Select standard file formats to facilitate access and contribution by humans and software (Control)

Use widely adopted, non-proprietary file formats to store and exchange review data, including meta-data, protocols, evidence tables, and outputs.

Why this matters

• Tool independence and longevity: Standard formats reduce lock-in and keep the evidence accessible across multiple tools, or even when individual tools evolve.
• Quality control and validation: Human- and machine-readable formats support automated validation (schema checks, content validation) while enabling transparent human oversight (e.g., through readable Git diffs).
• Reproducibility and traceability: Clean text-based formats integrate well with version control, so changes can be reviewed and compared efficiently.

Practical implementation

• Use BibTeX (.bib) for bibliographic metadata and record-level information.
• Use YAML (.yml / .yaml) or JSON (.json) for structured artifacts such as search strategies and extracted evidence, with a preference for YAML due to its readability in version-controlled workflows.
• Prefer plain-text formats that support stable diffs in Git, so even small changes remain visible and reviewable.
• Document expected fields and conventions (e.g., required BibTeX keys; JSON schemas for search strategies) to ensure consistency across contributors.

Resources

• The Turing Way: Research Data Management
• The Carpentries: Managing Data