Background: Establishing traceability from requirements documents to downstream artifacts early can be beneficial as it allows engineers to reason about requirements quality (e.g. completeness, consistency, redundancy). However, creating such early traces is difficult if downstream artifacts do not exist yet. Objective: We propose to use domain-specific taxonomies to establish early traceability, raising the value and perceived benefits of trace links so that they are also available at later development phases, e.g. in design, testing or maintenance. Method: We developed a recommender system that suggests trace links from requirements to a domain-specific taxonomy based on a series of heuristics. We designed a controlled experiment to compare industry practitioners' efficiency, accuracy, consistency and confidence with and without support from the recommender. Results: We have piloted the experimental material with seven practitioners. The analysis of self-reported confidence suggests that the trace task itself is very challenging as both control and treatment group report low confidence on correctness and completeness. Conclusions: As a pilot, the experiment was successful since it provided initial feedback on the performance of the recommender, insight on the experimental material and illustrated that the collected data can be meaningfully analysed.

翻译：背景：从需求文档到下游工件的早期可追溯性建立具有积极意义，工程师可借此对需求质量（如完整性、一致性、冗余性）进行推理。然而，当下游工件尚未生成时，建立此类早期追溯存在困难。目的：我们提出利用领域专用分类法建立早期可追溯性，提升追溯链的价值与感知效益，使其可在后续开发阶段（如设计、测试或维护）发挥作用。方法：我们开发了一个基于多启发式规则的推荐系统，用于建立需求与领域专用分类法之间的追溯链。通过控制实验，比较工业从业者在有无推荐系统支持下的效率、准确度、一致性和置信度。结果：我们组织七名从业者完成了实验材料试点测试。自评置信度分析表明，由于对照组和实验组均对答案正确性与完整性评价偏低，追溯任务本身具有极高挑战性。结论：作为试点实验，本次研究取得阶段性成功，获得了推荐系统的初步性能反馈，积累了实验材料优化见解，并验证了数据收集与分析的有效性。