Wikipedia talk:Categorization/Archive 7

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

←

Archive 5

Professions subcategorized by Nationality

General discussion

Note: This conversation started at Category talk:Film actors.

I would like people working in this category to consider adding ALL film actors back into this category. The decision to subcategorize everyone was made before there were category TOCs. Nationality is an artificial distinction to many actors who are multi-national. I could see the utility of being able to browse through a category of all actors. You would still have the choice of browsing by nationality. -- Samuel Wantman 21:15, 21 January 2006 (UTC)

I think that's crazy. I don't see how you're supposed to browse a category with thousands of articles in it. The table of contents helps if you know the name...which you don't need a category for in the first place. What you're proposing defeats the purpose of subcategories and goes against precedent in other categories. For instance, it's possible that I'd want to browse thru all American people, not just American botanists, or Californians. But we don't keep Americans in both categories, because being in the subcategory automatically puts you in the parent category. You can always browse the subcategories, too. Categories as big as film actors was last week are simply unbrowsable, but when subcategorized, they can still be browsed, while people can also check out Category:Mexican film actors and so on, getting more information without any more effort. NickelShoe 21:57, 21 January 2006 (UTC)

As per Wikipedia:Categorization#Guidelines. It does make sense, for instance, for award-winners to be both in the subcat and the parent cat, because award-winners are a small amount of actors, and because a browser might not have any idea an actor won the particular award. But all people have nationalities, and where the nationality is a question, then we can leave it here in Category:Film actors or in multiple subcats, such as Category:American film actors and Category:Canadian film actors. Duplication here doesn't seem any more useful than duplication at other occupations, and such a large occupation makes duplication more cumbersome here. NickelShoe 22:06, 21 January 2006 (UTC)

It is not as big as you think. Most of the film actor by nationality subcategories do not have many people in it. I suspect the combined listings would be about three times bigger than the American listings. It doesn't defeat the purpose of the subcategories, they would still exist. The past precedent was because of an inability to easily browse through large categories. With a table of contents, this would not be difficult to browse through. I am not proposing getting rid of the subcategories, I just proposing to add the complete list of film actors here. I do think there is a point when large categories would be unweildy, probably when there is more than about 600 names for each letter of the alphabet. I don't think this category would be that big. I see advantages, and I don't see any disadvantages. It has never been against policy to have articles listed in categories and their grandchild subcategories. The guidelines (which I just concluded facilitating the rewrite), say that these decisions should be made to help people browse through categories. Since I am just proposing to add to what we already have, I don't see how it makes things worse, and it might make it better for some people who would like to see the complete list. For example, let's say someone is trying to remember the name of the actor and all they remember is that it started with a B (or was it a D?) The way things are now, it would be hard to find if you don't know the nationality of the actor. It would be easy if they were all here. -- Samuel Wantman 22:20, 21 January 2006 (UTC)

It seems very awkward to me to have only the people with unknown or multiple nationalities listed on this page. It should be all or none. -- Samuel Wantman 22:21, 21 January 2006 (UTC)

I think the line of unbrowsability is much lower than that, and in any case, this category was that big. I've moved hundreds of people out of it in the past few days. I know several of the subcats are pretty small--like Category:Welsh film actors. But there's a lot of nationalities, and if you don't pull out some nationalities into subcats, this category has four hundred actors per letter (because there are plenty of people not in here before I started subcatting). Well, I wasn't counting, but you're free to look at my contributions to get an idea.

I can only speak for myself, but if I can only remember what letter something possibly started with, the category is actually unuseful to me around fifty. I'm sure others can use it up to two hundred.

I suppose leaving unknown nationalities here is a little weird...I was thinking as a temporary solution until they were either in multiple subcats or somebody actually included their nationality in the article. NickelShoe 22:30, 21 January 2006 (UTC)

Since you don't like browsing through more than 50 names you would not be forced to, you could browse the subcats. But many people might. It certainly helps deal with the people of unknown nationality, and in general might make for less work. The natural inclination is for editors to add people to this category. Why fight it? I can imagine that this would be a shock to you if you've just recategorized hundreds of articles. All I can say, is that I'll help you put them back. I would like to hear from others about this. And, BTW, I'd like to make the same change to film directors. Quick, can you tell me what the nationality is for Roman Polanski? -- Samuel Wantman 22:40, 21 January 2006 (UTC)

But...if I know his name, why don't I just type that in? (And it should go without saying that I'm going to hold off any recategorization up or down until there's some kind of consensus here.) NickelShoe 22:45, 21 January 2006 (UTC)

I guess I should be more explicit about my point about Roman Polanski. The point I'm making is that "by nationality" often constrains categories artificially. There is often nothing particually notable about an actor or director being from one country or another. They are, like in Polanski's case, often from one country, raised in another, got famous in a third and then moved somewhere else. My point is that while some people might find the distinctions of nationality to be notable, it often is not. Because of that, I'd prefer seeing categories populated at a higher level when possible. This would not apply to Category:People by nationality because that is the entire point of the category. Everywhere else, when nationality intersects with profession, I'd like to populate the smaller categories by nationality and larger categories by profession unless nationality is integral to the profession (like politicians). -- Samuel Wantman 23:05, 21 January 2006 (UTC)

I have also been doing some work on subcategorization. I guess my greatest concern is that a consensus be established and that the consensus then be posted at the top of the discussion page of the relevant categories. Otherwise, I believe, what we'd tend to get, (and what we probably had before the subcategorization effort) is some individuals only in the grandparent category and some individuals only in the grandchild (by nationality) category, and some in both. That, to me, is worse than the alternatives of either a very big grandparent category, or having to go to the grandchild categories to browse. Whichever way the consensus goes, I would help in the effort to make things consistent. I don't think the multi-nationals or unknown nationalities are a problem. Multi-nationals can be listed in multiple nationality categories. Unknown nationalities can be researched to determine their nationality. In general, I tend to prefer very small categories :P but I guess that with the TOC, 200 names (1 page) per letter, or less, is reasonable. -- LiniShu 23:28, 21 January 2006 (UTC)

I decided to do some analysis of this category. There are about 3300 Entries in Category:Film actors Almost all don't seem to be listed in the nationality subcategories. For the subcategories there are:

1113 Americans
3 Argentines
203 Australians
5 Austrians
83 British which overlap with 107 English 17 Scottish and 8 Welsh
59 Canadians
2 Danes
23 French
5 Germans
13 from Hong Kong
7 Irish
4 Israelis
4 Italians
71 Mexicans
2 Poles
5 Russians
13 Singapore
3 Spaniards
6 Swedes

At most there would be about 5000 total names in Category:Film actors which would not make the category very much more crowded than it is now. I also notice that there is already Category:Actors by nationality which makes me wonder why we would need Film actors to also be broken up by nationality. Film actors is already a subcategorization of Actors. -- Samuel Wantman 00:12, 22 January 2006 (UTC)

I support the inclusion of e.g. American film actors in the Film actors category. It's true that many people might find the category too big to browse, and so would use the subcategories instead--but if people are interested for whatever reason in film actors in general, then as big as that is, that's what would be best for them.

Look, my copy of the Video Hound book has thousands of movies listed in it, and sometimes I'll browse through it to see what I'll find. Other times I'll be interested in finding a specific kind of film and I'll browse through the book's categories. Why does what works in a printed reference work not work for an online encyclopedia?

