public class MyObject
{
    private MyExpensiveObject _myExpensiveObject;

    public MyExpensiveObject ExpensiveObject
    {
        get { return _myExpensiveObject ??
                     (_myExpensiveObject = new MyExpensiveObject()); }
    }
}

Now that the actual port is done, I plan on cleaning up the code. Like most software written in C in the 80s, it relies on way too many global variables, uses an abundance of switch statements on object types, and has no abstraction between the game logic and the UI. Since it’s on GitHub, that should leave a nice history of the refactorings I try to apply to it. Of course, before I can try to do any refactoring, I need to get some tests in place. That should be fun considering how much it relies on direct access to the curses library and a random number generator.

The repository is here. Feel free to fork it to help out if you like.

For Ruby programmers, this is easy: just split on underscores. For us .NET programmers, we need something a little fancier since we like to SquashOurMethodNamesTogetherLikeThis.

Here’s a little regular expression that can do exactly that:

(?<!^)(?=[A-Z])

That basically says to match right before every capital letter unless the capital letter is at the beginning of the string.

You can use the Regex.Split method to do the actual work of turning the method name into a string array:

Regex splitter = new Regex(@"(?<!^)(?=[A-Z])");
string[] words = splitter.Split(methodName);

My friend, Michael Kennedy, demoed in class recently how to use LINQPad to test regular expressions. It came in really handy while working on this:

I really need to click that “Activate Autocompletion” button so I can give Joseph Albahari the money he deserves for creating such a useful tool.

By the way, I used to think that “camel case” was for words like “camelCase” and “pascal case” was for words like “PascalCase”, but Wikipedia doesn’t make that distinction.

I decided to challenge myself by writing my solution as a single “line” of code:

public static class StringExtensions
{
    private static Dictionary<char, char> _charMap = new Dictionary<char, char>
    {
        { (char)8208,  (char)45 },
        { (char)8211,  (char)45 },
        { (char)8212,  (char)45 },
        { (char)8722,  (char)45 },
        { (char)173,   (char)45 },
        { (char)8209,  (char)45 },
        { (char)8259,  (char)45 },
        { (char)96,    (char)39 },
        { (char)8216,  (char)39 },
        { (char)8217,  (char)39 },
        { (char)8242,  (char)39 },
        { (char)769,   (char)39 },
        { (char)768,   (char)39 },
        { (char)8220,  (char)34 },
        { (char)8221,  (char)34 },
        { (char)8243,  (char)34 },
        { (char)12291, (char)34 },
        { (char)160,   (char)32 },
        { (char)8195,  (char)32 },
        { (char)8194,  (char)32 }
    };

    public static string UnWordify(this string value)
    {
        return new string((from c in value
                           select _charMap.ContainsKey(c) ? _charMap[c] : c)
                          .ToArray());
    }
}

OK, so I cheated and used a dictionary to create my map from the bad characters to the good ones, but the actual method is a single statement.