
In late 2025, the Generalist Repository Ecosystem Initiative (GREI) collaboration surveyed its member repositories to better understand their current data packaging practices and to identify their views on data packaging and available packaging formats. This document contains a comparison matrix of the characteristics of five popular data packaging formats identified in survey results as currently or potentially implemented by GREI repositories: Data Package v2 (Frictionless Data), RO-Crate, BagIt, CSV on the Web (CSVW), and Croissant (MLCommons). These formats were examined using nine criteria: 1) Specification & Governance, 2) Primary Purpose / Model, 3) Technical Capabilities, 4) Tooling & Ecosystem, 5) Adoption / Community, 6) Performance & Scalability, 7) Extensibility, 8) Implementation Considerations, and their support of 9) Tabular vs. Non-Tabular Data (CITE MATRIX ON ZENODO). The resulting matrix provides a baseline, side-by-side view of various data package standards across governance, capabilities, tooling, adoption, and extensibility. It is not intended to be an evaluation or ranking of data packaging formats.
