Return to Answer

I just noticed you referenced my article.

I'm the author of the blog post you're talking about. To clarify a few things:

Update

Update

A discussion in the comments warrants some clarification...

In C, you can write

void swap(int *x, int *y) {
    int t = *x;
    *x = *y;
    *y = t;
}

int x = 1;
int y = 2;
swap(&x, &y);

This is not a special case in C. Both languages use pass-by-value semantics. Here the call site is creating additional data structure to assist the function to access and manipulate data.

The function is being passed pointers to data, and follows those pointers to access and modify that data.

A similar approach in Java, where the caller sets up assisting structure, might be:

void swap(int[] x, int[] y) {
    int temp = x[0];
    x[0] = y[0];
    y[0] = temp;
}

int[] x = {1};
int[] y = {2};
swap(x, y);

(or if you wanted both examples to demonstrate features the other language doesn't have, create a mutable IntWrapper class to use in place of the arrays)

In these cases, both C and Java are simulating pass-by-reference. They're still both passing values (pointers to ints or arrays), and following those pointers inside the called function to manipulate the data.

Pass-by-reference is all about the function declaration/definition, and how it handles its parameters. Reference semantics apply to every call to that function, and the call site only needs to pass variables, no additional data structure.

These simulations require the call site and the function to cooperate. No doubt it's useful, but it's still pass-by-value.