Comment by twhitmore

A typical place this distinction is useful is that it determines whether the test-suite runs to completion with multiple tests, or panics/terminates hard on the first failure.