Measuring Technical Debt Of Requirement Change By Marginal Contribution

In the software life cycle,the requirement of software system is changing constantly,and new requirement appears frequently.When requirements change occurs,the impact on existing software systems is difficult to assess,especially for larger systems,it is difficult to determine what impact requirements change may have on the whole system or part of the system.If these effects are ignored,the implementation cost of requirement change may be several times higher than expected.Although requirement engineering is critical to the success of software projects,most requirements decisions often depend on the preferences and balances of developers,lacking a systematic and clear management method.In order to measure and evaluate the impact of requirement changes,we introduce the concept of technical debt.Technical debt is a trade-off between short-term and medium-term benefits of software development projects and long-term software quality.Software technology debt uses the concept of "debt" in economics to describe the technical compromise implemented in software development in order to realize the short-term benefits of the project.However,from the long-term goal,technology debt will affect the quality,cost and development efficiency of the software,so it is necessary to manage it systematically and effectively.Technical debt initially only involves software implementation(i.e.code level),but it has gradually extended to software architecture,detailed design,even documentation,requirements and testing.Based on its interdisciplinary nature,namely software engineering and economics,the study of technology debt management should strike a balance between software engineering and economic theory.Aiming at the technical debt caused by the changing requirement in the software life cycle,this paper puts forward a definition of the technical debt caused by the changing requirement,quantifies the technical debt caused by the changing requirement,obtains the marginal contribution of the changing requirement with the help of the idea of "marginal contribution" in economics,provides the basis for the realization order of the changing requirement,and then uses the marginal benefit analysis method to provide the requirement.The realization value of change provides reference.Taking HADOOP as an example,this paper quantifies and ranks the marginal contribution of requirement change by defining and testing the technical debt of requirement change,and provides technical support for the realization order and mode of requirement change.It is applied to Spring Framework of large open source project.The availability and feasibility of the concept of marginal benefit of requirement change are verified by experiments.Finally,GDBT algorithm is used to study the history of requirement change in Spring Framework,and a method of analyzing marginal contribution ability of requirement change is proposed,and the importance of each field value to its marginal contribution is ranked.The results show that the analysis method can provide valuable results for requirement engineers to measure their workload and risk.