Repository Versioning

[hhkitsune]

Description

Gathering information to help refine our concept of versioning for the research data repository.

Live Examples

ScholarSphere (Penn State)

Live example was used for Discovery Versioning ticket; RDOS ticket #2.

• Versions are whole numbers with letter V (e.g. V1, V2, V3)
• The current version is the default
• Selecting a previous version will result in a yellow banner notification
• DOI does not change

Zenodo

Live example was requested by RDSS to show versioning for Zenodo.

• Has over 100 versions!
• New DOI for each version, but has an additional DOI to cite all versions
• Selecting a previous version will result in a yellow banner noticfication
• Versioning for Zenodo explained
  • "...cite either specific versions of a record or to cite, via a top-level DOI, all the versions of a record."

Considerations

RDA Data Versioning (WG & IG; as noted/paraphrased from presentation)

Reasons identified for versioning:

• Reproducibility
  • relies on precise ID of extract of data that is used
• Authority, Identity and Ethics
  • unambiguous ID; understand which version has authority (example: most recent version)
  • correct citation is ethical and prevents confusion
• Data Quality
  • can attach data quality measures to different versions (see possible example in video at 45:11)
• mentioned other reasons possible

Resources

• Klump, J., Wyborn, L., Wu, M., Martin, J., Downs, R.R. and Asmi, A., 2021. Versioning Data Is About More than Revisions: A Conceptual Framework and Proposed Principles. Data Science Journal, 20(1), p.12. DOI: http://doi.org/10.5334/dsj-2021-012
• RDA Data Versioning WG (note: this group uses the FRBR model which has terms that can be confused with terms from other models. I would advise focusing on what this group is defining, rather than becoming fixated on the labels):
  • Revised Report of the RDA Data Versioning WG
  • Revised Data Versioning Use Cases
  • RDA DVIG to W3C (mapping)
  • RDA DVWG Webinar
  • RDA DVIG Webinar