A modified function point method for CAL systems with respect to software cost estimation

Abstract

Algorithmic models for software cost estimation usually assume that software size is the major factor that influences software development effort. Function point analysis (FPA) is a widely accepted method to measure software size. However, when we used FPA to size computer assisted learning (CAL) systems for development effort estimation, we found that FPA has a number of difficulties. The study examines the special characteristics of CAL systems and analyses the additional considerations needed if FPA is used to size CAL systems for effort estimation. The concept of effective size is proposed to account for software reuse. An integrated software cost model previously introduced by the authors is adapted to the CAL environment. Real world data are used to test the adapted model, and the results and implications are discussed.