Wikidata

Wikidata:Requests for comment/Use of dates in the descriptions of items regarding humans

< Wikidata:Requests for comment
An editor has requested the community to provide input on "Use of dates in the descriptions of items regarding humans" via the Requests for comment (RFC) process. This is the discussion page regarding the issue.

If you have an opinion regarding this issue, feel free to comment below. Thank you!

The theme of the use of dates in the descriptions of items regarding humans (instance of (P31)human (Q5)) has been, as of now, debated many times and mostly with inconclusive results; descriptions in different languages contain dates in higher or smaller percent mainly according to the activities of single bots and to the personal preferences of individual editors. As of now, no hard rule generally imposes or banishes the use of dates in descriptions. While I tend to think that reaching some binding conclusion is definitely very very hard and maybe wouldn't effectively be desirable, at the same time it is true that periodically massive additions of dates or manual removals of dates have caused conflicts and edit wars. Thus, I think that at least some degree of consensus about this practice is advisable. I will try to sum up the previous arguments raised about this practice and to pose some general questions about it.

The longest and most participated discussion I know about this theme, although I can miss many others (please mention others if you know them!), was Wikidata:Project chat/Archive/2021/12#Should the description of a person include year of birth and death, if applicable?. Just a little earlier I started one at the Italian PC, Wikidata:Bar/Archive/2021/11#Descrizioni in italiano per elementi di persone. I will try to extract from these two discussions, especially the first one, some points (I mark objections/remedies in orange):

n° for the use of dates against the use of dates
1° standard method (see authority files) for disambiguation:
  • necessary for people with same name, same occupation and same citizenship;
  • somehow useful also for people with the same name;
  • for others, waiting until for an item regarding the second person with the same name to be created and continuously update descriptions may not be efficient
2° better than leaving descriptions blank, when very few information about the person is known (e.g. persons from genealogical databases)
3° duplicating information already stored in statements (and descriptions are not definitions but are meant for disambiguation)
4° usefulness for Wikidata users (mainly in Special:Search, both searchbox and search results), for tool users (e.g. PetScan, OpenRefine; it may help reconciliations) and for external data reusers objection: display issue, it could be addressed by software (see e.g. autodescriptions in Mix'n'match and possibly, in the future, multilingual descriptions on Wikidata based on the items' statements) and by data reusers themselves
5° proposed remedy for uncertain dates: omit questionable dates or mark them somehow (c., ca., ? etc.) legitimating questionable dates (posing an uncertain, or simply not-referenced, date in the description may legitimate it far beyond the actual supports it receives from the sources)
6° proposed partial remedy: queries could be used to spot cases of disagreements between description and statements and generate list of items deserving an intervention for the realignment of the two potentially leading to disagreements between description and statements; newbies could end up correcting dates only in descriptions (because it's easier)
7°
  • proposed partial remedy: sometimes solvable moving the dates leftwards (e.g. "politician (1910-1950), mayor of X");
  • otherwise, it could be tried a standardization in order to have dates in final position always representing birth-death dates (but that seems very difficult and still not devoid of ambiguity);
  • a third solution would be moving the dates always at the start of the descriptions (e.g. "(1910-1950) mayor of X")
 sometimes being confusing if always collocated at the end, e.g. for positions (e.g. "mayor of X (1910-1950)" is ambiguous, birth-death or start-end of the position?; "husband of X (1910-1950)" is ambiguous, birth-death of the man or of the woman or start-end of marriage?)

--Epìdosis 17:38, 19 May 2022 (UTC)[reply]

Questions edit

The main questions I think should be answered are the following ones:

  1. a hard rule of some sort (e.g. "always dates" or "dates only for homonym persons" or "dates only for homonym persons with same occupation and same citizenship"; others are possible) applicable to descriptions in all languages should be established? if it should be, which one?
  2. if a hard rule for all languages is not established, could a single language community on Wikidata establish a hard rule of some sort applicable to descriptions in its language?
  3. if a hard rule for all languages is not established, how should conflicts between users adding and users removing dates be solved?
  4. while each language has its peculiarities, in general (or, say, at least for English) should dates be placed: always at the start of the descriptions, or always at the end of the descriptions, or usually at the end of the descriptions but with exceptions for ambiguous cases?

