Commons:Requests for comment/Categories of photographs by country by date
An editor has requested comment from other editors for this discussion. If you have an opinion regarding this issue, feel free to comment below. |
Background
[edit]We have a bunch of categories of the type Category:Germany photographs taken on 2024-05-04. They all have the same format, they are hidden, and they are actually meant to be used by templates such as {{Taken on}}.
However, these categories often get diffused -- subcategories to the level of administrative divisions such as Category:Saxony photographs taken on 2024-05-04 or even cities such as Category:Dresden photographs taken on 2024-05-04 get created, and the files from the main category get moved there (and thus become not readable by the templates). This is being done by a number of users, and has been discussed at various village pumps. The latest discussions can be found at Commons:Village pump#Category diffusion, again and Commons:Administrators' noticeboard/Blocks and protections#Sahaib, battleground mentality, and edit warring, with the suggestion that an RfC should be opened.
Note that some users apparently confuse the above categories with the categories such as Category:May 2024 in Germany, Category:May 2024 in Saxony, and Category:May 2024 in Dresden. These are real unhidden categories and may be diffused down to villages if needed. I have never seen them diffused down to a single day, but can not exclude that this is happening as well, we do not have policies prohibiting such diffusion. The category tree can be sometimes confusing though with hidden categories being subcategories of non-hidden categories. This category group is being discussed since 2022 at Commons:Requests for comment/2022 overhaul of categories by period, and that discussion will certainly benefit from having more opinions, but it is out of scope for this RfC.--Ymblanter (talk) 10:38, 10 May 2024 (UTC)
- take a look at Commons:Village pump/Archive/2023/07#Category:2020 photographs of Hannover. i dugged up the history of these cats last year. RZuo (talk) 11:19, 10 May 2024 (UTC)
- Thanks, the link contains indeed more discussions of the issue.--Ymblanter (talk) 11:32, 10 May 2024 (UTC)
- Just another case happening right now. Note that categories of the type "Month in Foo" are being removed; these will be very difficult to restore even if the photos get moved up the category tree by a bot.--Ymblanter (talk) 12:36, 12 May 2024 (UTC)
- Thanks, the link contains indeed more discussions of the issue.--Ymblanter (talk) 11:32, 10 May 2024 (UTC)
RfC questions
[edit]0. Are categories of photographs by country by date necessary?
[edit]The current understanding is that since the categories by country are use by templates/bots, they are necessary, and every photograph (if applicable) must be in one of these categories (even if it is also included in the subcategories). If you oppose this understanding, please discuss in this section.--Ymblanter (talk) 12:39, 12 May 2024 (UTC)
- "Must" is an overstatement. Often we do not know the exact date of a photograph. For older photographs, we may not even be completely confident of the decade. - Jmabel ! talk 17:31, 14 May 2024 (UTC)
- Further note: these "mechanized" categories (usually maintained through {{Taken on}}) should all be hidden categories. - Jmabel ! talk 17:37, 14 May 2024 (UTC)
- Yes, I think they are useful. One way I use them is to see whether there are other photos of the same subject made that day and either harmonize their categories or create a new category. --JopkeB (talk) 03:44, 19 May 2024 (UTC)
- We need a category by date, but there shall not be a categorization by country and date. Reason: that would be 7300 categories per year on the country level (nearly; there are only 195 contries). The US have 50 states. Would give us approx. 18.600 categorie for the federal state level in the US only each year. There might by countries with lesser provinces and states like Canada or Australia, perhaps the average bis about 15, would result in 365 x 200 x 15 each year on secons level entities, roundabout 1.1 million categories for this level each year. France has about 100 departments, Germany more than 300 landkreises, the U.S. more than 3000 countys, would result in 1.24 million categories for these three countries only. And we still did not reach the municipality level. Germany has about 11.000 municipalities. Gives us 4 million categories on the municipality level in Germany only. Do we need categories by date? Yes. Do we need categories by date and area? Hell, nooo! --Matthiasb (talk) 04:31, 21 May 2024 (UTC)
- What is exactly the problem? Why are those numbers a problem?
- In my experience we only create a category by country (or a subdivision or any other criterium) and date if there are files for it. And for some countries I am already glad that there is a category structure by year for that country (without the subdivisions at all). The alternative is overcrowded categories by country by year, making them unusable. JopkeB (talk) 07:05, 21 May 2024 (UTC)
- What is exactly the problem? Why are those numbers a problem?
