一句话：提出将架构恢复工具系统应用于开源仓库，以构建大规模架构数据集的愿景。
软件架构研究日益依赖数据驱动方法，但大规模、多样化的架构数据集稀缺；专家标注数据质量高却昂贵且覆盖有限。本文提出通过系统性地将架构恢复工具应用于开源软件仓库，实现架构数据民主化的愿景。作为扩展摘要，文章讨论三方面：大规模架构数据集可能开启的新研究方向；自动恢复架构的数据质量保障策略；以及构建可扩展的持续数据收集流水线的方法。若该愿景实现，架构研究有望从数据稀缺的案例研究转向大规模经验科学。
The software architecture research community increasingly relies on data-driven approaches in research, yet faces a key bottleneck: the scarcity of large-scale, diverse architecture datasets. While expert-curated architecture data provides high quality, it remains expensive and limited in scope. We propose a vision for democratizing architecture data by systematically applying architecture recovery tools to open-source software repositories. This extended abstract discusses three key aspects: (1) the potential impact of large-scale architecture datasets on enabling new research directions, (2) strategies for ensuring data quality in automatically recovered architectures, and (3) approaches for building scalable pipelines for continuous data collection. Realizing this vision could fundamentally transform architecture research from data-scarce case studies to large-scale empirical science.
@inproceedings{paperbot11514,
title = {Scaling Software Architecture Data: A Vision for Leveraging Architecture Recovery on Open-Source Projects},
author = {Yan He and Guangjie Li and Yonghui Wu and Changqin Li and Yi Tang and Yufu Wang and Ge Niu and Jie Bai},
booktitle = {ICSE},
year = {2026},
doi = {10.1145/3774748.3795869}
}