Promises And Problems
Keith J. Laidler · 1993
Abstract This chapter examines the metrics of the Software Science model, cyclamate complexity, and information flow in more detail. These are selected on the basis of their popularity within the software engineering literature, and the significance of the claims made by their progenitors. Claimed benefits are summarized. Each metric is then made the subject of an in–depth critique and all are found wanting. The important point made by this chapter is that this is not due to mischance, but is indicative of deeper problems of methodology employed in the field of software metrics. The preceding chapter has provided an introduction to the area of software metrics and has, at the same time, outlined some of the major problems of the subject. The aim of this chapter is to look at three widely used metrics and use them as examples to illustrate more fully these problems. The first metric – or rather family of metrics – is that derived from Halstead’s Software Science which is almost certainly the widest studied and ‘validated’ metric. Of the three metrics examined in this chapter this has, over the last ten years, been the recipient of most criticism; we summarize these criticisms here.