- I'd say for countries and maybe the next level down from that by date categories are necessary but it gets super obtuse outside of that. Especially if it has to do with another subject besides administrative boundaries. For instance a lot of the subcategories in Category:Postmarks by day are totally pointless and just make it impossible to sort images of postmarks any other way. Like the subcategories in Category:8 June postmarks. It's totally pointless to organize things down to that small of a level. --Adamant1 (talk) 07:15, 21 May 2024 (UTC)
- This discussion is only about photographs by country by date, not about other subjects (so it should be about photographs). If you encounter problems with postmarks, please start a seperate discussion about that subject. JopkeB (talk) 15:05, 21 May 2024 (UTC)
- Matthiasb: I think there is a misunderstanding: If there are no "2024-05-04 photographs from Landkreis Buxtehude", that category is entirely unnecessary and will not even be created. If there are a bunch of photos from that place and that date, the category can be potentially useful, especially if some category by-month already has thousands of images (like Berlin, Paris or London: I expect that no single day-category of these cities is going to remain empty). But please see also my opinion on Question 1. --Enyavar (talk) 15:21, 21 May 2024 (UTC)
- This discussion is only about photographs by country by date, not about other subjects (so it should be about photographs). If you encounter problems with postmarks, please start a seperate discussion about that subject. JopkeB (talk) 15:05, 21 May 2024 (UTC)
- I'd say for countries and maybe the next level down from that by date categories are necessary but it gets super obtuse outside of that. Especially if it has to do with another subject besides administrative boundaries. For instance a lot of the subcategories in Category:Postmarks by day are totally pointless and just make it impossible to sort images of postmarks any other way. Like the subcategories in Category:8 June postmarks. It's totally pointless to organize things down to that small of a level. --Adamant1 (talk) 07:15, 21 May 2024 (UTC)
- @Matthiasb: I don't quite understand -- if the high potential number of categories for "subdivision by date" or "municipality by date" is a problem (which I agree with), why should "country by date", with a very much lower number, necessarily be a problem too? Simply we should establish "Stop at country level" as policy, that's it... --A.Savin 10:11, 21 May 2024 (UTC)
- Oppose Again: not for every subdivision and municipality categories by date will be made, only for the dates we need, and that usually might be a few by year per municipality (my experience: at the most), except for very large and popular cities like New York. The parent categories can be "Photographs taken in Land X on yyyy-mm-dd" and "YYYY in Municipality Y", you do not need categories like "subdivision by date" or "municipality by date", except for perhaps very large cities and/or countries. And so be it, why bother about it? JopkeB (talk) 15:18, 21 May 2024 (UTC)
- @A.Savin: If a "Stop at country level" policy would work, that will be fine for me but does it work? Would not a "november 31, 2006 anti government protests in Berlin provoke the creation of Category:2006-11-31 in Berlin? I guess people tend to give such categories dates beause of there is an anti government protest every week, maybe and I see such a part of tree to eolve from the bottom, not from the top down as usual.
- Therefor in the German WP we do have a strict anti place-time-combination-category policy. I agree that at commons there is a different reasoning wether a category is useful or not but guess we let the genie out of the bottle and nobody will be capable to control what happens. For example because of people begin to agree, perhaps me included, that 2024 in Germany might not be adequate to host all 2024-05-22 in German files. Dunno. Matthiasb (talk) 22:04, 26 May 2024 (UTC)
- I don't think there has been any objection to the topical categories that get down to a rather local place and a month. It's the hidden categories for a specific day that we are discussing. Note that no topical category for an event should ever be a subcat of "YYYY-MM-DD PLACE photographs", whether PLACE is a country or something more local, because a topical category is not inherently limited to photographs. - Jmabel ! talk 14:42, 27 May 2024 (UTC)
- @Matthiasb: I don't quite understand -- if the high potential number of categories for "subdivision by date" or "municipality by date" is a problem (which I agree with), why should "country by date", with a very much lower number, necessarily be a problem too? Simply we should establish "Stop at country level" as policy, that's it... --A.Savin 10:11, 21 May 2024 (UTC)
- I don't oppose all of the understanding of the question fundamentally, but I think that in my opinion it's also okay if an image is in a sub-category. Duplicate-catting on the 0-level is then not a "must". I'd like to further say that most contemporary images are not even categorized by date (yet) - if we restrict the possibilities of sub-categories now, we are forcing unsorted hidden categories with hundred thousand photographs into existence; and those hidden categories are NOT going to be useful, if we can't also categorize by location. --Enyavar (talk) 15:21, 21 May 2024 (UTC)
- No. Use structured data queries. Why do we have structured data if we're not going to use it. Category:Black and white photos taken in Turkey in 1932 depicting an adult female human in a sitting pose wearing no socks needs to go! It's time to stop treating categories like a query language and start using an actual query language. It's not only an inefficient solution, but it actually makes files harder to find, for example, if you just want to see all the photos of people wearing no socks (and I know you do). Nosferattus (talk) 04:11, 30 May 2024 (UTC)
- i think there's only 1 argument for diffusion to "date x city" level -- overcrowding of the parent cat. diffusion makes it easier to navigate, especially for areas that attract a lot of media, e.g. nyc/london/berlin.
