Why all rings should have a 1

We argue that the definition of ring should require the existence of a multiplicative identity 1 because this requirement is part of what associativity should be. We list counterarguments in order to rebut them.