The point I want to stress is that the Film actors category ought to have some use, otherwise it should be a category empty except for the subcategories. Who would be interested in the set of "Film actors who are not listed in a subcategory"? To whom is that a useful group?

When a consensus is reached, putting the guidelines on the top of each category is of course a great idea.

The question of whether you need actors by nationality and film actors by nationality is a good one. In terms of what people might be looking for, it's not clear why having separate subcategories for say French actors and French film actors would be helpful. I've thought the same thing about musicians by state and musical groups by state--I spent an evening recently separating out the individual musicians from the musical groups into the two respective subcategories for Massachusetts, and now I'm wondering if someone interested in seeing the Massachusetts musical scene at a glance wouldn't be better served by a single category that combines both the individuals and the bands. Sometimes there's an obvious way to break articles up that really is not very useful.

Nareek 04:42, 22 January 2006 (UTC)

The more general question is "When is a category too big?" The push to subcategorize started before the implementation of category table of contents. I don't know if this question has been really addressed since then. My take on this is that categories get subcategorized into tiny pieces often because it is not possible to do a database selection. These small categories like Category:Polish film directors have minimal browsing value compared to being able to browse through Category:Film directors. These multi-attribute categories are often too small. As a guideline, it seems reasonable that categories be populated at the same level as articles about the subject. We don't have an article about Polish film directors but we do have one about Film directors. It also seems that above about 6000 articles a category would become very unwieldy. -- Samuel Wantman 08:40, 22 January 2006 (UTC)

Thanks, Samuel, for taking the time to do some numbers analysis for this discussion :) And also, thanks for raising the question of "When is a category too big?" I think it would be helpful if we also had some consensus on that. Not a hard and fast rule, and I could imagine there being exceptions in some specialized areas of knowledge, but, I would appreciate, for example, when I look at Category:American actors and think, "wow, this category is really big", being able to go back to the consensus that has been reached, and to answer myself "ok, it's not too big, yet." Your suggested guideline of categories being populated at the same level as articles about the subject is a good one, with the possible exception of when that category is just absolutely too big per the consensus that I hope we'll arrive at.

Another point I'd like opinions on - this discussion began with the question of whether articles should be in both the grandparent category Category:Film actors and the grandchild categories (Category:American film actors, for example); I would like to know whether the articles in Category:American film actors should also be listed in their other grandparent, Category:American actors. That is a category that might be over 6000 if all entries were listed there. (Sorry, I haven't counted yet how many are there right now). Actually, as I think about it, there may be many articles in Category:American actors or Category:Actors that are not yet listed in any of the subcategories of Category:Actors by medium So, the count of about 5000 names for Category:Film actors might be higher if everyone was listed there who should be listed there. -- LiniShu 12:44, 22 January 2006 (UTC)

Category:American actors has 4355 entries at this time. This would not count those that are only listed in a grandchild American actors by medium subcategory (whether recategorized by myself and others, or listed directly in the grandchild subcat in the first place.) Category:American actors currently has a too large tag on it. If you read the discussion page, it looks like there was previously a too large tag, which Samuel removed, noting so, with his reasons, on the discussion page, in August 2005. The tag was added again, without discussion, on 14 January 2006. I began, that same day, slowly, to depopulate the category into the American actors by medium subcategories. --LiniShu 13:21, 22 January 2006 (UTC)

What exactly does "unbrowsable" mean? You can browse through an unabridged dictionary or a multi-volume encyclopedia, and plenty of people do. If you mean it's not practical to look through the entire category to find someone specific whose name you've forgotten, then that might be a pretty low number (depending on your patience). But there are plenty of other reasons you might be browsing a category like "actors." Maybe you want to see what kind of names actors have. Maybe you want to see how many actors are in Wikipedia. Whatever--if there aren't any reasons people might be interested in actors in general, then why have anyone in the category? The fact that a particular size might be cumbersome for a particular use is an argument for addding subcategories, but it's not an argument for eliminating the category. It's not like it's taking up physical space.

Nareek 16:34, 22 January 2006 (UTC)

My $.02: First, since previous upgrades to the Wikimedia software, large categories no longer pose the technical problems that they once did. So the size of a category alone should not be the rationale for breaking into subcategories. Second, "profession by nationality" is an extremely arbitrary breakdown -- many professionals simply do not have any intrinsic association with a nationality. So while I have no objection to creating these subcategories for those professionals who do have a strong association by nationality, these subcats should not mean that the primary category is de-populated. older ≠ wiser 16:57, 22 January 2006 (UTC)

So it sounds like there is general consensus for the notion that categories should be fully populated at the "topic article level". I still don't know the criteria for where the upper limit is. Is it Category:Film actors, Category:Actors, Category:Entertainers, Category:Celebrities or Category:People. I strongly agree with populating Film actors, not certain about Actors and tend to think anything higher is too much. One criteria is to not put an article into every possible category because that clutters the article with categories. I'd propose that we only populate to the level of notariety (there's probably a better way to say this). If people are notable for being Directors, that is the highest level of populating film people categories. There are still grey areas with this, Poets or Writers? Film actors or Actors? Any ideas? What criteria do we use to make these decisions? -- Samuel Wantman 21:06, 22 January 2006 (UTC)

I favor categories that are not subdivided into a hundred subcategories. Unless one already knows that a professional is "Scottish" rather than simply "British", he or she might not be found by browsing at all. Categories are already sorted by alphabet. It'd be much simpler, overall, if people were identified as "Scots" and "Actors", rather than as "Scottish actors". -Will Beback 09:08, 23 January 2006 (UTC)

Points that I think we're on the way to establishing, so far:

There really is no such thing as a category that is too big, by any kind of absolute standard.
It is reasonable to keep categories farther up the heirarchy populated, to the "topic article level" and/or "to the level of notoriety" (which could maybe be defined by the opening sentence of the articles?) Our articles say things like "Clark Gable was an American film actor..." or "Sir Alec Guiness was an Oscar-winning English actor..." - note the use of the word actor rather than entertainer, celebrity, or person :)
The contributors to this discussion have different preferences for what level of categorization/ subcategorization they find helpful for browing, and presumably our encyclopedia users will also. Even to the same individual, different methods of browsing may be helpful on different occasions, depending on the object of the browse. We can accomodate all by using the subcategories that currently exist or others that might be requested, and also by keeping categories farther up the heirarchy populated too, according to the criteria in the bullet point immediately above, or as requested.
The possibility of "cluttering" articles with too many categories and subcategories should be considered, but is a secondary concern to having useful categories, possibly at multiple levels.