- but even such cats can sometimes contain more than 1000 files. if that argument applies, why not diffuse them further to "nyc photos taken on 2011-11-11 AM/morning..."?
- diffusion has to stop at a certain level, even if some categories will still contain >200 files.
- if we look at each diffusion, from century to decade to year to month, the distribution is always 10x (year->month 12x), but month->date is 30x, so this diffusion is much "stronger". diffusion to month level means for a place for a given year files are contained in 12 cats, but diffusion to date level will be 365 cats.
- i think that's the reason why diffusion to date level is needlessly granular and undesirable. RZuo (talk) 09:12, 30 May 2024 (UTC)
1. Should subcategories of the categories of photographs by country by date be created
[edit]In other words, do we need categories such as Category:Saxony photographs taken on 2024-05-04?--Ymblanter (talk) 10:38, 10 May 2024 (UTC)
Note: I request the closer, in case the close is "no consensus", also indicate what is the status quo situation (basically, what should and what should not be allowed in the future).--Ymblanter (talk) 10:41, 10 May 2024 (UTC)
- If it were up to me, I'd say "no", but if we do go that way I'd like to see it be non-diffusing below the country level: add another parameter to {{Taken on}} and maybe to {{Taken in}} so we can support both the country level and something more specific. These are hidden maintenance categories, so OVERCAT doesn't necessarily apply. - Jmabel ! talk 17:41, 14 May 2024 (UTC)
- Thinking about this for a while, the suggestion of Jmabel seems very good: "Taken on (date)" and "Taken in (city/province/country)" sound like very good parameters for automated reasoning, as long as Commons or WikiData provide the ontological models of "location X is part of region Z in country Y on continent W" and "day W in month X is part of year Z in decade Y". IF our servers can take it, users could use then adjust time-level and location-level up and down to search for "2020s photos of Liechtenstein" or "2024-05-05 photos of Brooklyn".
Failing to implement that idea, we may not need subcategories about country subsubdivision, but I don't think we should move against them, as long as they adhere to the overall structure of the tree and are not empty. If really needed, all images of the lower subcats can be also included in the country-level subcat. (But I admit just seven files are a really low number.) --Enyavar (talk) 19:14, 17 May 2024 (UTC) - Yes, I think we need them, even if only to relieve overcrowded categories and find easier what you are looking for, just as we do with other subjects. --JopkeB (talk) 03:57, 19 May 2024 (UTC)
- No, it should be strictly prohibted, see my reasonatingin 0. --Matthiasb (talk) 04:32, 21 May 2024 (UTC)
- Yes, though not just as an arbitrary parsing of so-called overcrowded categories, but in cases where there is a clear delineation and sufficient content to make delineation meaningful. Josh (talk) 08:15, 24 May 2024 (UTC)
- Absolutely not! Kill with fire and salt the earth! Nosferattus (talk) 04:13, 30 May 2024 (UTC)
- I am against any restrictions. Cats can be hidden, I am okay with that, but I think structuring photographs by date is rather useful --PantheraLeo1359531 😺 (talk) 16:10, 9 June 2024 (UTC)
2. To which level is the diffusion allowed?
[edit]If the answer to question 1 is "no", this question becomes moot. If you strongly feel the subcategories should not be created, you do not need to reply here. However, if you think they should be created, or you are on the fence, indicate here what level down they can go. First level administrative divisions? Any localities above certain population level (say 1M)? Any level assuming the number of photographs is appropriate?--Ymblanter (talk) 10:41, 10 May 2024 (UTC)
- A cutoff on the 1-million population level means that even the "United Nation of Malta + Iceland" wouldn't deserve a national category. So no, I think all levels can be justified, as long as there are enough photos created for that specific day. Somewhere between 50-100 images is a good minimum number, I think? Even a smaller town can occasionally qualify. --Enyavar (talk) 19:32, 17 May 2024 (UTC)
- Though in 0. I have argued against each sorting by date and area I think we cannot prevent this at all. Pitifully we have categories like 2024-05-21 anti-goverment protests in Mannheim and these categories will have to go somewhere... --Matthiasb (talk) 04:40, 21 May 2024 (UTC)
- Categories like "Russia photographs taken on 2024-05-21" are important because they de-load the core categories like "May 2024 in Russia" and "2024 in Russia". However, for such a large country as Russia, it would be bad to mix photos taken in different regions in "Russia photographs taken on 2024-05-21". This should not be done for at least 2 reasons. 1. For example, it will be much more interesting for residents of the Moscow region to see photos taken in Moscow region on this date but not in all of Russia. For me, "Moscow photos taken on 2024-05-21" are much more interesting than, for example, "St. Petersburg photos taken on 2024-05-21". The one from St. Petersburg will probably say the opposite. 2. Can you imagine how many photos will be collected in this category, for example, in 100 years? Can you imagine how many photos will be taken in Russia in 21 May 2024? Copyright is not eternal and probably a lot of these photos will be uploaded to Commons, early or late. Thus, category for example, "Russia photographs taken on 2024-05-21" will ultimately cointain hundreds or more likely, thousands photographs. The same is true for other countries, of which we have vastly more photos (and will be even vastly more) on Commons: US, UK, Germany, Spain, Italy, France etc and for large cities like Moscow, St. Petersburg, Kyiv, New York, Berlin, Paris, London, Beijing. I'm not 100% sure for smaller cities like Dresden, Hannover and Cologne. However, Commons has so vast number of German photos, so such diffusion may be acceptable as an exception. Countries that should not be expected to have a large number of photos on Commons like, for example, Suriname, Mali, Zambia, Albania etc. categorization like "Mali photographs taken on 2024-05-21" is enought imho. Юрий Д.К 11:33, 21 May 2024 (UTC)
- Largely Agree, very good points. But even for Paramaribo (capital of Suriname) there might be one or more days that many photos will be taken and uploaded on Commons, for which a "Category:Paramaribo photographs taken on yyyy-mm-dd" might be needed. So I would not exclude any country or subdivision. Just trust your fellow editors that they only create categories when they are needed and otherwise you can always nominate redundant categories for deletion. What we can do, is set a minimum number of files in a subdivision by date category.
