Predicting Adaptability Level of Object-Oriented Software using Metrices and Threshold Values

Adaptability, a sub-characteristic of software quality has become so important in recent years and a key characteristic of well designed applications because modern software systems are expected to service swiftly revolutionizing business environments. Much effort have been invested to the development of object oriented design metrics to measure software properties such as coupling, cohesion and inheritance, but metrics to be used by software developers to predict, at the coding phase, the extent to which object oriented software adapt is needed. In this work a software analyzer has been developed using JAVA programming language to measure the values of some internal properties of object oriented software using combination of associated object oriented design metrics. Threshold values of these object oriented design metrics are grouped into levels taking into consideration the desirable values of the internal software properties.The measured values are matched against these threshold values and the level of each attribute is noted. Decisions rules are formulated and used in conjunction with binary logic combination of the possible internal software properties to aid predict adaptability level (Poorly Adaptable 1; Fairly Adaptable-2; Adaptable -3) of a given software. The analyzer is implemented on 12 open source JAVA projects and their adaptability levels are noted. The analysis revealed that source codes with low coupling, high cohesion, low inheritance and low complexity easily adapt to new operating environment with little or no modifications to the source codes.Department of Computer Science, University of Benin, Benin City, Nigeria, Email: vakwukwuma@yahoo.com. Department of Computer Science, University of Uyo, Uyo, Nigeria, Email: edwardudo@uniuyo.edu.ng