User:Beireke1/sandbox/Intentional Sitelinks

Guidance for splitting a conflated Wikidata item

edit
  • The disentanglement (or splitting) of conflated Wikidata items requires carefully sorting out external identifiers so they are accurately matched either to the organization item or to the building item. These external identifiers should not be blindly matched to Wikidata items simply based on their names.
    • While authority files such as VIAF, ISNI and GND are most likely to hold records for names of organizations, they may also hold records for names of places. Links to authority files should be validated with regard to the class of the item. For example, GND has distinct normative data for organizations (Normdaten: Organisationen) and geographics (Normdaten: Geografika).
    • Some external reference databases (authority files) do not specify whether they are referencing an organization or an architectural structure (example). These external identifiers may be qualified with {{statement||P4390|Q39893184}} to indicate they are not an exact match.
  • The splitting of a Wikidata item involves sorting sitelinks: which Wikidata should be linked to which Wikipedia article(s). This is important to ensure connectivity both between different Wikimedia projects and between Wikipedia language versions. Depending on the main focus of each article, the various articles in different languages that originally pointed to the same conflated Wikidata item may in the future point to different Wikidata items. As a consequence, some interlanguage links between Wikipedias may be broken. The ideal situation that should be obtained when sorting sitelinks is:
    1. Ensure that each Wikipedia article links to the Wikidata item that is the main focus of the article.
    2. Ensure that each Wikidata item links to a Wikipedia page on which the concept is described, even when the article covers multiple Wikidata concepts and even when the main topic is covered by a different Wikidata item. As a Wikipedia article page cannot be sitelinked from more than one Wikidata item, this causes a problem that can be solved with an intentional sitelink to redirect (Q70894304). Simply said, it means that first a redirect Wikipedia page should be created to link the item to. The Wikipedia redirect page should point to the main Wikipedia article that covers multiple Wikidata items. The Wikidata item that is the main focus of the article should have a normal sitelink to the Wikidata article page. The second Wikidata item should have an intentional sitelink to redirect (Q70894304).
  • The splitting of a Wikidata item may also require sorting out links with Wikimedia Commons. Wikimedia Commons is more likely to hold images representing buildings than images of organizations, but it is better to verify the content of Commons categories and galleries before making a decision on a link with Commons.
    • If a category doesn’t have clear focus and holds mixed content representing both the building and the organization, then it may be necessary to split it into two distinct categories at the same time as correcting Wikidata and Wikipedia.
    • If necessary,   and   statements should be moved to the Wikidata item that matches most closely the focus of the category or gallery.
    • The sitelinks to Commons should be updated to match these statements. If there is a   or   statement but no sitelink, a sitelink should be added. Sitelinks are more important than P373/P935 because they are bidirectional and they are used to power infoboxes. Sitelinks to Commons are found under the “Multilingual sites” heading, at the bottom of Wikidata items.
    • Wikipedia articles may include manual links to Commons categories. Those sitelinks should be edited or removed if necessary.
  • This modelling issue was discussed in the Project chat in December 2021. You can read it at: Wikidata:Project_chat/Archive/2021/12#Could_a_new_constraint_help_dealing_with_conflation_issues?
  • Help:Conflation
  • Help:Split_an_item