- For countries: at least one.
- For subdivisions/provinces/municipalities/populated places/and so on: at least ten (or any other number we agree on).
- JopkeB (talk) 15:44, 21 May 2024 (UTC)
- Largely Agree, very good points. But even for Paramaribo (capital of Suriname) there might be one or more days that many photos will be taken and uploaded on Commons, for which a "Category:Paramaribo photographs taken on yyyy-mm-dd" might be needed. So I would not exclude any country or subdivision. Just trust your fellow editors that they only create categories when they are needed and otherwise you can always nominate redundant categories for deletion. What we can do, is set a minimum number of files in a subdivision by date category.
- They can go down as far as it makes sense based on two things. If you are creating a sub-cat, is the scope of that sub-cat sufficiently distinct both from the parent category and parallel sub-cats to be meaningful? And, is there content present to support the new category (i.e. something to put there)? If these are both true, then the new sub-cat may be warranted. I don't think any particular number of files needs to be set to permit or bar such categories, as it depends on more than the category itself. For example, if you have a particular country on a date and make a sub cat for a city in that country for which there is only one image, that may not really be useful and probably isn't the best plan. On the other hand, if the country already has a network of city sub-cats by date, and some of them maybe have only a single file, that is probably fine. Josh (talk) 08:24, 24 May 2024 (UTC)
- Comment what concerns me here -- and I've seen it happen over and over -- is that once we say something is allowed at all, the "splitters" win every time. So what I'm afraid we'll see is (for example) someone goes to some relatively small town like Merrick, New York, takes a dozen photos, and puts them in a special category for a day and place, making it far less likely that they get well categorized in any other way. (There are other similar issues, but I don't want to go on at length.) - Jmabel ! talk 20:04, 24 May 2024 (UTC)
- Comment We also have to differ between cities and districts --PantheraLeo1359531 😺 (talk) 16:50, 21 June 2024 (UTC)
- @PantheraLeo1359531: What do you mean? In what way should we differ? JopkeB (talk) 04:58, 22 June 2024 (UTC)
- When we have New York, we can treat this as a city or a state. The United States has 50 states, so the diffusion would be far weaker than diffusion to cities. Germany has several districts; some of them are cities at the same time. When we have little countries like Liechtenstein or Vatican City, we would have a diffusion to country-level, but there are many areas in the world that are larger (in manners of area and population), but have a lower level than a country. Hmm. I might be a nerd when it comes to structuring, but date-cats should be hidden at least. Greetings --PantheraLeo1359531 😺 (talk) 09:10, 22 June 2024 (UTC)
- I'm guessing you meant differentiate (make a distinction), not differ (disagree). - Jmabel ! talk 16:20, 22 June 2024 (UTC)
- Yes, at least I did (I just copied it), sorry.
