Recent Advances in Software Reliability Assurance

Abstract

Achieving true service agility requires development teams to be able to continuously integrate and deliver software every few weeks. This requires capabilities for test automation, modelling and prediction of software reliability. In this paper, we report on our recent experiences applying a simple and novel curve shifting technique to software defect prediction for a continuous integration, continuous delivery project. The technique transforms the defect arrival curve from a given previous release using the user story (or feature)1 development plan so as to predict defect arrival for the required release. We also discuss the different views on software defects from a quality vis-a-vis project management perspective and how the proposed technique applies to either.