Improvement Of Software Size Metrics Based On Fpa Method

As the popularization of computer application, the size and complication of application software system is more and more great. To get the directive understanding to the project at the early development phase, the software size measurement become-one of the key technology of software engineering. One of the measurement method is the objectfor research of this article--FPA(Function Points Analysis). By researching, the FPAmethod has the following shortages. Firstly, the result of measurement is not very exact to the concrete type of software. Secondly, though the FPA method is suit for the project of MIS type, the method still has some limitations. For example, there are some errors in complicate page when we compare the measurement result by using FPA to the actual development time. So we should improve on the shortages of FPA method.In this article, author advanced some improvement about the shortages of FPA method. The first, get a particular partition about the original Data Function Points and transaction Function Points computing matrix. The second, put the function points computing matrix to numerical value analysis, and replace the function points computing matrix with a conversion function which convert RET, DET to function points. In this article, we compare the original FPA method and the improved FPA method with an actual software project. What the result is that the improved FPA method is more accurate in the page with complicate business logic and large number of data element than the original FPA method, and the result is closer with the actual development time.