Object oriented metrics useful in the prediction of class testing complexity

Abstract

Adequate metrics of object-oriented software enable one to determine the complexity of a system and estimate the effort needed for testing already in the early stage of system development. The metrics values enable to locate parts of the design that could be error prone. Changes in these parts could significantly, improve the quality of the final product and decrease testing complexity. Unfortunately only few of the existing Computer Aided Software Engineering tools (CASE) calculate object metrics. In this paper methods allowing proper calculation of class metrics for some commercial CASE tool have been developed. New metric, calculable on the basis of information kept in CASE repository and useful in the estimation of testing effort have also been proposed. The evaluation of all discussed metrics does not depend on object design method and on the implementation language.