Beginners question - Why does C use * to declare a pointer and not &?

Question

So let's say:

int n = 50;
int *p = &n;

Why do we use * and not & in C? What's the key difference between * and &?

Answer 1

Using * to declare pointers is mostly a matter of convention, but there is a reason of consistency: the * in the declaration int *p means int is the type of *p.

It might seem more consistent to write int &p = &n as p is initialized to the address of n but this convention would not hold for double pointers: int **pp defines pp as a pointer to a pointer to an int, yet pp cannot be initialized with &(&n).

Note that int& p = n; is a valid definition in C++ for a reference to an int, which is a pointer in disguise. Modifying p would then modify n. References are implemented as pointers without the indirection notation.

Answer 2

C declarations follow a convention often described as “declaration mimics use”. The structure of a declaration matches (as closely as possible) the structure of an expression in the code.

For example, let’s say you have a pointer to an int named p. To access the value stored in the pointed-to object, we dereference the pointer with the unary * operator, like so:

printf( "value of what p points to is %d\n", *p );

The expression *p has type int, so we declare p as

int *p;

This conveys that the variable p has type "pointer to int", because the combination of p and the dereference operator * in the declarator yield an expression of type int.

And this is primarily why pointers aren't declared using the & operator in C, because the type of the expression &p wouldn't be int.

C++ uses the unary & operator to declare references, which are not the same thing as pointers:

void foo( int &ref )
{
  ref = new_value(); // writes a new value to the actual parameter ref references
}

int main( void )
{
  int x = 1;
  std::cout << "x before foo = " << x << std::endl;
  foo( x );
  std::cout << "x after foo = " << x << std::endl;
  return 0;
}

References kind of break this system - there's no &ref expression that has type int, it just notes that ref resolves to the same object as some other identifier.

---

C declarations are made up of two parts - a sequence of declaration specifiers (storage class specifier, type qualifier(s), type specifier(s)) followed by a sequence of (possibly initialized) declarators. Pointer-ness, array-ness, and function-ness are specified in the declarator:

            declaration specifiers           declarators
                     |                           |
+--------------------+----------------+ +--------+--------+
|                                     | |                 |
static const volatile unsigned long int a[10], *p, f(void);
   |   |            | |               | |      |   |
   |   |            | |               | |      |   +––––––– function declarator
   |   |            | |               | |      +––––––––––– pointer declarator
   |   |            | |               | +––––––––—–––––——–– array declarator
   |   |            | +–––––––+––––––—+
   |   |            |         |
   |   |            |         +––––––––––––––——–––––––––––— type specifiers
   |   +––––––+–––––+
   |          |
   |          +–––––––––––––––––––––––––––––––––––––––––––– type qualifiers
   +–––––––––––––—––––––––––––––––––––––––––––––––––––––––– storage class specifier

This is important - there's no "pointer to" type specifier. The pointerness is specified by the declarator. If you write something like

int* a, b;

it will be parsed as

int (*a), b;

and only a will be declared as a pointer.