Rationale

In the context of Lionel Martin’s challenge, I have been involved in the implementation of multiagent network games meant to be played by bots.

After implementing several games (spaceships in 2014, aquar.iom in 2016) I came to the following conclusions.

• Implenting the network server is tough.
• Handling the clients correctly (errors, fairness, not spamming slow clients…) mostly means that most of the development time is in the network game server, not in the game itself.
• The games in this context are quite specific (fair, turn-based, visualizable, no big performance constraint), which means the development effort can be shared regarding the network server.