Help:Deprecation
This page in a nutshell: Often, property values in Wikidata should be ranked as deprecated, not removed. Consider whether including the data is still useful. |
Statements in Wikidata should be ranked as deprecated (and not removed) if they are:
- superseded (as opposed to "outdated"; see note on 'end date', below)
- now known to be wrong, but were once thought correct
Warning: All statements, including deprecated ones, must be verifiable. Deprecation is not an option for information that can not be publicly sourced, e.g. non-public personal information about subject of an item. Wikidata:Living people applies to all statements, including deprecated ones. |
Background
editThe Wikidata web frontend has a different role than the Wikipedia web frontend has. At Wikidata, data users typically do not use the web frontend (they use the Wikidata Query Service (Q20950365) or Wikipedia templates/modules), thus the web frontend is basically a tool for editors only and it is perfectly fine and desirable to have "wrong" or "multiple" values in items, if ranks and sources are properly used. In contrast, Wikipedia readers use the same frontend as Wikipedia editors do, which implies to remove "wrong" values from articles in most cases.
Marking statements as deprecated instead of simply deleting them has several benefits:
- it allows other users to know not to re-add the value to the item
- such statements ("known to be untrue", "is not") are necessary in open-world assumption (Q851949) systems, such as Web Ontology Language (Q826165), by removing "trivial" claims we reduce usage of Wikidata for other purposes
- it provides a mechanism for representing the evolution of theories and ideas and thereby creates a richer context for understanding human knowledge. This is especially valuable in automated classification systems along with different from (P1889) and similar properties
- it upholds and establishes the integrity of Wikidata as a secondary knowledge base (that collects and links to references), rather than a primary database of facts. Wikidata simply provides information according to specific sources; those sources may or may not reflect contemporary thought or scientific consensus
At this point it should also be mentioned that deprecated statements are not visible for data users, unless they are explicitly asked for. This applies to the Query Service as well as to Wikipedia parser functions (such as {{#property:}}
and {{#statement:}}
). Superseded or wrong data, which is correctly tagged with deprecated rank at Wikidata, does therefore not pollute Wikipedia infoboxes, etc.
Reason for deprecation
editA deprecated value should always have a reason for deprecated rank (P2241) qualifier - see values already used for this (some may be sub-optimal). Some of the potentially useful ones are:
- cannot be confirmed by other sources (Q25895909) - general use
- not been able to confirm this claim (Q21655367) - general use
- conflation (Q14946528) - general use. Refer to Help:Conflation of two persons for guidance on how to use this Wikibase reason for deprecated rank (Q27949697).
- conflation of depiction and metadata from different objects (Q115099570) - more specific subset of conflation (Q14946528)
- incorrect value (Q41755623) - general use
- withdrawn identifier value (Q21441764) for identifiers
- person found to be alive (Q21124171) for date of death (P570)
- election result invalidated (Q25235916)
- withdrawn award (Q24629887)
- award subsidiary in rank to later award (Q41787617) - awards in an order of chivalry
- possibly invalid entry requiring further references (Q35779580) - general use
How to apply ranks
editSee Help:Ranking.
Outdated statements and 'end date'
editIf a value becomes out-of-date, for example:
- population change measured by a new census
- spouse no longer applies due to divorce
- position held no longer applies to due to retirement, or after an election
the value should not be deprecated, but instead an end time qualifier should be added.
Examples
edit- Karl Zilles (Q23015723) has two values for ORCID iD (P496); one is deprecated, with the reason withdrawn identifier value (Q21441764)
- Honoré de Balzac (Q9711) has two values for date of death (P570): 18 and 19 August 1850. According to La Mort de Balzac (Q681680) and most other sources Balzac died in the evening of August 18. The reference supporting August 19 claim is Integrated Authority File (Q36578) by German National Library (Q27302), which is usually a reliable source and which does not specify where the information come from. August 19 claim is tagged as deprecated, with the reason incorrect value (Q41755623). Interestingly August 17 is also mentioned as date of Balzac death in trivia-library.com, which is not mentioned in Honoré de Balzac (Q9711) as not considered a reliable source.