Question

I do not understand how this piece of code (from Wikipedia) works:

template <int N>
struct Factorial 
{
    enum { value = N * Factorial<N - 1>::value };
};

template <>
struct Factorial<0> 
{
    enum { value = 1 };
};

// Factorial<4>::value == 24
// Factorial<0>::value == 1
void foo()
{
    int x = Factorial<4>::value; // == 24
    int y = Factorial<0>::value; // == 1
}

• What is this weird template that
  takes <int N>?
• What is this second
  weird template <>?
• What are the
  enumerations for?
• What is the advantage
  of using this rather than normal
  runtime factorial calculation?
• How often do you people use this? I have been using C++ for a while now, but never used this before. How big a part of C++ was I missing out on?

Thanks!

Answer 1

• What is this weird template that takes <int N>?

In C++, template arguments can either be types (prefixed with class or typename) or integers (prefixed with int or unsigned int). Here we are in the second case.

• What is this second weird template <>?

template<> struct Factorial<0> is a complete specialization of Factorial class template, which means that 0 is considered a special value to which corresponds its own version of Factorial.

• What are the enums for?

enums are the way to compute values in metaprogramming C++

• What is the advantage of using this rather than normal runtime factorial calculation?

The reason why this code was created in the first place is to create a proof of concept that calculus can be done using metaprogramming. The advantage is that generated code is extremely efficient (calling Factorial<4>::value is equivalent to simply writing “24” in your code.

• How often do you people use this? I have been using C++ for a while now, but never used this before. How big a part of C++ was I missing out on?

Such functionality is rarely achieved using this method, but metaprogramming is used more and more nowadays. See Boost meta-programming library to get a hint of what can be done.