--LiniShu 12:09, 23 January 2006 (UTC)

I've asked Brion Vibber (a developer) about whether there are any technical concerns about having large categories. -- Samuel Wantman 07:13, 24 January 2006 (UTC)

Thank you --LiniShu 11:28, 24 January 2006 (UTC)

This is the response from Brion Vibber. "THERE IS NO ISSUE WITH THE SERVERS DUE TO LARGE CATEGORIES". So the only issue is the awkwardness of browsing large categories. -- Samuel Wantman 02:04, 25 January 2006 (UTC)
There is certainly no need for any pause to the improvements to the accuracy of categorisation. Things are much better than they were a year ago, but there is still lots to do. No search system is as good as a category system which does some of the work for people. I have found and categorised hundreds and hundreds of articles which were only in the subject category or only in the national category, and the combined categories filter articles through so they are in both. It shouldn't be forgotten that for many occupations and nationalities there are quite small numbers of articles, and without the precise categories they would belong in two big categories, so the six "Fooian xers" would be lost among say 1,252 Fooians (upmerged from 200 different categories) and 2,213 Xers (upmerged from a 150 different categories. Both of these categories would be a useless seas of unfamliar names for most people. It is not a question of whether random browsing is possible, but rather a matter of helping people to target their browsing. People who know exactly what they are looking for can use the search box. By contrast the category system is a natigational tool which helps people to identify groups of related topics. This is especially relevant to people from smaller countries, as the few articles on people of each occupation of their nationality would get lost in a sea of foreigners. In any case starting the conservation with a category where nationality is not quite as relevant as in some other occupations (though it is still very relevant, especially for actors from outside the English-speaking world) has put this discussion on a somewhat false footing. CalJW 05:10, 25 January 2006 (UTC)

I don't think anyone is advocating that we remove any of the categories that have been added over the last year. There would still be all the "Fooian fooers" categories. This discussion is about adding all the fooers from all countries back into a single category. Some people might look at these large categories and find them useless, but they could still move to the smaller subcategories. I often find the small categories useless and wish they were combined. This option to look at the contents is small pieces or in one large category would add flexibility to the categorization scheme and would serve the needs of more users. -- Samuel Wantman 06:18, 25 January 2006 (UTC)

I am also very much opposed to dividing categories of people into subcategories solely based on nationality. I doubt that everyone browsing Film Directors really cares to look through the lens of the person's nationality. Why limit the ability to look at the category in different ways? It doesn't effect the subcategories at all, those are still listed at the top of the page and readily accessible. Cacophony 09:10, 25 January 2006 (UTC)

Proposed consensus statement for professions subcategorized by nationality

In cases where subcategories of the type "Fooian fooers" exist, it is allowable and acceptable to have the individual fooer articles also listed on the "fooer" category page. This can be done on a case by case basis as editors/ users express the desire to be able to browse articles at the "fooer" level and as there are editors willing to work on the repopulation of the "fooer" category if necessary.
The object is to accomodate different preferences in browsing as expressed in the discussion on this topic.
The question of the effect of large categories (1000's of entries) has been raised with developers, and we are told that the nature of the category interface is such that large categories are not a drain on the server.
The intention is that "ancestor" categories be populated up to the topic article level and/or the "level of notability" as indicated by the identification of the individual's profession in the opening sentence of typical articles in the category. This can be addressed in specific on a case by case basis. We would refrain from automatically populating every category up the heirarchy to Category:People, for example, so as not to clutter articles with categories.

--LiniShu 02:11, 26 January 2006 (UTC)

I would go even further and say "it is recommended to have the individual articles also listed in the larger categories." Also, I'm wondering if this applies to other categories besides "fooers" and subcategorization other than "fooian". I'll look around the category structure and see what I find. -- Samuel Wantman 01:54, 26 January 2006 (UTC)

Can we look at this from the other direction? Try categorizing a musician. If I followed the suggestion to put him or her in all the categories below the level of musician I might put him in Musician, American musician, American jazz musician, American saxaphonists, and California musicians, not to mention what bands he might belong or belonged to. I don't think many people writing an article on a musician will put him in all the relevant categories if this is the rule, leading to incomplete parent categories and more work for those trying to maintain the system. There's also the drawback of dozens of category links at the bottom of the page.

Perhaps a technical solution would be best. Put everyone in the most restrictive categories and have a way to specify that you want to see all the pages in the current category and all child categories (but I'm in danger of bringing in the circular categories discussion into this one so I'd better stop here). --JeffW 06:00, 11 March 2006 (UTC)

Category maintenance is a very real concern, and there is a technical solution that is already possible. It would be possible to create a bot that would populate categories by looking for specific templates. This is similar to what is already done when categories are renamed. After consensus at WP:CFD the template {{categoryredirect}} is added to the category. A bot looks for categories with this template and moves all the articles to the renamed category. We could create a template, perhaps calling it {{Category duplicated}} that would add a message that says, "all articles found here are also part of category xxx" Xxx would be the parent of the category that gets repopulated. So for example Category:Polish film directors would have the message that says "all articles found here are also part of Category:Film directors". Whenever anyone adds an article to Polish film directors, it would automatically get added to film directors. Likewise, a bot could look for film directors without a nationality and add them to a Category:Film directors of uncategorized nationality. These are tasks very well suited for a bot. -- Samuel Wantman 09:06, 14 March 2006 (UTC)

Action item(s)

Possible general action items if the consensus statement is accepted:

Is there cause to update Wikipedia:Categorization/Categories and subcategories section on "Reasons for duplication"? Is our "desire to browse articles at the topic article level" a distinct reason from those already mentioned? --LiniShu 02:21, 26 January 2006 (UTC)

Comment - There have been a relatively small number of editors involved in the discussion to this point, although it is in a central location as far as categorization is concerned. I would recommend that, after leaving the discussion open a little longer, we then proceed with the effort of making the categorization of individuals in the film, television and theatrical professions consistent. I would not be surprised, though, if the general topic was revisited in other times/places. Hopefully, though, the current discussion may be good groundwork for any future discussions. Thanks, Samuel, for all of your due diligence on the issue of category size. --LiniShu 02:40, 26 January 2006 (UTC)

It sounds like a good plan to me--wait a bit, and if there are no further objections, use the actors category as an experiment to see how well a more-or-less fully populated category tree works. Then, if it seems like a good idea, we can meet back here and use it as an example of why the recommendation maybe should be changed. I'm kind of new around here; does that sound wikikosher?

I would argue for consistency in dealing with categories up to Actors. (I could see including a populated category of People, for a guide to all biographies in Wikipedia, but that would be outside the scope of doing one branch as a test case.) I would say that Actors should be fully populated, as well as Film actors, American actors and American film actors. That's a minimum of four categories for every actor--that does not seem like too many to me. Thus: Category:Actors; Category:NATIONALITY actors; Category:FIELD actors; Category NATIONALITY FIELD actors.

Is there another kind of category that every actor should have? If we want to fully populate the tree, adding another category would mean adding four more categories to every actor (e.g., DECADE actors, NATIONALITY DECADE actors, DECADE FIELD actors, DECADE NATIONALITY FIELD actors). But we don't have to be strict about this: If we wanted to do the decade thing, for instance, we could say we'll only do it for a few specific types, like American film actors and British film actors. Then only those selected types would get an extra category, and those only one (DECADE NATIONALITY FIELD actors).

Nareek 05:24, 26 January 2006 (UTC)

Okay, but keep in mind that most actors aren't simply "American film actors"--they're also "Jewish American actors", "American film directors", "Entertainers who died in their fifties", and "People from New Jersey". I'm not saying this is the end of the world, but let's keep in mind we're not always talking about four categories. NickelShoe 05:47, 26 January 2006 (UTC)

Very true, many will have more than four categories. But for purposes of this project, there should be a set of categories that should be given to every actor. Are those four it, or is there some other grouping that everyone should have?

