How it works...

In the console window, you will see two very different results. Simply put, in the first line, the variable a is the variable a, the variable b is the variable b, and the variable val is the variable val.

In the second line, the variablea is the variable a, the variable b is the variable b, and the variable refVal is the variable b. This is the whole crux of the ref keyword. In the first GetLargest() method, we returned the largest value into the variable val. This value was 20. The variable val and the variable b had no relation to one another as they were allocated different spaces in memory.

In the second GetLargest() method, we returned the largest variable itself (which was b) into the variable refVal. The variable refVal therefore becomes an alias of the variable b because they both point to the same space allocated in memory. To illustrate this even more clearly, let us have a look at the memory addresses for the variables.

From the Project menu, go to the Properties of the current project. In the Build tab, check the option to Allow unsafe code and save the properties.

Add the following code to your console application:

unsafe
{
  IntPtr a_var_memoryAddress = (IntPtr)(&a);
  IntPtr b_var_memoryAddress = (IntPtr)(&b);
  IntPtr val_var_memoryAddress = (IntPtr)(&val);

  fixed (int* refVal_var = &refVal)
  {
    IntPtr refVal_var_memoryAddress = (IntPtr)(refVal_var);
    WriteLine($"The memory address of a is {a_var_memoryAddress}");
    WriteLine($"The memory address of b is {b_var_memoryAddress}");
    WriteLine($"The memory address of val is {val_var_memoryAddress}");
    WriteLine($"The memory address of refVal is
              {refVal_var_memoryAddress}");
  }
}

This code is not really related to the recipe on ref returns and locals, so I'm not even going to go into it in any detail. If you want to learn more about pointers in C#, start with the MSDN article on Pointer types (C# Programming Guide): https://msdn.microsoft.com/en-us/library/y31yhkeb.aspx.

Run your console application and take a look at the memory addresses listed:

You will notice straightaway that variable b and variable refVal have the same memory address of 11531252, while variable b and variable val have different memory addresses.

So now for the million dollar question: Why is this feature in C# 7.0 even useful? Well, simply put, it can improve performance. Many developers mention that it will be quite useful for game programmers, who can now pass these aliases around to reference large data structures. This means that they don't have to make a copy of a large array (for example) in order to work with it. Using ref, they can create an alias that points to the original memory location of the array and read or modify it directly. Thinking of it this way suddenly brings the usefulness of this C# 7.0 feature into perspective.

Will I ever use it? I don't really know. Perhaps not often but, as with local functions, this feature of C# 7.0 is really a great addition to the developer's toolkit. It solves some really tricky problems when you want to get away from copying around large structures in your code.