Object oriented design metrics framework based on code extraction

This paper proposes a framework to compute and evaluate metrics for object oriented design (MOOD). A new overall maintainability index is proposed to measures the design effectiveness of an implemented software system. The new proposed maintainability index is based on four famous primitive metrics. The proposed framework works on both the object code and source code of the software system. Object code extraction is more powerful than source code extraction in the situation of legacy third party systems when the source code is not available. The framework manually integrates with a commercial tool to obtain two MOOD metrics. The framework is implemented to measure two other MOOD metrics. It merges the four metrics to calculate the proposed maintainability index (PMI). The framework could be extended with a machine learning component that assign weights to the MOOD metrics´ measures according to software experts´ evaluation. A case study was applied to a software product through five versions of evolution. The results of applying the PMI of each version were compared. The results showed exact compliance and alignment with experts´ reviews and postmortems results.