Page MenuHomePhabricator

Umbrella for user stories: Wikidata <-> external sources
Open, Needs TriagePublic

Description

This task is problem stated that will be broken down to smaller User cases

Problem statement
When working with linked data in a complex environment we need to have a transparent process

  1. Wikidata modelling
    1. modelling decision taken
      1. it should be easy to understand why we have two objects with same name and same location. Compare called ensemble of buildings Q30311220 and Q10661178 same name and same location why do we have 2 objects with different instance of?
  2. Metadata about data sources
    1. quality of the source
      1. for more "popular people" we get
        1. more sources ex. stating the birth date. All external sources are not primary sources and have the same quality somehow this needs to be communicated to the reader
        2. precision on dates creates confusing data. Why is fact with a reference saying a person is born YYYY and another one YYYY-MM-DD would be nice for the reader to understand quality, decision taken or not taken when importing data
    2. the trustworthiness of the source
      1. document some metadata about a source if they have a qualityprocess, changeprocess....
    3. why we trust source A better than B
  3. the change process
    1. how do we handle changes with this external data source
    2. for a specific fact see when and why it was changed with a reference to the change process of that data provider and if possible with a lognumber of the external system