An integrated software cost model based on COCOMO and function point approaches

Software cost estimation is important in information systems management and is commonly taught in software engineering courses. The paper re-examines the key features of two well-known "public domain" methods, COCOMO and function point analysis, and attempts to combine the advantages of these two approaches into a single model. To address the problem of language dependency, language-weighted function point is introduced. In addition, continuous, rather than discrete, "cost drivers" are used. A data set from 81 software projects was used to test the model. Cost estimates so obtained compared favourably with results from other more elaborate methods. It is argued that the conceptual simplicity of this integrated model makes it suitable for inclusion in software engineering courses. Also the model renders itself flexibly in tailoring to different realistic cost estimation environments.<>