A critical analysis of current OO design metrics

Chidamber and Kemerer (C&K) outlined some initial proposals for language-independent OO design metrics in 1991. This suite is expanded on by C&K in 1994 and the metrics were tested on systems developed in C++ and Smalltalk(TM). The six metrics making up the C&K suite can be criticised for a number of reasons. This does not make them bad metrics; on the contrary the C&K work represents one of the most thorough treatments of the subject at the current time. However, the authors explicitly state, ...there is no reason to believe that the proposed metrics will be found to be comprehensive, and further work could result in additions, changes and possible deletions from this suite. This analysis will serve to make other researchers and practitioners aware of some of the problems that may arise from using these measures. As a by-product, the axioms of E. Weyuker (1983) come under scrutiny in terms of their applicability to object-oriented metrics.