?

Why shouldn't we test for equality with floats?

 Because computers use binary, floats are actually very close approximations of the actual values. Testing for equality can result in an unexpected error, so it's better to determine whether two numbers are close enough for our purposes rather than precisely equal.