- Thanks, PantheraLeo1359531, for the clarification. That is exactly why I propose to create categories only when there are enough files for, no matter whether it is on country level, city, state or region. JopkeB (talk) 16:29, 22 June 2024 (UTC)
- Thanks, Jmabel :D, yes you're right. As not-native speaker, I misuse some words sometimes :). And I agree with JopkeB. One not so large city can hold very many pictures (because there are many cultural heritage monuments whatsoever) with subcats, while larger areas do not need that much subcats (by date etc.) --PantheraLeo1359531 😺 (talk) 16:39, 22 June 2024 (UTC)
- I'm guessing you meant differentiate (make a distinction), not differ (disagree). - Jmabel ! talk 16:20, 22 June 2024 (UTC)
- When we have New York, we can treat this as a city or a state. The United States has 50 states, so the diffusion would be far weaker than diffusion to cities. Germany has several districts; some of them are cities at the same time. When we have little countries like Liechtenstein or Vatican City, we would have a diffusion to country-level, but there are many areas in the world that are larger (in manners of area and population), but have a lower level than a country. Hmm. I might be a nerd when it comes to structuring, but date-cats should be hidden at least. Greetings --PantheraLeo1359531 😺 (talk) 09:10, 22 June 2024 (UTC)
- @PantheraLeo1359531: What do you mean? In what way should we differ? JopkeB (talk) 04:58, 22 June 2024 (UTC)
3. What is a "country" in this context
[edit]Withdrawn, see the discussion below--Ymblanter (talk) 12:40, 12 May 2024 (UTC)
- The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Is a country just a member of Category:Countries of Europe by name or similar (meaning for example that Scotland is not a country and Gibraltar is a country, Åland is a country and Abkhazia is not)? Or are there other suggestions?--Ymblanter (talk) 10:41, 10 May 2024 (UTC)
- The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
4. Do we need categories by date but not by location
[edit]See for example Category:Railway photographs taken on 2006-01-25, currently there are more than 7,800 similar subcats. The added value for users in having all kinds of railway photos sorted by date is not obvious at all. --A.Savin 15:33, 10 May 2024 (UTC)
- These are mainly useful for people who curate content here, and I for one do find them useful. Quite often, if you are having trouble making sense of where/what a given photo is, these are helpful for finding another photo in the same photo shoot. But definitely these should be "hidden categories": of interest to contributors, not to general users. - Jmabel ! talk 17:30, 14 May 2024 (UTC)
- I agree that these can be useful categories, and likewise that Jmabel is correct, they should be administrative categories (i.e. 'hidden'). Contents here should not be removed from topical (i.e. non-hidden) categories to be placed in these. Josh (talk) 08:06, 24 May 2024 (UTC)
- No. Used structured data queries. Nosferattus (talk) 04:18, 30 May 2024 (UTC)
Preliminary discussions
[edit]- A special case may also be Category:Railway photographs by date + its numerous subcats, for which I think we also need clarification if it's meaningful and desirable as is (IMHO not). --A.Savin 11:45, 10 May 2024 (UTC)
- Good idea, I will add a question (#4) later, or feel free to do it yourself.--Ymblanter (talk) 15:12, 10 May 2024 (UTC)
- Why are we attempting to make search of dates redundant in favour of categories which only serve to hide files from view and easy thumbnail comparison? Broichmore (talk) 12:12, 10 May 2024 (UTC)
- @Broichmore: Question What are you referring to when you say 'hide files'? Josh (talk) 15:47, 10 May 2024 (UTC)
- Simply we may have 10 photographs (that's optimistic) for a particular week, devolved into 7 daily categories; most, with just 1 file in each. The photos are effectively hidden. There are few notable dates, for filing at this micro level. Example 9-11, death of JFK, man lands on the moon. Broichmore (talk) 18:25, 10 May 2024 (UTC)
- @Broichmore: Question What are you referring to when you say 'hide files'? Josh (talk) 15:47, 10 May 2024 (UTC)
- if there isnt a rule, common sense is that a category is subdivided only if there are too many files (>400?).
- so for example San Marino should only be diffused to Category:2017 in San Marino maximally (country x year), whereas a city like NYC can be diffused to Category:New York City photographs taken on 2024-03-07 (city x date).
- so i think when we consider a "rule", we should think about practicality of commons maintenance.
- the quantity of files related to a place is primarily affected by these factors: population, local participation of commons/wikipedia, cultural/political significance.
- for example, Dortmund had 587k residents and Hannover 534k in 2020, but Category:2019 in Dortmund has 467 elements under it and Category:2019 in Hannover has 1973, probably because Hannover is the capital of a state.
- Category:2019 in Fukuoka prefecture has 1516, even though it has almost 10 times the population of Hannover, probably because there are way fewer active users in japan than in germany.
- so if we come up with a rule with those places that have a lot of files in mind, then when this rule is applied to places like Fukuoka or less developed areas in africa/asia, categories will be overly subdivided. but if we come up with a rule with these places in mind, then nyc/london/berlin categories will be bloated.
- imo, it's not a problem if the 2300+ files under Category:January 2023 in New York City are put there and not diffused. (i support the lowest level to be "city x month".)--RZuo (talk) 14:02, 10 May 2024 (UTC)
- Day categories are all supposed to hidden categories intended for internal use. I believe we should not push those below the country level (and I live in a very large country, the U.S.).
- Month categories can be pushed down as far as we want. I'd have no objection to (for example) a particular individual month in a particular public market, as long as it's the sort of place where we normally have a dozen or so photos taken each month.
