Wikidata/Notes/Bibliographic data

Wikidata support for bibliographic metadata

Berlin Hackathon 2012, 2012-06-03

Participants

• Daniel Kinzler (Wikimedia DE)
• Daniel Mietchen (OKFN DE)
• Dario Taraborelli (Wikimedia Foundation)
• Martin Fenner (Public Library of Science)

References

• https://meta.wikimedia.org/wiki/Wikidata/Notes/References
• http://lists.wikimedia.org/pipermail/wikidata-l/2012-May/000540.html

Questions

• What is the main use case for having source metadata accessible in Wikidata?
  • providing source references for property values on wikidata
  • providing source references in Wikipedia text
• Can we use wikidata to store source metadata and address individual sources as entities?
  • yes.
  • it's unclear if it's desirable to have lots of bib metadata on wikidata if it's not used to source claims
• What are dependencies that will need to be addressed before we can store source metadata?
  • one major issue is the ability to address items using arbitrary unique identifiers (like DOI or ISBN)
• Should we focus on generic sources or a subset of sources with unique ids like scientific papers for DOIs?
• Will it be possible to address a source in Wikidata via its DOI?
  • that is desirable, but DOI should not be a special case. there should be a generic mechanism for arbitrary unique identifiers.
• How will source metadata be seeded to Wikidata (on demand or by bulk import)?
  • on-demand is the typical wiki way. often this would be done semi-automatically
  • bulk imports require the ability to exclude trusted imports from the review queue
• How will Wikimedia projects migrate from the current Template:Cite templates to Wikidata links?
  • probably by bot
• How will provenance for bibliographic metadata be handled in wikidata?
  • if metadata is imported from an authoritative source, how does this fact contribute to the provenance info
• What kind of curation model will apply to metadata provided by the main DOI-provider authority (CrossRef)?
• What kind of properties (and at what level of granularity) should source entries store on top of bibliographic metadata?
• Will we be able to represent license and access information to source entries?
  • yes
• What kind of policy could support the introduction of source entries in wikidata?
  • it's likely that there will be restrictive policies about modifying any information imported from an authoritative source
• Does it make sense to implement "data templates"?
  • like item references, but (partially) shown inline... lots of technical issues to consider.
• Should users be able to modify statements (property values) imported from authoritative shources?
  • yes, but any modification of an existing statement should be brought to the special attention of reviewers. This should only happen in the rare case that information from the original source was misrepresented (imported incorrectly). Other cases would typically be handled by adding a statement, not modifying one.
• Will wikidata provide statement-level data protection?