Data-driven prototyping via natural-language-based GUI retrieval

Abstract Rapid GUI prototyping has evolved into a widely applied technique in early stages of software development to facilitate the clarification and refinement requirements. Especially high-fidelity shown enable productive discussions with customers mitigate potential misunderstandings, however, benefits applying prototypes are accompanied by disadvantage being expensive time-consuming requiring experience create. In this work, we show RaWi , data-driven approach that effectively retrieves GUIs for reuse from large-scale semi-automatically created repository mobile apps on basis Natural Language (NL) searches improve its productivity leveraging vast knowledge embodied repository. Retrieved can directly be reused adapted graphical editor . Moreover, present comprehensive evaluation methodology (i) systematic NL-based ranking methods through novel high-quality gold standard conduct an in-depth traditional IR state-of-the-art BERT-based models ranking, (ii) assessment extensive user study practical environment.