- - Jmabel ! talk 15:00, 10 May 2024 (UTC)
- I tried to make this distinction in the background section, but feel free to edit. Ymblanter (talk) 15:13, 10 May 2024 (UTC)
- i fail to see why there's this distinction between day cats and month cats, and why day cats should only be diffused to country level if month cats can be diffused to town level.
- my preference is to delete all "Category:xx photographs taken on yyyy-mm-dd". (roughly 140k+, https://commons.wikimedia.org/w/index.php?search=intitle%3A%22photographs+taken+on%22+intitle%3A%2F%5B0-9%5D%7B4%7D%5C-%5B0-9%5D%7B2%7D%5C-%5B0-9%5D%7B2%7D%2F&ns14=1 ) or, at the very least, prevent future creation of such categories and leave the existing ones as a legacy to be cleaned up slowly.
- as @Achim55 said in Commons:Categories for discussion/2016/10/Category:September 2007 Finland photographs, "When hopefully one day the category intersection mechanism will be implemented into the wiki software hundreds of thousands of these categories will be thrown away."
- as @Auntof6 said in Commons:Categories for discussion/2016/12/Category:Switzerland by day, "The subcategories here are for individual days, which is more detail than needed. Many of the subcategories have only a few entries, some as few as one. I think categories by month and year would be enough to cover what there is here." RZuo (talk) 10:53, 21 May 2024 (UTC)
- If there are already 140k+ categories like "Catgory:xx photographs taken on yyyy-mm-dd" there isn't anything to regulate anymore. Actually we're through. (I did not know that it is like that.) Matthiasb (talk) 09:26, 27 May 2024 (UTC)
- For my part, I strongly Oppose any rule depending on how crowded a category currently is. Either "by date" categories are relevant for cities, or not. If yes, nothing is speaking against such a category even if only one file fits so far. If no, they should not be created at all, not even for 1,000 files. --A.Savin 15:06, 10 May 2024 (UTC)
- Limiting the level of sub-categorization based on an arbitrary line, whether it be by level (e.g. down to country, city, etc.) or by quantity of files, is going to cause problems. RZuo was right to point out the disparity it would cause. One line may make a lot of sense for a highly-populated (# of images) place, but if enforced on a sparsely populated place, it would not be appropriate, and the same in reverse. Thus any new rule to try and draw such a line would, as well-meaning as it may be, need to pass a high bar of necessity to warrant the perhaps-unintended consequences. As for the specific questions above:
- Sub-categories of 'country photographed on date' categories are completely fine, so long as they are compliant with Commons category policies . If the files in the parent category are sufficiently distinct to be sortable and there are sufficient quantity to warrant the effort, than specific subs can be created. Those subs should match the hierarchy of the main topic (country) so far as it is applicable to the contents.
- The diffusion is allowed as deep as the above applies for the given topic. Once a level is reached where no meaningful distinction can be made amongst the files, that is the limit to sub-categorization.
- What is a 'country' in this context should be identical to what is a 'country' everywhere else in the Commons category scheme, i.e. what is within the scope of Category:Countries. That is a matter of its own discussion, but this tree should not adopt a unique definition of 'country'. This is important to comply with the Universality Principle .
- If categorizing files with a 'country' topic is reasonable, I see no reason why it wouldn't be as reasonable for other topics, providing there are enough files to warrant it. Personally, my energy for sorting by time usually runs out at the year level, and the topics I usually work on don't often need more specificity than that, but I wouldn't artificially preclude topics other than 'country' for sorting by date.
- The primary problem cited by Ymblanter appears to be that as these sub-category trees become more developed, they attract movement of files to them from the topical categories, resulting in 'hidden' files. It seems that as something such as 'Nairobi photographed in X' is created, files under Nairobi end up getting moved to this new sub-cat. It is a non-topical category (thus hidden) and thus those files are now no longer in the topical category where they belong.
