AN EMPIRICALLY TESTED DESIGN PATTERN SELECTION FRAMEWORK BASED ON DEVELOPER EXPERIENCE
Eddy Khonica, Imran Medi, Muhammad Ehsan Rana · Zenodo (CERN European Organization for Nuclear Research) · 2020
In most IT projects, software maintenance had always been a difficult task to perform especially when the software is not designed properly. This has led to numerous changes all over the places despite a minor change request. Design patterns are widely known for its proven solution to recurring problems. If fully utilized, design patterns can prevent such incident from happening as changes can often be added without affecting existing components. However, due to numerous design patterns that are available these days, it is difficult to select the right pattern. In addition, existing studies still lack a proven guideline to select the right pattern. This difficulty is faced by many beginners and experienced developers. The aim of this research is to propose a framework to select the most suitable design patterns in software development. The research methodology used is quantitative research where primary data are collected using survey and questionnaires. This is mainly to find out developers’ attitudes towards the use of design patterns. 25 out of 48 papers regarding design patterns are referenced in this paper. The framework developed is evaluated and empirically tested using expert judgement and controlled experiment. The result shows that the framework does help in evaluating and selecting the most suitable pattern. Therefore, the authors highly recommend developers to make use of this framework in software development.