Journal of Software Reproducibility Guide
Establishes experiment reproducibility, data and code availability, and artifact standards for submissions to the Journal of Software (软件学报).
The Journal of Software Reproducibility skill provides comprehensive guidance for researchers submitting computer science papers to the Journal of Software (《软件学报》, JOS). It helps authors avoid weak 'available upon request' statements by establishing clear data and code availability declarations, fixing environments with containerization and lockfiles, controlling random seeds, and locking dataset provenance or LLM API outputs. Designed to make empirical evaluations trustworthy and verifiable, it includes self-check checklists, directory structure templates, and specific advice on articulating reproducibility across methodology, experiments, and threats-to-validity sections.