- Wikidata aims to provide a centralized data repository that will feed all the wikipedias, as well as any other platform that wants to use the data. (Currently, each wikipedia - English, French, Russian, etc. - is maintained independently.)
- Two main goals:
- Ease curation and support smaller wikipedias
- Enable queries like "What are the ten largest cities with a female mayor"?
Phase 1: Language links recently: every language links to every other language wikidata: one page for each entity, list representations in each language wikipedias pull language links from wikidata
Phase 2: Infoboxes currently: each article calls an infobox with values wikidata: centralize the values wikipedias will call the infobox, and populate with values from Wikidata
Phase 3: Queries
- Why not use SMW?
Wikidata aims to do for data what WikiCommons does for images.
SMWs assumed to be of one language. Possible to build with I18n, but hairy. Wikidata has internationalization built in from the beginning.
- Advantages of SMW:
- Querying capability - http://semantic-mediawiki.org/wiki/Help:Inline_queries#Parser_function_.23ask
- Semantic Results Format - http://www.semantic-mediawiki.org/wiki/Semantic_Result_Formats
- Mixing of structured and unstructured data.
- Advantages of Wikidata:
- Built-in provenance
- Can we satisfy our primary use case with SMW?
Probably, by using Approved Revs
- Overlap of data models
- One page = one thing.
- Information is being stored, mutatis mutandis, as triples: <the subject of a wikipedia page> <has property> <some value>
- Overlap of personnel: Overlap of personnel: Denny Vrandecic, Jeroen de Dauw, Markus Krotzsch
- Overlap of development paths: https://meta.wikimedia.org/wiki/Wikidata/Notes/SMW_and_Wikidata