Software maintenance cost estimation with fourth generation languages

Abstract

This thesis addresses the problem of allocation of software maintenance resources in a commercial environment using fourth generation language systems. The activity of maintaining software has a poor image amongst software managers, as it often appears that there is no end product. This image will only improve when software maintenance can be discussed in business terms, one of the main reasons being that the maintenance costs can then be compared to the costs of not maintaining the system. Software maintenance will continue to exist in the fourth generation environment, as systems will still be required to evolve. Cost estimation is an imprecise science, as there are many variables such as human, technical, environmental and political which can effect the ultimate costs of software and the resources required to maintain it. Some of the factors appear more obvious than others, for example an experienced programmer can achieve a specific task in less time than an inexperienced one. To fully estimate software maintenance costs these factors need to be identified and weights assigned to them. This thesis examines a means to identify these factors and their weights, and produces the first cut of an equation which will enable the software maintenance resources in a fourth generation language to be estimated.