Software cost estimation; Tata Information Systems Limited

Abstract

In an age when software forms a major part of any computer system, the importance of accurately estimating the effort, time and cost of a software system cannot be overemphasized. Software estimation has come of age. Managers are increasingly using scientific methods instead of relying on hunches. This project is an effort to understand software cost estimation techniques thoroughly and adapt some of the popular estimation techniques to the Indian conditions. It will also suggest a suitable format for collecting project data which will be used for continuously recalibrating the estimation models. The models which have been recalibrated are: 1. The COCOMO 2. The Bailey-Basili Meta Model Though it was decided to recalibrate the Putnam Model too, lack of information on the model forced the model to be dropped. Another objective was to develop a computer-based software cost estimation system for TISL, to be used for estimating its projects. A simple system has been developed which may be used for cost estimation. This system may be modified to lend it more aesthetic looks and user-friendliness. The project involved a literature survey, collection of project data from the software industry, analysis of the data, design and development of the computer-based system.