Question

I have a question about pure virtual functions. I am not clear about how it works and when we need to use pure virtual functions. This is the example that I do not understand:

file.h

class A
{
public :
            A();
            ~A();

            virtual void func1(void) = 0;
            virtual UINT32 func2(void) = 0;
            UINT32 initialize(void) = 0;
}


file.cpp

UINT32 A:initialize (void)
{
            func1();
            func2();
            return (result);
}

Can anyone explain in detail what this example actually does and what the result is?
I really appreciate your help and knowledge. Thank you very much.

Answer 1

(Note, the declaration for initialize() should not be virtual, and the implementation of initialize() should return func2() probably. As this is an example, it doesn’t really matter what initialize() does, but it should compile correctly.)

The primary purpose of virtual functions is to achieve polymorphism.

Class A defines two pure virtual methods, and initialize calls them. This allows code in the program to initialize something of type A without being aware of the subclass. There may be many subclasses of A, and each may do something slightly differently within func1() and func2(). Since the code that only knows about A is sometimes initializing objects of different types through A, A can be referred to as a polymorphic type.

class B : public A {
    void func1 () { std::cout << "B::func1" << std::endl; }
    UINT32 func2 () { return 1; }
};

class C : public A {
    void func1 () { std::cout << "C::func1" << std::endl; }
    UINT32 func2 () { return 2; }
};

void call_initialize (A *a) {
    std::cout << a->initialize()  << std::endl;
}

B b;
C c;
call_initialize(&b);
call_initialize(&c);

Results in the output:

B::func1
1
C::func1
2

This is an example of polymorphic behavior, because the output is different depending on if A was subclassed by B or by C.