Java：声明空构造函数以从另一个类初始化对象的替代方法

Question

I am creating a simple, text-based interactive application, where the user is prompted for input. Each input corresponds to a specific command, which invokes a certain method. Because there are many methods, I have chosen to distribute them between different classes. So the application works like this:

The main class is responsible for reading the user input (using a Scanner object). The user input is then passed as a parameter to a method B in class B that decides which command the input corresponds to. After deciding, the method invokes the correct method, which can be in any of the other classes (in this example, method C in class C)

This means that I have to initialize object instances in order to avoid null pointer exception. One of the classes that I have to initialize an object from only has a constructor with parameters, but I don't actually want to initialize an object with values in this case, but only to use it as a reference pointer to the object so I can invoke methods from that class. Now I am using an empty constructor in order to solve this, but is there any alternative, better solution than declaring an empty constructor? I have written code that I think demonstrates my problem:

public class MainClass {
    ClassB classB = new ClassB();

    public void methodA {
        classB.methodB(userInput);
    }
}


public class ClassB {
    ClassC classC = new ClassC();

    public void methodB {
        classC.methodC();
    }
}


public class ClassC {
    String name;
    int age;

    public ClassC(String name, int age) {
        this.name = name;
        this.age = age;
    }

    public ClassC() {
    }

    public void methodC() {
        // Do something
    }
}

Answer 1

but I don't actually want to initialize an object with values in this case, but only to use it as a reference pointer to the object so I can invoke methods from that class

The problem is that those methods normally need state of the object to make sense (ie the fields to be initialized). If in your case they don't, maybe it makes more sense to make the method static (and call ClassC.methodC() ) or put these methods on some other class with a default constructor.

Answer 2

The real answer is: don't do that. You see, if that ClassC is about storing details of a player ... first of call it Player (even when putting up examples: use meaningful names so people understand your intent).

As soon as we talk about players, it becomes clear: a Player object is probably supposed to model a (human or not) player. There are no humans "without" a name or age. That tells you: wrong approach. Your classes and models should reflect reality.

You don't create empty object upfront; you create them at that point in time when it makes sense (and then you provide the necessary information to them, so you can ideally make the objects immutable, by declaring all your fields to be final).

So, long story short: step back and rethink your design. If there is "something" to do that doesn't require the presence of a populated Player object, then that should probably be static method and it should probably not sit on the Player class.