Is My Software "Good Enough" to Release? - A Probabilistic Assessment

We present the basics of a probabilistic methodology to assess the overall quality of software preparatory to its release through the evaluation of process and product evidence, the 'Good Enough' to Release" (GETR) Methodology, in this paper. GETR Methodology has three main elements: a model whose elements represent activities and artifacts identified in the literature as being effective assessors of software quality, a process for populating certain parts of the model, and methods for analyzing the importance of contributions made by individual evidence to the determination of overall system quality. First, the methodology's components are briefly introduced. A demonstration of how the methodology can be applied is then given through two case studies reviewing release assessments for inhouse developed analytical tools. The robustness of the model is also illustrated by the results of the case studies.