
230 ◾ Software Metrics
is is one reason why the normal distribution can only be a crude
approximation to the “true” distribution.
If you are using a continuous function as a probability distribution, you
need to understand that, unlike discrete probability distributions, it is not
meaningful to talk about the probability of a (point) value. So, we cannot
talk about the probability that a person’s height is 158 cm. Instead, we always
need to talk in terms of a range, such as the probability that a person’s height
is between 157.9 and 158.1 cm. e probability is then the proportion of the
curve lying between those two values. is is shown ...