Enabling automated requirements reuse and configuration

TLDR: Zhang et al. as mentioned in this paper propose a methodology with tool support, named as Zen-ReqConfig, to tackle the problem of automatically developing structured and configuration-ready PL requirements repositories.

Abstract: A system product line (PL) often has a large number of reusable and configurable requirements, which in practice are organized hierarchically based on the architecture of the PL. However, the current literature lacks approaches that can help practitioners to systematically and automatically develop structured and configuration-ready PL requirements repositories. In the context of product line engineering and model-based engineering, automatic requirements structuring can benefit from models. Such a structured PL requirements repository can greatly facilitate the development of product-specific requirements repository, the product configuration at the requirements level, and the smooth transition to downstream product configuration phases (e.g., at the architecture design phase). In this paper, we propose a methodology with tool support, named as Zen-ReqConfig, to tackle the above challenge. Zen-ReqConfig is built on existing model-based technologies, natural language processing, and similarity measure techniques. It automatically devises a hierarchical structure for a PL requirements repository, automatically identifies variabilities in textual requirements, and facilitates the configuration of products at the requirements level, based on two types of variability modeling techniques [i.e., cardinality-based feature modeling (CBFM) and a UML-based variability modeling methodology (named as SimPL)]. We evaluated Zen-ReqConfig with five case studies. Results show that Zen-ReqConfig can achieve a better performance based on the character-based similarity measure Jaro than the term-based similarity measure Jaccard. With Jaro, Zen-ReqConfig can allocate textual requirements with high precision and recall, both over 95% on average and identify variabilities in textual requirements with high precision (over 97% on average) and recall (over 94% on average). Zen-ReqConfig achieved very good time performance: with less than a second for generating a hierarchical structure and less than 2 s on average for allocating a requirement. When comparing SimPL and CBFM, no practically significant difference was observed, and they both performed well when integrated with Zen-ReqConfig.