Commons:Requests for comment/2022 overhaul of categories by period

An editor has requested comment from other editors for this discussion.
If you have an opinion regarding this issue, feel free to comment below.

Problem

Currently, titles of "categories by period" have two different orders of words:

period first: by month, year, and millennium, e.g. Category:January 2022 in France Category:2022 in France Category:3rd millennium in Europe
topic first: by decade and century, e.g. Category:France in the 2020s Category:France in the 21st century

The inconsistency often makes categorisation waste time unnecessarily, because when users often mix up the order, they cannot find the existing cats (for example, Category:France in 2022 is a redlink as I write), or they would create duplicate cats in the other order (e.g. Category:2020s in France).

I do not know how this dual system came about. I cannot trace the origin of each cat tree.

There is a consistent structure on some wikis, e.g. enwp has en:Category:3rd millennium in Europe en:Category:21st century in Europe en:Category:2020s in Europe en:Category:2022 in Europe en:Category:January 2022 events in Europe.

As far as I know, there is no new way to handle categorisation on Commons in the near to medium-term future. We will still be using this [[Category:Example]] "code in square brackets on file pages" for categorisation.

Therefore, here are my 3 proposals. Which do you think is best for Commons (in the near to medium-term future)?--RZuo (talk) 07:03, 5 December 2022 (UTC)[reply]

Proposal 1

Harmonise all cats to use period first, that is, Category:January 2022 in France Category:2022 in France Category:3rd millennium in Europe Category:2020s in France Category:21st century in France.--RZuo (talk) 07:03, 5 December 2022 (UTC)[reply]

Support.--RZuo (talk) 08:44, 14 January 2023 (UTC)[reply]

Preliminary proposal

For all cats if and only if they are intersections of a period (millennium, century, decade, year, month) and a location (continent, region, country, province/state, city/municipality, county):

subcats be named "<period> in <location>"
create redirects specifically for yearly cats: Category:<location> in YYYY to Category:YYYY in <location>

For other cats, no suggestion for change.--RZuo (talk) 08:44, 14 January 2023 (UTC)[reply]

Support - Jmabel ! talk 18:34, 14 January 2023 (UTC)[reply]

Support Now this is a more clearly worded proposal I can get behind. (No idea what "preliminary" means in the context - I'd oppose the intention of migrating the postliminary of all other cats later) Enyavar (talk) 22:36, 14 January 2023 (UTC)[reply]

Proposal 2

Harmonise all cats to use topic first, that is, Category:France in January 2022 Category:France in 2022 Category:Europe in the 3rd millennium Category:France in the 2020s Category:France in the 21st century.--RZuo (talk) 07:03, 5 December 2022 (UTC)[reply]

Oppose - Jmabel ! talk 15:47, 5 December 2022 (UTC)[reply]

this just came to my mind. let's say we consider a small place, Category:Teton Village, for example. to me, "21st century in Teton Village" sounds weird. "Teton Village in the 21st century" sounds better.

also there's the problem of different prepositions used before a place, in or at, whereas for all month/year/decade/century... it's all "in".

so i

Support this.--RZuo (talk) 21:00, 2 August 2024 (UTC)[reply]

Proposal 3

Keep current systems. Create redirects from one order to the other, e.g. let Category:2020s in France redirect to Category:France in the 2020s, for all levels of the cat trees (would be a bot's job).--RZuo (talk) 07:03, 5 December 2022 (UTC)[reply]

Support, best solution as it doesn't require any overhauls and it would make the alternative optional to use, as redirects make it more easy to be discovered. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 22:48, 9 December 2022 (UTC)[reply]

Discussion

Please leave long replies below and under the proposals above only vote (with short sentences). Also feel free to propose other solutions.--RZuo (talk) 07:03, 5 December 2022 (UTC)[reply]

recent initial discussion of this problem: Commons:Village_pump/Archive/2022/11#Are_you_annoyed_by_the_"by_year"_and_"by_decade"_cats?.--RZuo (talk) 07:36, 5 December 2022 (UTC)[reply]

w.r.t. "in the near to medium-term future", i imagine (not representing anybody but just my imagination), cats could be handled as a property of the file/file page. a bit like com:SDC. with that we wont be using this "wikitext on file page" method to do categorisation. that would also allow cat titles to be multilingual instead of monolingual, mostly english. in that case, order of the phrase could be defined flexibly, and can take on both orders at the same time, just like a single label and multiple aliases for a single item on wikidata.

but this imagination of mine is obviously not happening anytime soon.--RZuo (talk) 07:03, 5 December 2022 (UTC)[reply]

