
Software Reliability ◾ 485
of novel faults. We can capture data to help us assess the short- and long-
term reliability by monitoring the time between failures. For example, we
can track execution time, noting how much time passes between succes-
sive failures.
Table 11.1 displays this type of data, expressing the successive execution
times, in seconds, between failures of a command-and-control system
during in-house testing using a simulation of the real operational envi-
ronment (Musa 1979). is data set is unusual, in that Musa took great
care in its collection. In particular, it was possible to obtain the actual
execution time, rather t ...