Using Gherkin for interaction design testing and monitoring

Abstract

In software testing, Controlled Natural Languages (CNLs), such as Gherkin, are sometimes used to help clients to communicate with developers and draw up specifications. These Gherkin specifications are then translated to tests. Since testing is not exhaustive, using runtime monitoring on critical systems helps ensure that programs run as they should while deployed. Runtime monitoring is not yet (widely) adopted by industry 13) and in this work, we are looking into automatically translating tests into runtime monitors. We take the case of safety critical number entry systems design in medical devices. Interaction designers and psychologists, study how to best design these interfaces such that the likelihood of human error is reduced. Non-technical designers then use a CNL to communicate their findings with software developers/testers, who will then draw up unit tests. In this abstract we propose to automaticaJiy translate these tests to runtime monitors in order to ensure that the system works correctly when deployed, without any additional cost to the manufachirer. Finally, we show the corresponding automata for the runtime verification tool, LARVA