What about non-geographic categories by period, like in Category:Eiffel Tower by year, how would the year-subject order work there? --HyperGaruda (talk) 18:35, 6 December 2022 (UTC)[reply]

good question! now i remember a problem i had run into before. our current cat tree is Category:2022 in France (year) -> Category:January 2022 in France (month) -> smaller topic by month. somewhere down this tree, the order will have to flip, since i guess "2022 in Eiffel Tower" sounds unnatural. "2022 at Eiffel Tower" might be a bit better? but "Eiffel Tower in 2022" is still the best?

this problem has given rise to for example Category:2019 in HSBC Hong Kong headquarters building vs Category:HSBC HQ in 2020.--RZuo (talk) 23:21, 6 December 2022 (UTC)[reply]

if the topic is a person, i think the only way is Category:Tom Hanks in 2010. i cant imagine the year being in front. RZuo (talk) 23:32, 6 December 2022 (UTC)[reply]

How this structure comes to be; and why changing it will be hard

In general: I've done loads of categorizations in the past year, specializing on maps. When categories don't exist, I find orientation in pre-existing category structures, and that is how I imagine most categories were created, including Category:1856 in Paris as a sub-cat of Category:Paris in the 1850s. Yes, this is a bit maddening, but I found that almost all category trees were probably created like this. A LOT of this has apparently to do with how the templates were programmed, and in which way the authors there wanted the navigation bars to get accessed.
<Period-over-topic> makes a lot of sense in the historical periods: First, there are five old maps in "Old maps of London", and we don't need further categorization. Suddenly there are 500 "Old maps of London", and you need to divide them up by century, and "old" is replaced by "", and later by decade as the categories fill up. That's how we get to Category:1850s maps of London
<Topic-over-period> makes sense in other contexts, when the parent category was framed differently: We start with "Tour de France", and the first person to divide it up goes by "Tour de France, 1975" - emphasis on the event, not the year. Meanwhile however, other uploaders already filled "1975 in Paris". I imagine that in such cases, the person or group who made the first widespread template, wins the race, and defines the future structure.
<Topic-over-period-over-topic>: "Maps of the history of England" get logically subdivided into "Maps of 15th-century history of England" (with 15th century the period depicted, however we shortened this to Category:Maps of 15th-century England). Contrast that with "18th-century maps of the history of England": the 18th century means the date of map creation, a vastly different thing. Which means, that "19th-century maps of 2nd-century Roman Empire" (i.e. <period-over-topic-over-period-over-topic>) is a valid structure, and I'm positive that we could fill that thing pretty quickly some maps. The hardest part is not imagining the category, but going through with it and then enlarging the structure consistently: After all, there are also "18th-century maps of 1st-century Roman Empire" and "20th-century maps of 3rd-century Roman Empire", and "21st-century maps of 1st-century BC Roman Empire". We could do this, the media is available right here on Commons already, but I can only say this confidently because I have a basic overview over what is there.
And that is why I oppose proposals 1 and 2. These are non-starters, because you have no idea who is possibly going to object. I'm just a map-person, but there are also sport-persons, architecture-persons, biography-persons and fashion-persons, and I'm sure they could name you other reasons why one-over-the-other is not going to work for them. Which leaves Proposal 3. It sounds good at first superficially but there are obvious drawbacks: First, who creates all the redirects? We are going to need millions of those. Second, who maintains all the redirects, who changes and deletes them once they stopped being useful?
The obvious solution is (I think) structured metadata. With my maps, it would in theory be easy to create categorizations out of metadata. Practically, the structured-data input interface makes it extra hard to define {Spanish-language topographic map of Andalusia, base map by the topographic service of Spain, overlay in February 2011 by User:XYZ, displays archaelogical sites of the Roman era} or {Bilingual English-Japanese map of 9th-century history of China, made in 1934 by Wato Fujimuto, scan from the 1952 book "Title" by Perry Hillebrunk}. Instead of creating a dozen metadata claims, I use the imperfect category tree. Yes, that is sad and annoying to use.
*Sigh* Commons needs easy-to-use structured metadata, instead of even more hardcoded meta datastructures. --Enyavar (talk) 14:01, 10 December 2022 (UTC) [reply]

your heading seems to suggest you found the origin of these cat trees, but actually no...

this overhaul is expected to affect mostly the intersection of "period" and "place" cats. it's certainly futile to try harmonising more complicated cats for all countries. for example, under Category:Elections in 2014, there are all kinds of formats:

it's ok as long as they are consistent within themselves.

the cats that give rise to this discussion are not consistent by themselves. Category:3rd millennium in Europe Category:France in the 21st century, both need the article "the", but why is the order different??

as far as i can imagine, there's not much argument against having a consistent naming format other than "too much to change" "things have existed for years" (actually less than 20 for now).

benefits of a consistent structure are many. for users who work a lot on categorisation, who upload files often and have to find the correct cats to add, who design templates and modules... it's a relief to have a single structure. RZuo (talk) 18:30, 10 December 2022 (UTC)[reply]

My heading is that I suggest how the category tree came to be, so actually yes: I do know how a wiki works, and while I can't testify that things happened exactly the way I describe it (incremental uncoordinated growth, which then gets roughly organized by many conflicting users), this is how I think it happened. There is not a single "origin" of the cat trees.

