Versioning Data Is About More than Revisions: A Conceptual Framework and Proposed Principles

A dataset, small or big, is often changed to correct errors, apply new algorithms, add data (e.g., as part of a time series), etc. In addition, datasets might be bundled into collections, distributed in different encodings mirrored onto platforms. All these differences between versions need understood by researchers who want cite the exact version dataset that was used underpin their research. Failing do so reduces reproducibility research results. Ambiguous identification also impacts and centres are unable gain recognition credit for contributions collection, creation, curation publication individual datasets. Although means identify using persistent identifiers have been place more than decade, systematic versioning practices currently not available. this work, we analysed 39 use cases current across 33 organisations. We noticed term ‘version’ very general sense, extending beyond common understanding refer primarily revisions replacements. Using concepts developed software Functional Requirements Bibliographic Records (FRBR) conceptual framework, six foundational principles datasets: Revision, Release, Granularity, Manifestation, Provenance Citation. These provide high-level framework guiding consistent practice can serve guidance providers when setting up own revision protocols procedures.