<!-- canonical: efficientnewlanguage.org/ai/examples/007-mvp-tic-tac-toe | ai_layer_version: 0.1.0 | updated: 2026-07-18 -->

# Example 007 — MVP proof: a real, recognizable program ported to EML

`tic_tac_toe.eml` is a "does EML actually work for something real" test — not a new-language-feature regression fixture like the other `examples/phaseN-*` directories, but a demonstration that a genuine, independently-authored small program can be expressed end to end in EML: transpiled to Python, executed, and verified against real Python via the interpreter equivalence gate.

## EML

```eml
# Ported from Python-World/python-mini-projects (projects/Tic_tac_toe,
# MIT-licensed) as a "real, recognizable program" MVP proof for EML —
# not a new language-feature test, but a demonstration that a genuine
# small program can be expressed end to end. The original is interactive
# (input()); this port replaces that with a fixed, scripted move sequence
# (including one deliberately invalid move) so the whole program is
# deterministic and verifiable against real Python (see tic_tac_toe.trace.jsonl,
# generated by `eml trace --run`).
class TicTacToe:
    def __init__(self):
        [" ", " ", " ", " ", " ", " ", " ", " ", " "] => self.squares
        [[0, 1, 2], [3, 4, 5], [6, 7, 8], [0, 3, 6], [1, 4, 7], [2, 5, 8], [0, 4, 8], [2, 4, 6]] => self.win_lines
        "X" => self.current

    def play_move(self, position):
        if self.squares[position] != " ":
            raise ValueError("square already taken")
        self.current => self.squares[position]

    def has_won(self):
        for line in self.win_lines:
            {self.squares[line[0]], self.squares[line[1]], self.squares[line[2]]} == {self.current} ? 1 : 0 => hit
            if hit == 1:
                return 1
        return 0

    def next_turn(self):
        self.current == "X" ? "O" : "X" => self.current

TicTacToe() => game
moves^+[0, 4, 1, 5, 0, 2]
winner^+""

for move in moves:
    try:
        game.play_move(move)
    except ValueError:
        continue
    game.has_won() => won
    if won == 1:
        game.current => winner
        break
    game.next_turn()

game.squares => final_board
final_board^0
winner^0
```

## Python (deterministic transpilation)

```python
class TicTacToe:
    def __init__(self):
        self.squares = [" ", " ", " ", " ", " ", " ", " ", " ", " "]
        self.win_lines = [[0, 1, 2], [3, 4, 5], [6, 7, 8], [0, 3, 6], [1, 4, 7], [2, 5, 8], [0, 4, 8], [2, 4, 6]]
        self.current = "X"
    def play_move(self, position):
        if self.squares[position] != " ":
            raise ValueError("square already taken")
        self.squares[position] = self.current
    def has_won(self):
        for line in self.win_lines:
            hit = 1 if {self.squares[line[0]], self.squares[line[1]], self.squares[line[2]]} == {self.current} else 0
            if hit == 1:
                return 1
        return 0
    def next_turn(self):
        self.current = "O" if self.current == "X" else "X"

game = TicTacToe()
moves = [0, 4, 1, 5, 0, 2]
winner = ""
for move in moves:
    try:
        game.play_move(move)
    except ValueError:
        continue
    won = game.has_won()
    if won == 1:
        winner = game.current
        break
    game.next_turn()
final_board = game.squares
print(final_board)
print(winner)
```

## stdout (executed)

```text
['X', 'X', 'X', ' ', 'O', 'O', ' ', ' ', ' ']
X
```

## Round-trip

`ok: true` — round-trip fixpoint reached (python1 == python2)

## Trace event types

eml:run:start · eml:classdef · eml:call · eml:assign · eml:return · eml:output · eml:run:done