Tic-tac-toe is a game where you play in 3x3 squares, with 2 players in turn, and whoever first connects 3 squares in a line wins. Writing a tic-tac-toe game in a logic programming language like miniKanren has some interesting effects. For example, it becomes really easy to know about “if this is the current board, what would be all the next possible moves?” and “if this is the current board, what would be the boards that can produce this board?”.
So what’s good about this? Well, people has studied the tic-tac-toe game and figured out multiple strategies that can make you never lose (for example, you should always mark the center if it is available), and those strategies, together with the ability to easily query around boards, make it really easy to create an AI to play the game with you. If you look at the code, the AI is actually implemented in less than 20 lines – that’s less than one A4 paper if printed.
Interested? Fork the source code.