There are attempts to standardize election categories, of course. At some point in the future, the elections by year will all be consistent within themselves and within the "Elections" category tree. I'm mostly active in the "Maps" category tree, but when I find uncategorized election maps, I'm sorting them correctly by "year elections in country".

I am NOT against harmonizing and I do see the benefits of a consistent structure. Which is why I am looking up the predominant, prevalent structures in categories I touch, and then I follow the lead. For example, we have vastly more categories about "maps of the world" than we have "world maps". So my new categories are "maps of the world", and I did already move some "world map"-categories into the other scheme.

Yes, I am harmonizing on the micro-level (bottom-up), while your suggestion concerns the overarching macro-level (top-down). But doing a macro-level change always needs to consider the micro-level consequences. That is what I was generally missing in your proposal, and that is why I was giving some feedback about problems with the top-down approach. --Enyavar (talk) 15:31, 28 December 2022 (UTC)[reply]

yes? so what're the first categories created in each format? give the answers or get lost.--RZuo (talk) 08:21, 14 January 2023 (UTC)[reply]

So both the "year by country" and "country by decade" trees were largely created in 2008 and 2009 by a pretty diverse group of editors. Infrogmation and AnRo are really standing out there in doing a lot of work it seems, for both formats at the same time and for a long time since they started (though they didn't create the first templates either), and both editors are still active today. Ask them, or any other who established category structures? As far as I can see, they didn't coordinate via the Village pump, but it hardly matters in my opinion: The seniority of one "system" or template doesn't invalidate the other. --Enyavar (talk) 00:07, 15 January 2023 (UTC)[reply]

Months as numbers rather than English names?

If major effort is going to be put into reworking chronology categories, can we get rid of forcing everyone to use English language month names when numbers would do the same job more clearly and in less space (eg rather than "January 2022 in France", "2022-01 in France" or similar). Thoughts? -- Infrogmation of New Orleans (talk) 16:36, 21 December 2022 (UTC)[reply]

if it were up to me to decide, i would prefer an all number notation following yyyy-mm-dd.

moreover, i think a name format of "period - topic" would resolve most problems we have at hand. for example:

2nd millennium - France
20th century - France
2020s - France
2020 - France
2020-02 - France
2020-02-29 - France
2020-02 - Eiffel Tower
2020-02 - Tom Hanks
2011-05 - Barack Obama - France = Category:Barack Obama in France, May 2011

this way we could (1) avoid the problem about what should come first (2020 in France vs Eiffel Tower in 2020) (2) avoid the problem with preposition (2020 in France vs 2020 at Eiffel Tower) (3) there's no need to use sortkeys "01" "02"... for monthly cats. all cats will sort nicely by themselves in numerical order.

but i guess this approach is probably not gonna be welcomed by most users. XD

there'd be a problem for day cats like Category:12 February, though. many people might interpret 02-12 as 2nd of Dec, Dec 2002, Feb 2012...--RZuo (talk) 07:41, 28 December 2022 (UTC)[reply]

I'm coming from an incrementalist position here, i.e. favoring gradual changes: When we're talking individual dates, the main categories are already formed in ISO 8601 (Category:2020-02-22). This should be implemented consequently, first (just look at Category:Russo-Ukrainian War: March 2, 2022, smh). In a second step, we could then discuss using ISO 8601 on the yyyy-mm level like Infrogmation suggests. The mm-dd level is indeed highly confusing, I would implement that as the last thing.

Also, hyphenized categories? Hate them already just by thinking about it. Is it "Category: Architecture - France" or "Category: France - Architecture"? Multi-topic categories really need the qualifiers, just check out the related category trees, you'd need to re-sort EVERYthing: "Category:2020 - Books - United States" vs "Category:Books - Biology - United States"? It boils down in how to logically prevent people from wrongly putting files into "Category:United States - Books" when these files belong into "Category:Books - United States" ... yeah no thanks. --Enyavar (talk) 15:03, 28 December 2022 (UTC)[reply]

that's not a real proposal if this user cannot understand. (bolded a sentence to highlight it for this user.)--RZuo (talk) 08:21, 14 January 2023 (UTC)[reply]

Examples of absurdity

Category:Military in 2022 vs Category:2022 in politics. topic first or year first?!!! this is so dysfunctional.--RZuo (talk) 19:52, 28 October 2023 (UTC)[reply]