Lines 12–15 declare a constructor with three default parameters. We did not define a constructor with an empty parameter list, so for class Time2 the constructor at lines 12–15 is also considered to be the class’s parameterless constructor—you can call the constructor without arguments and the compiler will provide the default values. This constructor also can be called with one argument for the hour, two for the hour and minute, or three for the hour, minute and second. This constructor calls SetTime to set the time.

Lines 18–19 declare another Time2 constructor that receives a reference to a Time2 object. In this case, the values from the Time2 argument are passed to the three-parameter constructor at lines 12–15 to initialize the hour, minute and second. In this constructor, we use this in a manner that’s allowed only in the constructor’s header. In line 19,

: this(time.Hour, time.Minute, time.Second) { }

: this followed by parentheses containing arguments indicates a call to one of the class’s other constructors—in this case, the Time2 constructor that takes three int arguments (lines 12–15). Line 19 passes the values of the time argument’s Hour, Minute and Second properties to initialize the Time2 object being constructed. Any initialization code in the body of the constructor at lines 18–19 would execute after the other constructor is called.

Using this as in line 19 is called a constructor initializer. It enables a class to reuse initialization code provided by a constructor, rather than defining similar code in another constructor. If we needed to change how objects of class Time2 are initialized, only the constructor at lines 12–15 would need to be modified. Even that constructor might not need modification—it simply calls the SetTime method to perform the actual initialization, so it’s possible that the changes the class might require would be localized to SetTime.

Line 19 could have directly accessed instance variables hour, minute and second of the constructor’s time argument with the expressions time.hour, time.minute and time.second—even though they’re declared as private variables of class Time2.

Method SetTime (lines 23–28) invokes the set accessors of the properties Hour (lines 31–47), Minute (lines 50–66) and Second (lines 69–85), which ensure that the hour is in the range 0 to 23 and that the values for the minute and second are each in the range 0 to 59. If a value is out of range, each set accessor throws an ArgumentOutOfRangeException (lines 41–42, 60–61 and 79–80). In this example, we use the exception class’s overloaded constructor that receives three arguments:

• the string name of the item that was out of range

• the value that was supplied for that item and

• a string error message.

It’s common to include in an exception’s error message a variable’s or property’s identifier. This information can help a client-code programmer understand the context in which the exception occurred. Prior to C# 6, you had to hard code these identifiers into your error-message strings. As of C# 6, you can instead use the nameof operator (lines 41–42, 60–61 and 79–80), which returns a string representation of the identifier enclosed in parentheses. For example, the expression

nameof(value)

in line 41 returns the string "value" and the expression

nameof(Hour)

in line 42 returns the string "Hour".

Time2’s properties are accessed throughout the class’s body—SetTime assigns values to Hour, Minute and Second in lines 25–27, and ToUniversalString and ToString use properties Hour, Minute and Second in line 89 and lines 93–94, respectively. These methods could access the class’s private data directly. However, consider changing the time’s representation from three int values (requiring 12 bytes of memory) to one int value representing the total number of elapsed seconds since midnight (requiring only four bytes of memory). If we make this change, only code that accesses the private data directly would need to change—for class Time2, the bodies of properties Hour, Minute and Second. There would be no need to modify SetTime, ToUniversalString or ToString, because they access the private data indirectly through Hour, Minute and Second. Designing a class in this manner reduces the likelihood of programming errors when altering the class’s implementation.

Similarly, each constructor could include a copy of the appropriate statements from method SetTime. Doing so may be slightly more efficient, because the extra constructor call and the call to SetTime are eliminated. However, duplicating statements in multiple methods or constructors makes changing the class’s internal data representation more difficult and error-prone. Having one constructor call the other or even call SetTime directly allows any changes to SetTime’s implementation to be made only once.