Nareek 06:03, 26 January 2006 (UTC)

Assume that every topic level category is broken into subcategories by nationality. According to this proposal, each article ends up in two categories. If it is also broken by another attribute, that puts the article in 4 categories. 3 attributes yields 8 categories, etc... with the result that N attributes yield 2^N categories. For this reason, restraint is necessary. Eventually, hopefully, the software will be modified to tag articles with these attributes in a better system, or allow database searches. Until that day arrives (and if it arrives) we cannot expect this system to handle multiple attributes. For this reason, I'd like to propose the following. That duplication of categorization does not go above the lowest possible topic category (for example Film actors), and that subcategorization of that lowest level should only be done with parallel subcategories, which means that they would all be subcategories of the topic category and no attributes would be combined. So for example, if the topic level category is Category:Film actors, then Category:Film actors by decade it should not be combined with Category:Film actors by nationality to create Category:American film actors of the 60's. The point of this proposal started with trying to make it possible to browse through topic level categories, not to create and populate every possible subcategory. Let's do this a small step at a time and see how it goes. For now, let's just re-populate Category:Film actors and Category:Film directors and take it from there... -- Samuel Wantman 08:23, 26 January 2006 (UTC)

Since the task involved is going to the articles in a category and making sure they're all in one or more other categories, it's not really harder to do four at once than two. If you go to "Category: American film actors" you'd open each article and paste in the following:

"Category: Actors Category:Film actors Category:American actors"

All these articles would already be in Category:American film actors. You might have to then delete some categories if there's duplication.

Since it's pretty much the same amount of work to paste in one category as three, it seems like we ought to do this the way we want it to turn out. Am I wrong in thinking that the consensus here is in favor of a fully populated tree?

Nareek 13:27, 26 January 2006 (UTC)

I'm attempting to kind of "redirect" the discussion on the specifics of the acting and directing professions to the section below :) Thanks, --LiniShu 13:35, 26 January 2006 (UTC)

Application to film, television and theatrical professions

Regarding the scope of the original discussion (and the area for which we have editors willing to work on consistency of categorization at this time): If the consensus statement proposed above is accepted, we have already established in the discussion that the following "grandparent" categories should be populated with all of the articles in their descendant "by nationality" categories: Category:Film actors, Category:Film directors, Category:Television actors, Category:Television directors, Category:Stage actors, Category:Theatre directors Are there any other similair "fooers" categories to add to this list at this time? --LiniShu 03:01, 26 January 2006 (UTC)

Other "ancestor" categories in question- We have Category:Actors, Category:Directors, and "Actors by nationality" categories such as Category:American actors. Should any of these be populated with individual articles? If no one has a strong opinion, we could go with the status quo for now; current general usage based on numbers of articles, is to not have "Actors" and "Directors" populated with individual articles, but we do have the "Actors by nationality" categories populated with individual articles. --LiniShu 03:13, 26 January 2006 (UTC)

I'm taking the liberty of pulling out a couple of comments from the section above, relevant to what we do here and now with the film, television, and theatrical professions (--LiniShu 13:31, 26 January 2006 (UTC)):

From Nareek 05:24, 26 January 2006 (UTC) - Populate Category:Actors; Category:NATIONALITY actors; Category:FIELD actors; Category NATIONALITY FIELD actors with individual actor articles.

From Samuel Wantman 08:23, 26 January 2006 (UTC) - Let's do this a small step at a time and see how it goes. For now, let's just re-populate Category:Film actors and Category:Film directors and take it from there...

My preference - My "pet question" :P I really want to know what to do with Category:American actors - because that's the avenue by which I became involved in this discussion. I can go through the category making sure each actor is in the appropriate Category:FIELD actors and Category NATIONALITY FIELD actors, but do I then remove them from "American actors" or leave them in? And do I add back those I had removed before this discussion began? I'd like to make the usage consistent, one way or another. Status quo and lesser amount of edits required would probably be to populate rather than depopulate. --LiniShu 13:31, 26 January 2006 (UTC)

I think we should move this discussion to Category talk:Actors so that if there's anyone following that page, they can have a chance to participate in the discussion. I'm going to copy and paste some of the actor-specific talk here to there--maybe some of it should then be deleted from this page? It's getting kinda unweildy.

Nareek 15:13, 26 January 2006 (UTC)

Proceeding from here

At this time, among this group we have clear consensus on: do re-populate Category:Film actors and Category:Film directors, do not re-populate up the heirarchy tree above Category:Actors. There is less clear agreement on "Actors" itself, and the NATIONALITY actors categories. We also have the "cousin" categories of the Film categories - Stage and Television. I've been thinking this over the past couple days - I think that those who are interested in helping with the repopulation effort for the "MEDIUM(Film,Stage,Television) Actors" (or directors) categories should go ahead and begin, maybe focusing on the film category first (one has to choose somewhere to start), but at the same time, as we "touch" any articles to add in MEDIUM Actors category tags, why not make sure, at the same time, as Nareek suggested, that the NATIONALITY Actors and Actors category tags are there also? Of the group that has given consensus on populating categories farther up the heirarchy tree than the NATIONALITY MEDIUM Actors categories, no one showed significant opposition to populating NATIONALITY Actors or Actors, and our "populate to the level of notability" criterion could be applied either at the MEDIUM Actors or Actors level, and, interest has been expressed in having Category:Actors populated for browsing.

As we begin this work, it would be good to incorporate Samuel's new "Allincluded" template on the appropriate category pages.

I am planning to proceed as I have described above, if there is strong objection, or a different consensus is reached, my changes can always be undone.

Nareek, I am adding this latest to the Category talk:Actors as well; wasn't sure where it would be most likely to be observed, here or there?

NickelShoe, you were part of this discussion at the beginning; I know that the repopulation of the higher level categories is a new direction from the way we were going before, hopefully, after the discussions above, and the checking of whether categories can be too large (from a technical standpoint), this conclusion is something that you can live with? Your opinion is valued; you do a lot of work improving the quality of Wikipedia, and you had already put in a lot of effort in organizing the Actors categories according to the generally held application of WP:MOS prior to this discussion.

Feedback from all is appreciated, as always. Thanks,Lini 13:25, 29 January 2006 (UTC)

Yeah, I think it makes the process of categorizing articles more complicated (rather than simply using the most specific category, one has to figure out when to also use general categories), but I trust that if others say it's helpful to be able to browse categories I find unwieldy that they're telling the truth. So I wouldn't stand in the way. Hopefully some day we'll have cross-referencing cats. NickelShoe 19:31, 29 January 2006 (UTC)

I'm against any change. An actor's nationality is important. It effects the type of parts they get etc. We don't need a huge category, some of them are more than big enough as it is. Golfcam 22:41, 24 February 2006 (UTC)

I very strongly object to a swing to dogmatism in the opposite direction. Extrapolating from a special case such as film actors (not that I agree with any change even there) is a wrong headed proceeding. There will be many cases where this sort of multiple categorisation is very messy because it will push many of the subcategories off of the first page. It will also generate immense category clutter on some article pages. And in any case, how do you decide which categories should have topic articles? There are many which don't exist, but could or should. This proposal is for a scattergun approach, when we should concentrate on flexibility and accuracy based on what is best in each area. CalJW 05:46, 2 March 2006 (UTC)

