Burstiness of interrupted Bernoulli process

The author first rectifies an error that has been propagating through the literature concerning the squared coefficient of variation of an interrupted Bernoulli process; secondly, the author shows that under the practical constraint that the mean length of the idle period of an interrupted Bernoulli process is finite, the squared coefficient of variation (and hence the burstiness) has a finite maximum value. The relationship between the transition probabilities of the Markov chain to give this maximum is derived