Code-based analysis of the development effort of a large scale courseware project

This paper investigates proposed code-based measures for multimedia courseware as part of an on-going program of research into effort estimation models. The proposed code-based size measures are analysed using source code and project data from WinlEcon courseware written in the ToolBook authoring language. Statistical analysis of a priori, ToolBook specific and portable subsets of 47 code measurements are presented along with proposed predictive models of courseware development effort based on a priori, best subset an’d portable subset modelling assumptions.