This is discussed in the article. The standard is to use infinity, not zero.

> A hardware example of exception handling is the IEEE 754 standard for floating point arithmetic which defines values for various exceptions (e.g. infinity for divide-by-zero)