QUESt: new practices to represent hypotheses in experiment-driven software development

Abstract

Recent studies have proposed the use of experiments to guide software development in order to build features that users really want. In this context, product assumptions should be taken as hypotheses to be tested through experiments. User stories (US) are broadly used in the agile context but current guidelines to write them, like INVEST (Independent, Negotiable, Valuable, Estimable, Small and Testable), are not suitable in a experiment-driven context. In this paper, we present the first cycle of a design science research to tackle this problem. We proposed QUESt, a quality guideline recommending US to have a Questioning sense, be Updatable, Evaluable and Straightforward, and a new template to write user stories. To evaluate these new artifacts, we performed a think-aloud evaluation with software development practitioners. Although the results do not confirm the artifacts effectiveness, they indicate that they have value and should be tested within a comprehensive framework complemented by other practices.