Actors by time period

A wild and crazy idea - for those who prefer smaller subcategories for browsing - for American actors, even the most precise subcategories that we have now Category:American film actors, Category:American stage actors, and Category:American television actors are not small categories. "American film actors" at the time of this writing has 1331 entries; more will be added in the effort to consistently categorize "American actors", and the category will only keep growing with time. With the above discussions about the disadvantages of "over-subcategorizing", I'm not in a hurry to create these new subcategories, and I would not want anyone reading this to rush out and create them, but... what about the idea of actor subcategories by time period, as we have for musical groups? Some actors like George Burns, of course, would have quite a few decades of activity, but others would have only one or two. Advantages would be: a.) we'd have subcategories that at some point stop growing very much b.) we could see people grouped with their contemporaries c.) an actor's time period implies something about the cultural climate in which they worked. What do you all think? (If you are tired of more and more subcategorization, "please don't shoot me" for suggesting this :) - just an idea) Thanks, --LiniShu 03:33, 26 January 2006 (UTC)

The only problem would be the number of categories added to individual articles. You would probably be better off creating a list of film actors by time period. —Mike 20:33, 12 February 2006 (UTC)

Theoretically yes, but this is a wiki and it is very clear from precedent that categories are better maintained than lists in most cases and this is becoming more and more the case. For example most of the country-topic lists have been virtually abandoned, and the exceptions are mostly maintained by a single enthusiast, whereas many people contribute to category maintenance because it can be done as an adjunct to visiting an article. CalJW 06:02, 2 March 2006 (UTC)

Category:Architects is easier to maintain than List of architects, and the category is made up of articles or stubs with real content. The list tends to accumulate vanity listings, people just tack their name on the list and contribute no article. In general, the structure of architecture is an example for this entire discussion. Architects are listed (as an 'instance') in the format discussed above, namely: Category:NATIONALITY architects, in parent Category:Architects by nationality, fooian fooers. This can go a bit to far, perhaps, in what seems to be extremely narrow categories attached to Alvar Aalto (i.e. Category:Natives of Southern Ostrobothnia).
The time period for an architect is found in Cat:STYLE architects, such as Category:Baroque architects or Category:Postmodern architects. There's also the ambitious Timeline of architecture and parent Category:Years in architecture. Infoboxes can also be used for showing the span of years that an individual worked or played music (60's, 70s, 80s, etc.). The List of architects is arranged by decades, with some overlap between the 1990s and 2000s, which can be revised.
From discussion above: "Populating categories with articles... ": Articles in the 'notable level' of Cat:Architects can be reduced to a set of "household name" or "notable" architects (who will be disputed, no doubt). Otherwise, the category will be hard to browse - a mix of notables and, less notable. If, for example, a user misspells a name in the search box (and there's no redirect), they might browse the categories for answers. But, to find Aalto, the user would have to know his country Category:Finnish architects. So the duplicate listing (in 'grandparent' and 'grandchild') would be helpful to a general audience, looking for a world famous architect at the level of Category:Architects. -- Dogears 09:11, 3 March 2006 (UTC)

But why limit it to just the notables? Categories serve as Wikipedia's index and table of contents. I'd put ALL architects in Category:Architects (duplicating the architects by nationality subcategories) and put the notables in List of notable architects which can be annotated as to why they are notable. This list can be "piped" so that it is at the top of the category as List of architects already is. A user looking for the famous can find the list of notables, someone wanting to browse all the architects could look through Category:Architects, and others could look through the architect categories by nationality. This would make it more useful for more people.
There are many reasons why someone would want to browse through all the architects. Perhaps they can't remember the name of an obscure architect (Starts with "G"?) Or perhaps they are an expert on Architecture and are scanning the category to see what architects are missing. They might just be looking to compile a list of architects. -- Samuel Wantman 10:17, 3 March 2006 (UTC)

Articles for specific dates

I just noticed a new trend: articles for specific dates, such as January 18, 2005. Not sure I like it (note the related AFD discussion), but lots of them now exist. For the purpose of this talk page, I'm wondering how they should be categorized. The one I linked to above was placed in Category:2005, definitely not a good choice (I removed it). Most of the others seem to be uncategorized at the moment. So how should they be categorized? - dcljr (talk) 22:54, 24 January 2006 (UTC)

It should go in Category:2005, unless there's a better subcategory for it to go into. I mean, how is January 18, 2005 itself not a direct subset of the year 2005? -Sean Curtin 01:53, 25 January 2006 (UTC)
It's now in Category:Days in 2005, which is a pretty obvious solution. CalJW 05:20, 25 January 2006 (UTC)
- Perhaps the formatting of Category:Days in 2005 could be improved. Currently the days are categorized like this ...|04/10 which is good because everything will be in chronological order, but the first 9 months will be under the heading "0" and the last three will be under the heading "1". It might look better if they were all categorized like this ...| 04/10 By adding the space there won't be any headings and the entire year will be chronological. --Samuel Wantman 08:05, 25 January 2006 (UTC)

The one problem I see with January 18, 2005 is that it is really a subpage masquerading as an article. It doesn't have the normal intro paragraph, and I would expect to see some sort of navigational template. If these were added, it would probably screw up the formatting of January 2005. —Mike 20:49, 12 February 2006 (UTC)

The solution to this problem would be the noinclude tag. dcljr (talk) 23:37, 12 February 2006 (UTC)

Category purpose question

A question concerning the purpose of categories: Is the primary purpose of categories to:

Aid the reader in finding material which may be of interest, or relevant to a particular topic?
Producing a taxonomy; wherein being included in one or more categories is an indication--nay, a declaration by the Wikipedia community--that the subject of an article is an instance of the categories it is included in.

I seem to suspect the latter; and the article advises that "Categories appear without annotations, so be careful of NPOV when creating or filling categories. Unless it is self-evident and uncontroversial that something belongs in a category, it should not be put into a category.".

However, one side-effect of that rule is category fights, wherein two (or more editors) will have a dispute over whether a topic belongs in a particular category, or not--or constantly move the article back and forth between two different categories. Some categories are seen as marks of prestige (such as Category:physics, whereas others are seen as badges of dishonor, like Category:Pseudoscience); I can think of several articles in the realm of physics which have been the subject of such disputes.

Perhaps a wider view of categories may be more appropriate--as in "this is a list of possibly relevant subjects; some of which may be of disputed relevance?" That might nip some of the edit wars in the bud. --EngineerScotty 23:01, 25 January 2006 (UTC)

I'm going to say that the first is definitely the purpose, however readers will still interpret it as a declaration, no matter what we say they should do. Therefore we can only be so loose with category usage. If a category is inherently a value judgment, it should be deleted or renamed. NickelShoe 02:10, 26 January 2006 (UTC)

The first guideline for categories is "Categories are mainly used to browse through similar articles". Even so, there are multiple taxonomies, and it is important that articles are not mis-categorized. Often categories with the word "related" in them are brought up for deletion at WP:CFD. Ineveitably they are deleted or renamed with the reasoning that EVERY category implies that the articles are related to the name of the category. Some categories ONLY contain related articles. See Category:George W. Bush for an example. This is frequently discussed and there is already language about this here. There is also discussion about this at the top of this page. At a practical level, I don't think we could ever reach consensus about a single taxonomy for all the articles in Wikipedia, and I don't see what use it would be. Categories are only good for browsing and discovering what articles exist here. I recently was involved in trying to mediate a controversy about whether the article Matthew Shepard should be in Category:Hate crimes. Technically, his murderers were never convicted of a hate crime, but the case is definitely related to the subject. The point of the category is to lead users to articles related to the subject, so it is reasonable that all articles related to hate crimes be in the hate crime category. -- Samuel Wantman 09:02, 26 January 2006 (UTC)

