A Process for Evidence-Based Engineering of Domain-Specific Languages
Michael Felderer, Fabian Jeschko · 2018
Domain-specific languages (DSLs) are mainly designed ad-hoc and gut feeling resulting in languages that are often not well suited for their users and engineers. In this paper we develop a process for evidence-based language engineering to design domain-specific languages based on empirical evidence to support decision in language engineering. The developed process comprises an iterative execution of the phases DSL engineering, issue identification, data collection and evidence appraisal. We exemplify the concept by designing a DSL for Gherkin, a language test-driven acceptance testing in Xtext. The required evidence is derived by mining and analyzing all GitHub projects until July 1, 2017 that apply Gherkin.