- I don't think artificial limits on sub-categorization of the 'by date' categories is going to stop this problem. It seems a bit like fighting shoplifting by not opening new stores. The real problem is a combination of reminding users to copy (not move) files to non-topical categories, and the basic consideration one should always have when doing any diffusing: to consider the ability of folks to find that file if they don't happen to be looking for the particular distinction you are diffusing it by. Josh (talk) 16:27, 10 May 2024 (UTC)
- Our main purpose is not diffusion of overcrowded categories at any cost, but usability for internal and external users in general. If we have an overcrowded parental category for something, that's annoying but surely not the reason to make it absurd. Without limitation, we are going to end up someday with categories like "Konrad-Wolf-Straße (Berlin-Alt-Hohenschönhausen) photographs taken on 2035-12-16" or "Human penis photographs taken on 2009-03-31", because the experience shows that there are always people around who have nothing better to do rather than sth. like this, with this level of absurdity that's surely not *my* Commons, sorry. Regards --A.Savin 17:27, 10 May 2024 (UTC)
- Thanks for the „kind“ note. Pardon me, but I have enough other things to do here. Now to the point: Once again, a discussion that was opened quite late. Cities like Dresden or London have long been more or less strictly structured according to individual shooting days. Should everything be reset now? I'm not really ready to discuss this again as long as categories like „Women wearing sunglasses in the City of Westminster“ or „Insects facing left“ are maintained on Wikimedia Commons. But anyway: The combination of location and shooting date has a great advantage: it saves additional subcategories and streamlines the number of categories per photo. For example, a picture from Herford taken on August 21, 2012 would be categorized as „2012 in Herford“, „August 2012 in North Rhine-Westphalia“ and „Germany photographs taken on 2012-08-21“. In my opinion, the shorter form „Herford photographs taken on 2012-08-21“ is much more elegant. What’s wrong with that? Who is afraid of such categories? Furthermore, we all know how things work here at Wikimedia Commons: First, a category "Germany photographs taken on .." is created, then all relevant photo material is put into it. Eventually, it is realized that there is a jumble of thousands of photos from Heidelberg to Berlin under this date, which are simply unmanageable in bulk. Then, a wise Wikimedian comes up with the idea to create a new subcategory. And so the cycle continues. I have no interest in having to go back and re-categorize all my photos every time. That's why I categorize from the bottom up and not from the top down as a principle. This saves me from constantly having to rework things. No question: Technical solutions can be introduced to make these processes easier. However, since it is no longer possible here to automatically extract geodata from metadata using bots, I don't have much hope. Unfortunately, a lot of manual adjustments are still necessary here. --J.-H. Janßen (talk) 19:54, 10 May 2024 (UTC)
- Well, you recategorize MY photos on a regular basis and clog my watchlist making it unusable. Apparently, you do not have any issues with this. I understand your unwillingness to "discuss it again", but it not that last time it was consensus that your actions are ok, quite the opposite. Ymblanter (talk) 20:00, 10 May 2024 (UTC)
- And this "jumble of thousands of photos" also applies for Königswinter, Insel Poel and Melle (Lower Saxony), to name a few? Yeah, just wonder who would believe that... --A.Savin 21:56, 10 May 2024 (UTC)
- Thanks for the „kind“ note. Pardon me, but I have enough other things to do here. Now to the point: Once again, a discussion that was opened quite late. Cities like Dresden or London have long been more or less strictly structured according to individual shooting days. Should everything be reset now? I'm not really ready to discuss this again as long as categories like „Women wearing sunglasses in the City of Westminster“ or „Insects facing left“ are maintained on Wikimedia Commons. But anyway: The combination of location and shooting date has a great advantage: it saves additional subcategories and streamlines the number of categories per photo. For example, a picture from Herford taken on August 21, 2012 would be categorized as „2012 in Herford“, „August 2012 in North Rhine-Westphalia“ and „Germany photographs taken on 2012-08-21“. In my opinion, the shorter form „Herford photographs taken on 2012-08-21“ is much more elegant. What’s wrong with that? Who is afraid of such categories? Furthermore, we all know how things work here at Wikimedia Commons: First, a category "Germany photographs taken on .." is created, then all relevant photo material is put into it. Eventually, it is realized that there is a jumble of thousands of photos from Heidelberg to Berlin under this date, which are simply unmanageable in bulk. Then, a wise Wikimedian comes up with the idea to create a new subcategory. And so the cycle continues. I have no interest in having to go back and re-categorize all my photos every time. That's why I categorize from the bottom up and not from the top down as a principle. This saves me from constantly having to rework things. No question: Technical solutions can be introduced to make these processes easier. However, since it is no longer possible here to automatically extract geodata from metadata using bots, I don't have much hope. Unfortunately, a lot of manual adjustments are still necessary here. --J.-H. Janßen (talk) 19:54, 10 May 2024 (UTC)
- Our main purpose is not diffusion of overcrowded categories at any cost, but usability for internal and external users in general. If we have an overcrowded parental category for something, that's annoying but surely not the reason to make it absurd. Without limitation, we are going to end up someday with categories like "Konrad-Wolf-Straße (Berlin-Alt-Hohenschönhausen) photographs taken on 2035-12-16" or "Human penis photographs taken on 2009-03-31", because the experience shows that there are always people around who have nothing better to do rather than sth. like this, with this level of absurdity that's surely not *my* Commons, sorry. Regards --A.Savin 17:27, 10 May 2024 (UTC)
- I am now thinking about removal of question 3, I agree it is redundant. I was afraid of users claiming that England, Scotland, and Wals are countries, but I guess this can be dealt with. I will add the clarification to question 1 instead.--Ymblanter (talk) 19:50, 11 May 2024 (UTC)
- My "rules" would be:
- Create a date category when you need it, just as you do for other types of categories, that is when you would like to keep a bunch of photos together or when the parent is overcrowded or would be overcrowded when you'd add a lot of photos about one geographic area. Exception: on the country level date categories may be created even if there is only one photograph, because it is usually expected that later on more photographs will be added for that day.
