544 ◾ Appendix: Solutions to Selected Exercises
d. A major weakness is that there is no notion of usage; subsystem
C may be used much more than others, and system B may be
never used. ere are also no notions of severity of defects nor of
complexity of subsystems. (KLOC is just a crude measure of size.)
ere is a blurring of the key distinction between fault and failure.
Why should a failure always be due to a single defect in a single
subsystem?
e. It would be extremely useful to get some measure of usage for
each subsystem, since this would enable us to use the existing data
to produce genuine reliability assessments of the dierent sub-
systems. Since system users are already logging failures as they
occur, it should not be too ...