Conditions For Sizing An Algorithm Using Function Points

Software estimation is an important activity of software development while the algorithm in the software has always been a challenge that the software estimation facing. The workload and cost estimation of algorithm design can be made in different stages while the software size measurement of algorithm is the basic work in them.This article uses the function points analysis to measure the scale of algorithm software (components) in the software. Firstly, it makes research on the basic characteristics and measurement problems of algorithm software (components) and states the measurement process and implementation rules of IFPUG method and COSMIC method. Deeply analyze the metrics adaptability of IFPUG and COSMIC methods on the algorithm software(components) and find out the basic method and frame work of the two methods measuring two methods to do the algorithm software(components) scale measurement.On the basis of the above, the thesis make improvement about the deficiency existed in IFPUG and COSMIC: Combining the advantages of IFPUG and COSMIC, make comparisons between the pros and cons before and after improvement by way of algorithm instances. The results reveal that the algorithm results after improvement are more reasonable and accurate.This article makes useful exploration about the improvement of traditional FPA method and provides a good reference for the comprehensive and scientific implementation of FPA method.