Anyway, if you think other aspects of this general topic should be discussed, feel free to add questions in this paragraph and to discuss related topics in the discussion. --Epìdosis 17:38, 19 May 2022 (UTC)[reply]

Discussion edit

  • I had missed replying to this RFC, apologies if this is too late - I would agree with "no hard rule" for #1. For #4, I think we do already have some de-facto standards (usually at the end, usually in brackets or after a comma, years not dates, use a hyphen not 'to', etc); I don't think there would be any harm in encouraging these.
Personally I am very much in favour of having them - as someone who deals with a lot of historic data, I agree with Jan's point about it being very useful to quickly get a sense of which time period someone is in, and honestly I think the dates may be more useful to me than eg the nationality, which we do routinely include. Autodescriptions would be lovely ... but we've been saying that for years and they're still a long way off! Andrew Gray (talk) 22:02, 19 October 2022 (UTC)[reply]
  • I'm very much a supporter of "no DOB/DOD in descriptions", simply because it is much simpler to maintain if we keep descriptions short. Most items about humans do well with citizenship+occupation without any DOB/DOD as a disambiguator, so we do not need to add additional information. So the only "hard rule" I'd support would be generally no such information in descriptions, as long as there is no specific conflict with another item to disambiguate.
    Re. conflicts: we are probably not able to find a general consensus, so I suggest not to fiddle with existing descriptions. Do not add or remove dates when they are missing/existing; do not make edits to the style or position within the description if there is such information present; only iteract with this part of a description if it is incorrect, or technically necessary to disambiguate. Uniformity of descriptions across any larger area is unattainable anyways. —MisterSynergy (talk) 22:21, 19 October 2022 (UTC)[reply]
  • So, while I was very much in favor of no DOB/DOD, the recent discussions on the uses and misuses of country of citizenship (P27), and more importantly ethnic group (P172) for those figures living in times and places where citizenship is a shaky concept, make me wonder if using dates instead of citizenship/nationality shouldn't be very much favored as something less political and more objective. Sure, there are countless Qid were exact dates are unknow, but they are also quite frequently Qid for which citizenship is unknow or wholly anachronistic. I find myself way more at ease as describing someone from Antiquity as "ancient + occupation" instead of "ethnicity + occupation". --Jahl de Vautban (talk) 12:26, 20 October 2022 (UTC)[reply]
  • Always provide the dates of birth and death when known, to keep things simple. However, no one should be chastised for omitting the dates. But, once the dates have been added by someone, they should not be removed. One should get a quick summary of the context of the person from the description, and the year range or century does provide part of that context. --Dan Polansky (talk) 08:33, 21 May 2023 (UTC)[reply]
    Let me add that I care above all about English, the lingua franca; if some cultures prefer to omit dates in their descriptions, I feel it would be fine to let them decide separately. That raises a broader question, on what level of unit should decisions be taken: 1) all languages, 2) specific language, 3) cultural language block, say all Latin-script European languages or all Germanic, Romance and Slavic languages as a supergroup regardless of script. I am new to Wikidata and do not know the conventional answer, and I do not know the subjectively best answer either. --Dan Polansky (talk) 04:28, 23 May 2023 (UTC)[reply]
  • Possibly relevant to this discussion is the existance of an enwiki hidden category which is triggered by a hyphen/en dash discrepancy in the date range separator used in a {{short description|[text]}} tag. There is clear guidance on enwiki that dates should be included in the short description:

    Dates or date ranges are encouraged when they enhance the short description as an annotation or improve disambiguation. As long as the formatting criteria are met, biographies of non-living people, articles on specific publications, and dated historical events generally benefit from dating, but since the description should be kept short, other information may need to take precedence.

    If the convention on English Wikipedia is also to use en dashes (rather than hyphens) as date separators; and a logical source for short description text is the corresponding wikidata item; then without agreement between wikidata and enwiki usage, there will be a persistent glitch on the enwiki side (if the wikidata description is used as an enwiki short description – which, again, seems to be a rational practice and a good use of the data we have here). -- Cl3phact0 (talk) 12:15, 6 August 2023 (UTC)[reply]
    Didn’t en.wp basically decide to get rid of Wikidata descriptions in mobile view? --Emu (talk) 12:56, 10 September 2023 (UTC)[reply]
    PS: See also: Wikidata:Project chat#Hyphen vs. En dash to separate years of birth/death? -- Cl3phact0 (talk) 13:40, 6 August 2023 (UTC)[reply]
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Requests_for_comment/Use_of_dates_in_the_descriptions_of_items_regarding_humans&oldid=1971712961"