Software reuse and information theory based metrics
William R. Torres, M.H. Samadzadeh · 2002
The main purpose of the research presented is to theoretically investigate the effect of reusing software on metrics that are based on the entropy function of communication information. R.N. Chanon's (1973) entropy loading and E.T. Chen's (1978) control structure entropy were applied to C and Ada programs obtained from the open literature. Four units of decomposition (statement, component, module, and program) along Chanon's definition of an object were introduced to classify software reuse units. A total of three versions for each of the three programs were considered (optimum reuse, intermediate reuse, and no reuse). The lines of code metric was utilized to quantify the amount of nonreusable code in each of the versions of the programs. Pearson product-moment correlations were computed between the information theory based metrics and the lines of code metric. The results of this study show that there are significant correlations between the information theory based metrics and software reusability.>