (which are known as "screens"). The clients talk to the "swarm server" natively using WebSockets, falling back to Flash-emulated sockets if the browser doesn't support them, and to AJAX if it doesn't support either. The game server talks to the swarm server using a simple TCP connection.

Okay kid, this is where it gets complicated. Since this is just a throwaway prototype, I'm going to bend a lot of rules from now on to speed things up. For example, the game logic was originally completely encapsulated (in fact I had the full game simulated 'in memory' before writing a single line of graphical code). But now to save some time, I'm going to mix a lot of the presentation logic into those same classes, and do the nasty with member variable visibility.

OK, so I got bored after dinner and decided to plunge on tonight regardless. Mostly graphical changes this hour: now using the Nokia font; made the images larger and redrew them; gave the snakes a different head depending on which way they're facing; added screen glare (to be redrawn later); ensure a certain number of rabbits are available "most" of the time (depending on number of players); and a few minor technical changes.

Hour 2 went well, after a minor technical glitch. The game is now "fully playable", though I've slowed it down to test. Each snake is independently controlled, and has its own colour, name and score. Snakes no longer grow infinitely, they now only grow when they eat a rabbit (which also increases their score). Powerups and skulls decay after a while. When a snake dies, it turns into skulls for a while too. Rabbits are randomly placed every so often. Maybe skulls should be too.

So, apparently you can write Snake in about an hour. I wonder if I've set myself enough of a challenge. Here it is, playing fine. Done so far: snake movement, snakes hitting snakes, snakes hitting themselves, snakes hitting walls, snakes hitting powerups, snakes hitting obstacles, snakes headbutting each other. Next up: scoring and control of multiple snakes.

This Friday there's a video games party at my atelier, celebrating and sharing video games, old and new. I've been meaning to write a new game for people to try out on the day, but of course I've been too busy to think about it and didn't remember until the weekend. Then I got drunk and didn't remember until today.