What is the difference between C structures and Java classes?

Question

I'm a newbie to Java, but somewhat familiar to C. I wanted to know -- what differences are there between C structures and Java objects and invoking their methods? Or are the totally equivalent?

For example, the Bicycle structure:

class BicycleDemo {
     public static void main(String[] args) {

          // Create two different Bicycle objects
          Bicycle bike1 = new Bicycle();
          Bicycle bike2 = new Bicycle();

          // Invoke methods on those objects
          bike1.changeCadence(50);
          bike1.speedUp(10);
          bike1.changeGear(2);
          bike1.printStates();

          bike2.changeCadence(50);
          bike2.speedUp(10);
          bike2.changeGear(2);
          bike2.changeCadence(40);
          bike2.speedUp(10);
          bike2.changeGear(3);
          bike2.printStates();
     }
}

The reason why I am asking is because they look so similar! Thanks!

Answer 1

If you leave method overriding out of the picture, then you can think of Java classes and methods as a pair of a C-style struct and a set of functions that operate on those structs. For example, if you have a class like this:

 public class MyJavaClass {
       private int x;
       public int getX() {
           return x;
       }
       public int setX(int value) {
           x = value;
       }
 }

This would be similar to writing C code to this effect:

struct MyJavaClass {
    int x;
};

int MyJavaClass_getX(struct MyJavaClass* this) {
    return this->x;
}
void MyJavaClass_setX(struct MyJavaClass* this, int value) {
    this->x = value;
}

The main idea is that a method is similar to a function that takes the receiver object as an implicit "this" parameter. In C, you have to explicitly pass the receiver as a parameter to the function, while in Java this is done implicitly through the object.method() syntax.

If you start introducing method overriding, this becomes a bit more complicated because the method that you invoke on an object depends on the dynamic type of the object, not the static type. One way of simulating this is using something called a vtable or virtual function table, so named because of C++'s virtual keyword. The idea is that each object stores a pointer to a table of function pointers, one per function that can be overridden, and when a method is called on the object the appropriate function pointer is selected out of the table and called. So, more properly, the above Java object might look something like this:

struct MyJavaClass_Vtable {
    void (*getX)(struct MyJavaClass* this);
    void (*setX)(struct MyJavaClass* this, int value);
};

struct MyJavaClass {
    struct MyJavaClass_Vtable* vtable;
    int x;
};

int MyJavaClass_getX(struct MyJavaClass* this) {
    return this->x;
}
void MyJavaClass_setX(struct MyJavaClass* this, int value) {
    this->x = value;
}

/* A global instance of the vtable for MyJavaClass */
struct MyJavaClass_Vtable MyJavaClassVtableInstance = {
    &MyJavaClass_getX,
    &MyJavaClass_setX
};

Whenever you created an instance of MyJavaClass, you'd set up its vtable doing something like this:

struct MyJavaClass* mjc = malloc(sizeof *mjc);
mjc->vtable = &MyJavaClassVtableInstance;

Then, when invoking a function like this (in Java):

myJavaClass.getX();

In C it would look like

myJavaClass->vtable->getX(myJavaClass);

So in a sense a Java class is just a struct with some extra metainformation. Of course, to the programmer it looks totally different - there's encapsulation, polymorphism, a stricter type system, etc. - but at the level of native code a regular C struct and a Java class probably look very similar.

Hope this helps!