Both and neither. I'd say the Wikipedia articles themselves are the best way to list the articles related, or partially related, to a subject, since they can qualify how the topic is related. After all, the articles, not the categories, are the center of this project. The best use of categories is to provide an overview of the breadth and depth of articles in ~~a broad field~~ Wikipedia so that the user does not have to slowly explore the branches of the tree of knowledge one twig at a time. You could say the categories provide a way to speed-read Wikipedia; they should include all the links to specific persons & objects and narrower topics & concepts found on each page. GUllman 19:03, 27 January 2006 (UTC)

Categories should be a Wikipedia-based system, not "factualness"-based. We would either have tons of uncategorized articles for valid but controversial ideas, ... or each category would have dozens of parallel categories for "some people believe" or "usually mentioned in association with" relationships. The real world is too complex to describe in terms of yes and no, so you shouldn't expect to learn about a subject by just reading categories. Want to know why an article is listed in a category? Simply click on it and read the article! GUllman 20:54, 27 January 2006 (UTC)

Can't see categories when editing

Why can't you see the Categories section of a page when editing a page? Cigarette 15:14, 27 January 2006 (UTC)

Yeah, somebody had to tell me too...they're all the way at the bottom of the page, under the edit box. Interwikis are over at the side. You mean when previewing, right? NickelShoe 17:22, 27 January 2006 (UTC)

If you mean you can't see the [[Category:]] references in the edit box, note that categories are sometimes "hidden" inside of {{templates}}. Maybe that's the problem. - dcljr (talk) 19:43, 7 March 2006 (UTC)

Template for categories that include the articles in its subdirectories

I have created Template:Allincluded for use in categories that include the articles that are in its subdirectories. Here is an example of its use from Category:Bridges in New York. {{Allincluded|bridges in New York State|the bridges}} produces the following comment:

For convenience, all bridges in New York State should be included in this category. This includes all the bridges that can also be found in the subcategories.

I think this template (or some similar comment should be used when subcategorization does not remove articles from topic categories as discussed above. -- Samuel Wantman 08:54, 28 January 2006 (UTC)

Yes, this new template should work very nicely for that purpose. Thanks! ----Lini 12:24, 28 January 2006 (UTC)

Help -- didn't alphabetize correctly

I linked an article about a person to a category, and when the name appeared on the category page it was in the wrong place. It ended up alphabetized under her first name, instead of the last name like everyone else in the category. How did this happen? How can I fix it?

(The article was Paola Giangiacomo and the category was Category:Television journalists.) — Michael J 17:37, 28 January 2006 (UTC)

Wikipedia:Categorisation_FAQ#How_do_I_sort_the_article_differently_on_the_category_page.3F Cacophony 17:59, 28 January 2006 (UTC)

Thanks for the info, and thanks for fixing the link! — Michael J 00:51, 29 January 2006 (UTC)

Sorting foreign names with special letters

I've notice that some people have been changing surnames starting with Æ, Ø or Å to different letters when the pipe (|) is used to sort the articles when categorizing. So Ø might have become O and I've seen Å changed to both A and Aa. Is there any guideline or policy on how to sort these names? Just randomly substituting the actual letters with some sort of home-made standard doesn't seem to be very useful. (I'm sure there are examples of other letters as well, but being Norwegian these are letters I've noticed the most. They are the last three letters in the Norwegian alphabet.) Tskoge 14:11, 5 February 2006 (UTC)

English speakers would have no idea where to look for these letters, so I suspect they should be put where people would look for them. I don't think there is a guideline for this, but I think it should be something like this:

If the letter is the same as a letter in the English alphabet with an added accent, it should be piped without the accent. Thus, "Êxample" would be piped as "Example".
if the name is outside the 26 letters of the English alphabet, the letter in the alphabet that most looks like the letter should be added to the beginning of the name in the pipe, so "Æxample" would be piped as "AÆxample". "Øxample" would be "OØxample", etc... This would put them in order after the listings without foreign letters.

Any other ideas? -- Samuel Wantman 19:32, 5 February 2006 (UTC)

I agree that in the English Wikipedia, we're catering to English speakers, who generally don't know what order letters go in foreign alphabets. I usually pipe the cats to sort by what letter I'd assume an English speaker would look under. (Not duplexed as suggested by Samuel.) It's also probably unclear to most English speakers when they're looking at a different letter as opposed to a letter with an accent mark, so I'd be afraid to make a distinction for the procedures in the different situations.

I think that it would make the categories pretty hard to use for a lot of people if we sorted them correctly according to foreign alphabets. I think Samuel's suggestion is pretty sensible, though I can imagine issues with letters such as Æ for which alternate spellings might begin with a E, not A. NickelShoe 20:47, 5 February 2006 (UTC)

I tend to agree, especially because each language has its own collation order. Many of these letters support differently depending on language. - Jmabel | Talk 05:26, 10 February 2006 (UTC)

...And the actual sort order will be based on Unicode, anyway, not individual language conventions (Unicode breaks up many scripts into different ranges of characters, so they wouldn't necessarily be sorted "correctly" even if you knew the language). - dcljr (talk) 23:41, 12 February 2006 (UTC)

They should definitely be sorted by the letter they most look like. Anything else is just no sorting at all to English speakers. It doesn't matter whether this look right to foreigners or not becuase they have their own Wikipedias. Golfcam 22:43, 24 February 2006 (UTC)

Nickel Shoe's suggestion about sording by different foreign alphabets doesn't make any sense at all. We need to sort on the English alphabet. In lengthy categories, that's what we have for shortcuts at the top--exactly 26 English letters (in rare cases numbers as well). Note that you also don't get the sort order of any foreign alphabet either, if you don't use index keys. You get some crazy, arbitrary Unicode number basis, not used in sorting any language. But how in the world are you going to find someone, say in a category such as Category:1893 births or Category:Living people if you have some crazy hodgepodge of 87 different language rules being applied, and where in the world are our editors ever going to discover what those 87 different language rules are. Actually, in some cases, there are variations in the indexing order in the same language in different countries, too. But that's irrelevant; what we need to worry about is English language sorting order.

Sometimes the differences in whether a character is sorted as one letter or two letters depends on whether it is a ligature or a diacritical mark--that's mentioned somewhere on one of the zillion subproject pages here. But often, it also depends in part on how that person is known in English; someone with a name like Böhm might choose to spell it Boehm or Bohm or even Behm in English. Sometimes the same letter will be used in several different languages, and will be more likely to be changed to two letters in English when it comes from some of those languages and to one letter in English when it comes from the another language.

But in almost every case, whether you sort it as one letter or two letters, you will be closer to the proper English sorting order than you were with the letter with squiggles on it, even for those people who disagree with you about whether it should be sorted as one letter or two.

