Question

So I made the following test for a class Board that would to be born:

[TestMethod]
public void Set_The_Origin_As_Violet_And_The_Query_Confirms_It() {
    Board board = new Board(10, 10);

    Color expected = Color.Violet;
    board.SetColorAt(0, 0, expected);
    Color actual = board.GetColorAt(0, 0);
    Assert.AreEqual(expected, actual);
}

I tried running the code, but the compiler signaled that Board didn’t exist. So I created it.

I tried to run the code again, but it was of no avail, as both SetColorAt() and GetColorAt() methods didn’t exist. I created them:

public void SetColorAt(int x, int y, Color color) {
}

public void GetColorAt(int x, int y) {
}

Still, not everything was fine, as I had to return Color.Violet. So I changed GetColorAt() to

public void GetColorAt(int x, int y) {
    return Color.Violet;
}

So I got a green light for the first time.

What I want my final code to be on the class board is something of the form:

public class Board
{
    private Color[,] board;

    public Board(int x, int y)
    {
        board = new Color[x, y];
    }

    public void SetColorAt(int x, int y, Color color) {
        board[x, y] = color;
    }

    public Color GetColorAt(int x, int y) {
        return board[x, y];
    }
}

My first question is….

how to get there? Can I consider that in the “refactor” phase of the Unit-Test I have shown above, when removing duplication I will eventually end up with this code?

If the answer is yes, I feel that my Unit-test is testing something way too “localized” for what the code actually does. You see, the test is just checking for 1 pixel and one color, while the code itself is way more complex and rich.

Maybe the solution would be to add more Unit-Tests? Which would you advise to make?

My second question is…

I know later I’ll want to have an IBoard. Should I just express that in the above Unit-Test? Should I let

Board board = new Board(10, 10);

as it is and still create the IBoard interface? How to deal with this?

Answer 1

public void GetColorAt(int x, int y) {
    return Color.Violet;
}

The return Color.Violet statement here can be considered Data duplication, so you can refactor this bit out. The simplest thing that could possible work though would be having a single Color value that gets set when you call SetColorAt

public class Board
{
    private Color theColor;

    public Board(int x, int y)
    {
    }

    public void SetColorAt(int x, int y, Color color) {
        theColor = color;
    }

    public Color GetColorAt(int x, int y) {
        return theColor;
    }
}

Now, you’ll need more tests, to show you can set seperate cells to different colors.