- "Overcrowded" is usually defined as "over 200 files" (see for instance Commons:Categories for discussion/2023/06/Category:Categories requiring temporary diffusion, where is stated that "there is already broad agreement on 200 as a baseline").
- No limits to the diffusion to a specific administrative level; any level that you need can be used, even geographic areas (like a national park or a small island) can be included.
- A file with a date category should always have at least one topic category.
General comments
[edit]Limitation of mediawiki search engine
[edit]i think the root cause of all these problems about "intersection" cats is the search engine of mediawiki.
i think all categories should just be about a specific quality. like sdc properties each of them only addresses one thing.
and when users need to find say "photo of trains in london on 2024-03-11", they can adjust the filters like
type=photo; depicted=train; location=london (subdivisions of london also implied); date=2024-03-11
then we dont need Category:London photographs taken on 2023-03-11 Category:Railway photographs taken on 2024-03-11 etc.
before we have something smarter to search through the huge repository, there's always argument for assembling photos about a certain topic over a certain period into a category. why do we do that actually? mostly because category pages have the thumbnail view.
on the other hand, super refined subcategories will not be a problem if we can just browse all content under a specific category easily. right now the only way to do that is deepcategory search which is limited to 256 cats. or we have to use petscan.--RZuo (talk) 17:57, 10 May 2024 (UTC)
- mw:Help:CirrusSearch#Deepcategory says it's configurable. can we have a higher limit for commons? RZuo (talk) 18:05, 10 May 2024 (UTC)
So if we agree to only sort photographs by date based on countries that would apply to the subdivisions of the United Kingdom as well, correct? (Northern Ireland, Wales, England and Scotland)? What about the Kingdom of the Netherlands which is comprised of the countries Aruba, Curaçao, the Netherlands, and Sint Maarten? I noticed there were some complains about the first in the project chat--Trade (talk) 19:52, 11 May 2024 (UTC)
- Indeed, Scotland is not listed in Category:Countries of Europe by name, and it is clearly a subdivision of the United Kingdom. Ymblanter (talk) 19:54, 11 May 2024 (UTC)
- How is Scotland anymore of a subdivision that the Faroe Islands is? The category doesn't very consistent Trade (talk) 20:19, 11 May 2024 (UTC)
- Scotland is of couse much more a subdivision than the Faroe Islans is, but for example I think that Aland should not be in the list. My point is however that we already have a lot of things on Commons which operate with the notion of a country, and if anyone wants to change this notion it is fine, but what is meant by "country" here must be the same as in all other contexts on Commons. Ymblanter (talk) 20:26, 11 May 2024 (UTC)
- I ran into a similar issue with P17 (country) on Wikidata a few days ago. Apparently Scotland is a country but it's also located in the country of the United Kingdom? I'm not really sure. Except that someone changed "Country: Scotland" to "Country: The United Kingdom." If you want another example look into how the various break away states that aren't intentionally recognized get treated on here. The point is, I don't think something like what makes something a "country" is universally agreed on or even really matters that much in the grand scheme of things. --Adamant1 (talk) 05:38, 21 May 2024 (UTC)
- @Adamant1: per Countries of the United Kingdom, Scotland, England and Wales are always referred to as countries whereas Northern Ireland is sometimes referred to as a province. Also should note that they is a very prominent Scottish independence movement since the Scottish National Party came into power in 2007, there was a 2014 Scottish independence referendum in which 44% wanted to succeed and a 2022 Supreme Court ruling that the Scottish Parliament does not have the power to legislate for a second referendum. Sahaib (talk) 06:58, 21 May 2024 (UTC)
- I ran into a similar issue with P17 (country) on Wikidata a few days ago. Apparently Scotland is a country but it's also located in the country of the United Kingdom? I'm not really sure. Except that someone changed "Country: Scotland" to "Country: The United Kingdom." If you want another example look into how the various break away states that aren't intentionally recognized get treated on here. The point is, I don't think something like what makes something a "country" is universally agreed on or even really matters that much in the grand scheme of things. --Adamant1 (talk) 05:38, 21 May 2024 (UTC)
- Scotland is of couse much more a subdivision than the Faroe Islans is, but for example I think that Aland should not be in the list. My point is however that we already have a lot of things on Commons which operate with the notion of a country, and if anyone wants to change this notion it is fine, but what is meant by "country" here must be the same as in all other contexts on Commons. Ymblanter (talk) 20:26, 11 May 2024 (UTC)
- How is Scotland anymore of a subdivision that the Faroe Islands is? The category doesn't very consistent Trade (talk) 20:19, 11 May 2024 (UTC)