1. An exception that occurs during floating-point operations, such as division by zero.
2. The number of digits a floating-point number can represent accurately.
