The backend is hosted on Scalingo and my trial expired.

Figure out if I want to keep paying for it or host the backend somewhere else :-)

The project has grown to a size where making involved changes can take a while before all the compile issues are fixed and where it would be very useful to have a unit test suite to better test changes incrementally.

Or they have 5m total to play.

People don't necessarily know how to play the game, and discovering how to play by yourself can be fun but also a bit frustrating...

Currently when presented with choices with the same weight the AI will always play the first option

Impact

When in a given position the AI will always play the same series of moves. This makes the AI predictable. If a player finds a way to wind they can play this way over and over again.

When a board has been won by a player, would be nice to draw a big-ass red X or O over the won board.

If a player clicks on an already filled cell that would send the next player to an already-won board if it wasn't filled, no mark is placed but the turn is considered played and valid. The next player can then choose anywhere to play.

Add some color to make it fancier there.

Add easy, normal, hard difficulty for AI.

Current endgame looks like this:

i.e. the board where the next move should happen were the game not finished is highlighted and other boards faded out.

The whole board should be highlighted instead.

When sending the opponent to the same board when winning this board, the opponent should be able to choose whichever board the want to play on (as is the case when sent to a board already full/won).

Right now, the opponent is sent to the just-won board, and cannot play.

Have a quick 2-min intro on how to play the game with step-by-step instructions and next buttons, to show the principle of sending an opponent to the next grid, and that the opponent can choose when the grid is full or won already.

Doable as either a bunch of overlayed pictures, or fancier, as an interactive tutorial.

In this case, when player O has just played to send player X at the center but this sub-board has already been won:

Player X can choose which sub-board to play on. Every sub-board can is fair game except the center one which has already been won. In this case every sub-board except the center one should be highlighted.

Have a proper README with:

• Explanations as to why tictactoe sucks
• And why ultimate tictactoe will save the world and the rules of the game
• And install instructions

This would help figure out what was the last move played by your opponent, especially useful in remote games or games against the AI.

Probably requires a model change to keep track of what the last move was.

On the view side the change might be a bit complicated as it will most likely require passing down what the last move was in order to do some style trickery to highlight it.

The menu is ugly as f*** - redesign it so that it's prettier.

Currently there's no way to figure out whose turn it is to play, which can be a bit frustrating.

Ideas:

• Place a "phantom" mark (cross or circle) under the cursor on hover where a mark would be placed for the active player -> This is cool because it doesn't clutter the UI which I want to avoid so the game still works as well on a small mobile screen, however this only works with a mouse and not with a touchscreen where there is no hover!

Use localStorage to save the game state and restore it.

Have an endgame screen, showing "Draw!" or "X wins!" or "O wins!" and with a "New game" button.

It should be easy enough to add an option to watch a replay of the game at the end.
Because Elm.

Currently the grid has a fixed size. It should respond to the current width/height (whichever is the smallest) of the page.

Just so I don't have to remember how to use travis