Model-Based Time and Cost Estimation in a Software Testing Environment
Bruce Benton · 2009
This paper presents a systematic approach for estimating time and cost information in a software testing environment. Models are powerful analytical tools, particularly well suited to solving complex, multivariable problems with time variant data. Software cost estimation certainly meets these criteria. This paper outlines how basic modeling techniques can be applied to this problem space. While there are some certain inescapable uncertainties in initial project estimations, by leveraging work repetition patterns in projects and the implementation of certain best practice estimation techniques it is possible to achieve improved quality and consistency of estimates, reduced estimate uncertainty earlier in the project, and increased accuracy of estimates over time. The model-based estimation system described in this paper is significantly less complicated than many formal estimation models (such as COCOMO II) and can realistically be implemented across organizations in a relatively short time.