COSMIC-FFP 기반 소프트웨어 개발노력 추정에 관한 연구

Abstract

A study on Software Development Effort Prediction Based on COSMIC-FFP : Focus on improved software development effort estimation model using weight

Sang-Ki Park

Interdisciplinary Program of Information System, The Graduate School, Pukyong National University

Abstract Successful project is to develop the system which user is satisfied to by using of restricted manpower, period and cost. We chose a software size as a unit to estimate correct development effort, development costs and development period estimation. It is possible for the accurate estimation of software size to allocate correct resource division and to definitely influence the software quality. Consequently, The accurate estimation for development effort, cost and period is the essential element to accomplish the project successfully. It is rather efficient that the estimation of software size is carried out at the early stage of the development. Line of code has been chiefly used to estimate the software size in korea because developers are easy to understand LOC and its meaning is clear. There are shortcomings in estimating LOC. First, It is hard to estimate LOC In the early stage of the software development. Second, LOC is influenced by the development environment. The recent trend of software size estimation method prefers FPA to LOC. There are the problems of FPA. First, there exist measurement error between the actual value and the estimates. Second, it is impossible to apply for real-time system and embedded system. However, The recent information system is getting larger in the portion of an embedded system or real-time system gradually. Therefore, COSMIC is the established to cover for these disadvantages of function point in 1998 and COSMIC-FFP is developed for estimating real-time system and embedded system. There have been many researches on the software size estimation methods based on COSMIC-FFP, but there are a few models for estimation the development effort and it is necessary to develop the model. The COSMIC-FFP technique has some advantages in appling several areas such as data management, real-time system, algorithmic software and others, but on the other hand, has some disadvantages in imposing the weight on function elements needed for estimating sizes. It is necessary for new method added complexity to COSMIC-FFP to quantify correctly the software size. Consequently, First, we measure FP and COSMIC-FFP through the case study and propose the transfer model with the above two measurement results in this paper. Second, we propose estimation model for development efforts based on COSMIC-FFP by using of data from the case study and the weight-added model based on complexity to raise the correctness of the proposed model. We estimate the COSMIC-FFP according to the proposed model and evaluate the estimation model of development effort based on the weight-added COSMIC-FFP. Since the proposed models in this paper are based on the results from carrying out the actual project, it is possible to estimate the development period, costs and efforts correctly. In the future, we will develop the general-use model through collecting various kind of an experiment data.