Application Research Of IFPUG Method On Embedded Software

IFPUG is one of the recognized functional size measurement methods by International Organization for Standardization,and there are already a large number of software metrics project using this method.The data shows that IFPUG is better than COSMIC in applicability of the embedded software project metrics.Through the measurement of specific cases,this paper analyzes the results deviation of IFPUG causes,and puts forward the solutions.This paper studies on the IFPUG measurement method for embedded software,mainly done the following work:1.Find the results of the IFPUG on embedded software metric accuracy is not high.This paper use the IFPUG method and COSMIC method to measure the same case,we found the results deviation of IFPUG reached 62.9% from the actual results.And through the statistical analysis of measurement results of multiple cases,we proved the universality of the problem.2.Research the reason of the measurement accuracy.Through research the characteristics and difficulties of embedded software,this article find three reasons from the IFPUG method such as Data Function,Transactional Function and complexity divided three aspects,such as the cause of inaccurate measurements.3.Put forward specific measures for method improvement.In combination with the idea of COSMIC and start with IFPUG itself,we analyzed the theory of pertinence and feasibility in method first,and then we improved the complexity of the functional factors,in order to improve the accuracy of IFPUG results on embedded software.4.Verify the feasibility and effectiveness of the improved method.By comparing the measurement results with the actual function of the software from the programs before and after,the feasibility and effectiveness of the improved method has been verified.The result shows that the traditional method of IFPUG results in measurement of embedded software,the deviation range between 24.0% and 73.4%.By using the improved method,the deviation range reduced to between 3.8% and 38.8%,the average deviation from 68.5% down to 19.2%.