Same goes for the initial letter. Almost all of our categories, especially those for people, should be indexed in a case insensitive way. That means if the indexed name starts with "de" or "von" it should be changed to "De" or "Von" in the indexing key. Gene Nygaard 03:41, 11 March 2006 (UTC)

I think you and Nickel Shoe are pretty much in agreement. --JeffW 05:13, 11 March 2006 (UTC)

Changing guideline

I've altered the guideline on the accompanying page to reflect the fact that there's at least a difference of opinion on the subject of populating the category trees. (See above.) There seemed to be a pro-population consensus in the discussion we just had, but perhaps it's too soon to change the guideline completely? In case I acted hastily, I left the old language up, commented out, to make it easy to restore. Nareek 23:45, 11 February 2006 (UTC)

I think the revised (inclusionist) way should be the guideline. Nobody has given me a good reason why categorization should be limited to the smallest denominator. Why put artificial limits on categorization? It only hurts the end users by creating (sometimes) unnatural divisions. Who cares if there is a few very similar category links at the bottom of the page? Tolerating "redundant" categorization is such a ultra-minor tradeoff to having meaningful (i.e. complete) categories. Cacophony 00:23, 12 February 2006 (UTC)

I agree with altering the guideline, but I think there is more to it. I don't think that categories should be populated ALL the way up a hierarchy. This would make it MORE difficult to navigate some categories. For example, Category:Bridges has articles about the different types of bridges along with other articles related to bridges, and all the subcategories have the artcles about specific individual bridges. This is a useful division of the articles. If Category:Bridges also had the hundreds of artilces about individual bridges, it would be very difficult to navigate the other articles about bridges. The discussion above talks about populating categories that were subdivided arbitrarily (like by nationality) and only up to the level of a topic article. I don't think we should condone anything more than that. -- Samuel Wantman 11:33, 12 February 2006 (UTC)

As I've said above I endorse a flexible approach, but I am very strongly opposed to multi-level categorisation as a standard policy. The exceptions should remain exceptions. I find the idea that categorisation by nationality is "arbitary" somewhat offensive. I value nationalities and they still play a major role in the world, even if some people like to thing it is cooler to treat them as anachromisms. They are not anachronisms. CalJW 05:51, 2 March 2006 (UTC)

Category sorting finer specifications

I added some additional specifications to Wikipedia:Categorization#Category sorting. Just mentioning here that most of that was resulting from a discussion here: Wikipedia talk:Naming conventions (people)#Naming convention for Dutchmen --Francis Schonken 20:02, 12 February 2006 (UTC)

Something for Category_talk:Peers#Sorting, Spanish, German, Chinese, Japanese, Korean personal names might be included as well. - User:Docu

I added something for Chinese names. -- User:Docu

Category_talk:Peers#Sorting is added as well. Talk:Iberian naming customs#Which_way_in_the_telephone_directory? might be used for Spanish names. -- User:Docu

Category talk:Peers isn't relevant for sorting in any other categories. Gene Nygaard 21:12, 3 May 2006 (UTC)

Why do you think that is so? The sorting as described on Category_talk:Peers#Sorting is the way they are sorted in other categories as well: e.g. birth/death categories. Do you have an other suggestion? None made any before, so I went ahead to add it. BTW what do you think of Talk:Iberian naming customs#Which_way_in_the_telephone_directory?. -- User:Docu

'Cycles should usually be avoided'

...but why?

Is there any reason why Category:Bill Drummond and Category:Jimmy Cauty shouldn't be both members and subcategories of Category:The KLF? Isn't this is a logical arrangement for the members of music groups who also have profiles as individuals? --kingboyk 18:33, 1 March 2006 (UTC)

A cycle occur when a category is both a parent and a child of another category. Cycles have nothing to do with the article space. Circeus 18:44, 1 March 2006 (UTC)

Let's approach this from another angle: could you please have a look at Category:The KLF and tell me not just if anything is wrong, but if it's wrong why it's wrong? It would be much appreciated. --kingboyk 18:49, 1 March 2006 (UTC)

I think it's a bit confusing for there to be both a subcategory and an article for the same person on the same category page. A naive user might not know the difference and the article can be accessed by going to the category first. On the other hand, it's not bad enough that I'd go in and edit it to remove the duplication. --—The preceding unsigned comment was added by JeffW (talk • contribs) .

There is actually a value to doing it this way if it is done correctly. For example in Category:Presidents of the United States there is Category:George W. Bush and the article George W. Bush, by having both, it communicates that there is an article about Bush and a category where articles related to Bush can be found. This is even more important for a category like Category:Bridges. In this case Suspension bridge is an article about a specific type of bridge. Category:Suspension bridges is full of all the articles about the many individual suspension bridges. Categories and articles are very different things. I find that often when both the categories and eponymous articles are both listed, it is usually the subcategory that is less appropriate to the category than the article. The precedent in the past was to have the category and not the article. But people's natural inclination is to put the article in the categories it belongs in (and rightly so). I think we are in a transition stage right now, and I suspect in the future, it will be more common to see the article and not the subcategory listed. These decisions, should be made by examining the parent category, and looking to see what the logic is that determines what belongs and what does not. If the sibling articles are there, the article belongs. If the sibling categories belong, the subcategory belongs. -- Samuel Wantman 21:06, 9 March 2006 (UTC)

Another example

I still have not seen a justification either. I'd love to see something besides, "Don't do it." Here are some excerpts from Category talk:Education: Categorization with respect to Science. and Category talk:Education#The classification of education and loops. Rfrisbie^talk 19:42, 1 March 2006 (UTC)

Categorization with respect to Science.

I was asked about the removal of Education from Category:Applied sciences and will respond here. My intent is to eliminate cycles from among the categories. Perhaps there is a better way to do this than my reversion, so I'll lay out the problem.

These are the portions of the Science tree and Education tree that apply:

Category:Applied sciences -> Category:Science ->

Category:Academic disciplines -> Category:Academia -> Category:Education

Note that "Science" links to "Acadmic disciplines." The goal is to connect these together in a way that doesn't cause a cycle. I removed, what in this diagram would be an arrow from "Education" back to "Applied sciences".

One must decide which of these categories is the most general, and thus not a subcat of any of the others. To me the candidates are "Science," "Academia" or "Education" (the most general in the current configuration). The one determined to be most general won't link back to the others.

I hope this makes some sort of sense. JonHarder 03:28, 28 February 2006 (UTC)

Your objective is clear (eliminate loops) but not your (or anyone else's) purpose (justification). It's stated in Cycles should usually be avoided:

Although the MediaWiki software does not prevent cycles (loops), these should usually be avoided. Here is an example existing in early November 2005: Category:Academic disciplines - Category:Interdisciplinary fields - Category:Social sciences - Category:Education - Category:Academic disciplines - Category:Interdisciplinary fields ....

No rational is given for not using the Mediawiki software's capability to use loops. In the real world, these also are known as self-referencing systems, in this case, education about education. Other examples include meta-evaluation (evaluating evaluation), meta-ethics (the ethics of ethics), and the history of history. The unfortunate and hopefully unintended consequence of this "no loops at all costs" approach is that the classification of education has been reduced to... Fundamental: Systems: Society: Culture: Education; and Fundamental: Systems: Education. Not only has education been removed from Applied sciences, It's been removed from Social sciences and Interdisciplinary fields as well. In effect, the message is "education is not science." Removing Academia from Education simply shifts the problem without addressing it. We could just as easily "solve" the loops problem by eliminating Science from Academic disciplines. From the Education classification perspective, that would be less disruptive.

