An interesting way to keep people interested in the game (during a fight, that is) would be to allow players to update the AI on the fly.
I'm thinking each player would submit an AI to start the game. At any point during the game players could "check-in" code (VCS-style) and cells created from that point on (via the splitting described in the article) would have this new code. Players would be allow, for instance, 2 check-ins per game. The upgraded cells would be represented by different shades of the player's color so that they could be kept track of.
I see this being useful for times when a player drastically misuses his cells initially or for an entirely different style of play (a different bracket on the ladder server).
You could try to put into the code the conditions that would cause you to change your strategy. Then when all those conditions are in place you send a signal to all cells to change strategy.
Your code then looks at the strategy variable to see which code path to execute.
I'm thinking each player would submit an AI to start the game. At any point during the game players could "check-in" code (VCS-style) and cells created from that point on (via the splitting described in the article) would have this new code. Players would be allow, for instance, 2 check-ins per game. The upgraded cells would be represented by different shades of the player's color so that they could be kept track of.
I see this being useful for times when a player drastically misuses his cells initially or for an entirely different style of play (a different bracket on the ladder server).