ANALYSING EFFECT OF REFACTORING ON SOFTWARE MAINTAINABILITY USING OBJECT ORIENTED METRICS

Abstract

Software maintainability is the ease with which a software components can be modified to rectify the defects or their cause, repair or supplant broken or exhausted segments without replacing the working parts, prevent unexpected working condition, maximize a product's useful life, maximize efficiency, reliability, and safety, meet new requirements, make future maintenance easier, or cope with a changed environment.

For vast programming frameworks, the maintenance stage has longer term than all the past life-cycle stages taken together, causing significantly more exertion. The time spent and exertion required to keep software product operational after deployment is exceptionally critical and expends to 40-70% of the total cost of the whole life cycle. Nice measure of software maintainability can enable better dealing in the maintenance stage exertion. In past writing, analysts and experts have proposed few machine learning calculations with a target to anticipate programming viability and assess them.

Maintainability model as described by S. Counsell [6] is used as base in this study. Maintenance is very important phase of software life cycle. And many researchers [1,2,3,4,5,16] have already shared their findings about object oriented metric and code refactoring has direct impact on maintainability. As per past literature, Maintainability [6,7,11] , C&K metrics [4] , other multiple OO metrics and Code-refactoring have some relation with each other.

Since refactoring was first investigated as a maintenance discipline in the late 1990's, it has moved toward becoming a vital part of an engineer's tool-set and generated numerous refactoring experimental studies. Seventy-two types of refactoring were described by Martin Fowler, in his book [14], which includes renaming, conditionalstatements, structural modifications and many more coding areas.

vii

The objective of this study is to calculate object oriented metrics[1,3,4,5,6,7,9,10,11] which can be further used with JArchitct tool and ref-finder to correlate it with software maintainability . In order to study, software repository of android application CALENDAR [22] is used. The motive is to generate a data set of a repository to measure the maintainability index of software based on object-oriented software metrics using the JHawk [19] tool. And using JArchitect2018.1.0 (demo) tool [20] for code smells and refactoring extraction. Then analysing the relation between object oriented metrics change with maintainability index and impact of refactoring on maintainability. The result shows that the dataset is successfully generated, and coderefactoring is more in those components which have low maintainability index.