Quantifying estimating index by using risk factors of new software development tools

Abstract

Project managers and software engineers know that using new software tools introduces more or less risk to their projects. However, the amount of risk introduced is hard to quantify exactly by just relying on existing software estimation models and techniques since most of these do not consider the risk of using new software development tools from the perspective of un individual engineer or software tool. Neither do these models and techniques provide the specific information on how much implementation effort would be required at an individual component level. Although software complexity metrics provide a finer level of granularity than software estimation of a software component level, they overlook technical risk factors involved in implementation. This thesis proposes an approach to quantifying the technical risk of using new software tools, or Technical Risk Factor (TRF), and to weighing software complexity by adding the quantified technical risk, or Weighted Complexity Index (WCI). A TRF represents the magnitude of increased effort for developing a component with software development tools by reflecting the quantified risks of individual tool and the experience of developing staff. A TRF is expected to be a predictor for software development effort and a WCI represents the overall risk of coding and testing a component by combining its respective TRF and complexity metric so that it is expected to enhance the predictability of effort more than the complexity metric or TRF alone can. A case study of a studio project of the MSE program at ICU is conducted to present an example of using the method and interpreting the result by a statistical analysis. The case study presented here demonstrates that the two metrics, TRF and WCI, have significant predictability in estimating implementation effort at a software component level when the software project has to depend heaving on new software technologies.