My main point is that eliminating loops in and of itself does not address the classification issue for Education. Clearly, leaving it out of the sphere of any type of science, as well as not acknowledging its self-referencing properties, are problematic. I hope we can keep working toward a solution that addressed all of our key issues. :-) Rfrisbie^talk 16:31, 28 February 2006 (UTC)

The classification of education and loops

The current classification of education is as follows:

Classification: Fundamental: Systems: Society: Culture: Education also: Fundamental: Systems: Education

Apparently, this is the result of an effort to eliminate category "loops." See the above discussion. An unfortunate side-effect of this effort is that education has been eliminated from categories such as category:Applied sciences and category:Social sciences. This directly contradicts education's place in the articles, List of academic disciplines and Social sciences. I'm going to place the following comment on the category page as a compromise, at least until something better comes along. Rfrisbie^talk 17:26, 28 February 2006 (UTC)

To avoid loops, the following classifications are not included in the education category structure.

Classification: Academic disciplines: Applied sciences: Education also: Academic disciplines: Social sciences: Education See also: List of academic disciplines, Social sciences

You're starting to confuse me now people. If someone can come up with a "why" in plain English please buzz me :-) --kingboyk 19:52, 1 March 2006 (UTC)

Kingboyk, I'm asking "why" too. I'm not trying to answer your question. :-) Rfrisbie^talk 20:17, 1 March 2006 (UTC)

Yes, I had worked that much out :) --kingboyk 20:22, 1 March 2006 (UTC)

I think it has to do with the fact the category structure is expected to be roughly semi-hierarchical:

categories do not form a strict hierarchy or tree structure, but a more general directed acyclic graph (or close to it, see below). (under "categories do not form a tree")

Generally, people can expect that the entire thing is organized roughly in a tree-like structure, and will be thoroughfully confused when encountering cycles. I know I was. Circeus 20:25, 1 March 2006 (UTC)

I don't see why. Visited links usually go a different colour. I was doing it this way because I felt it made navigation easier for newbies - there's no doubt to a casual surfer now that The KLF and Bill Drummond are interconnected and if you want to read about one you should probably read about the other! --kingboyk 20:31, 1 March 2006 (UTC)

I agree this is a "bad" rule. "Self-referencing systems" exist, such as in all the meta- fields. By definition, such systems create "cycles" or "loops." Rfrisbie^talk 20:50, 1 March 2006 (UTC)

I find category cycles are mildly confusing or surprising when I encounter them, but we should definitely use a little sense and not eliminate loops when the costs are higher than the benefits. I doubt anyone goes thru the cycle more than twice before they figure it out. NickelShoe 00:53, 2 March 2006 (UTC)

There are several reasons why I prefer to keep categorization tree-like, in addition to the confusion of cycles mentioned above:

Logically a category is more general than its subcategories. A cycle implies a category is both more general and less general than itself, or perhaps that every member of the cycle is of equal status.
A tree structure makes it possible, in the future, to be able to request every article on a certain topic and have software automatically build a collection of all the related articles, by traversing the category tree, without wandering into unexpected directions.
A disciplined approach to categorization helps the subcategorization process by identifying areas that relate to other trees. For example, it might be possible to link a subcategory to another field and not cause cycle, where linking the parent category would create a loop. Another way to say this is that even if a category placement causes a cycle, working instead with one of it's subcategories may not. And this in fact may turn out to be a better, more specific way to group things.

When categorizing it is helpful to consider whether X is a kind of Y or if X is a part of Y. If so, then X is a subcategory of Y. JonHarder 01:03, 3 March 2006 (UTC)

Indeed, but again I can't see how that applies to the pop group/individual person situation. I'm probably just being thick, granted :-)

I think I'll leave Category:The KLF how it is for now. If you can tweak the category membership of related pages and categories to make it better then please do so. Otherwise, I shall quietly WP:IAR as I'm not convinced that the current arrangement is 'bad' any way.

It might be that the advice needs to be softened too? --kingboyk 01:10, 3 March 2006 (UTC)

Those all are fine points above for reasons to avoid loops. So show me how to resolve this.

Classification: Education: Academia: Academic disciplines: Social sciences: Education:

Rfrisbie^talk 03:12, 3 March 2006 (UTC)

I'd like to take a step back for a second and look at the bigger picture. Wikipedia categories are NOT a pure classification system. Articles are in categories for many reasons. Sometimes they are just related to the subject of the category (for example Category:George W. Bush), sometimes they are members of the set defined by the category (example Category:Suspension bridges). Sometimes the category is very broad like Category:Education. The reason that there is a loop in the education example is because Academia is RELATED to Education, but Academic disciplines are a SUBSET of Academia. This sort of thing is to be expected. This is a wiki. The work that it would take to create a monolithic classification system that everyone could agree to would be huge and probably so frustrating that we'd all leave the project. The beauty of Wikipedia's categories is that multiple intersecting classifications can co-exist. The problems arise when all these hierarchies intersect. So sometimes cycles happen because things have not been thought out very well, but sometimes they just happen.

BTW, I am the original author of Wikipedia:Classification. My intent, in creating the page, was to create a way of showing the multiple classifications systems that exist here. For example, there is a hierarchy under Category:Bridges and another under Category:Toll bridges. By adding a classification, you can illustrate that these are two separate hierarchies that overlap each other. -- Samuel Wantman 09:55, 3 March 2006 (UTC)

Hi Samuel, thanks for your comments. Your "RELATED" argument makes sense IF it's TRUE. Not everyone, including me, will agree with you on that. :-) Some will say Academia IS a subset of Education. Be that as it may. I expect there must be a few cases SOMEWHERE that you would agree are self-referencing systems. The Meta- article comes to mind for some possible examples. Staying at the bigger picture, this page states that "Cycles should usually be avoided," yet nobody who strongly advocates removing them ever seems to acknowledge an instance where they should be maintained. They're only considered an inconvenience from sloppy classifications by others. So, can you give me ANY example(s) of what you would consider to be ACCEPTABLE category cycles? If so, they should be added to the page. If not, the section should be edited to say "Cycles should be avoided" (period). Thanks for working with me on this. Rfrisbie^talk 14:32, 3 March 2006 (UTC)

Perhaps "related" is not the correct word, but here is what I mean: Academia is a place where Education happens. If you move to the other end of the taxonomy, education is the thing that is being studied and taught in academia. So education is a subject of study and also is the process of learning and teaching. Since education has this broad meaning the cycle happens. I find this to be an acceptable cycle. Education does have this self-referential nature. But my bigger point is that this "sloppy"classification system is by design. It allows people to browse in broadly defined categories rather than in narrow ones. I suspect you could create categorization schemes that would have no cycles, but why bother? Categorization should consider the utility of the category for a user who wants to browse. It is not intended to be collaborative work on a perfect taxonomy of all knowledge, and I think that is a good thing. -- Samuel Wantman 07:33, 6 March 2006 (UTC)

Thanks Samuel Wantman. So, given everything about everything, how would you classify and categorize Category:Education and the related subcategories the create the loops? Rfrisbie^talk 15:14, 6 March 2006 (UTC)

I would just leave the loops. I find the comments about the loops on the category pages to be confusing and jarring. The only way remove the loops is to try and separate the different meanings of education into separate categories, and this would make the categories much less useful for browsing. I don't think loops upset many casual users of Wikipedia. They are probably most upsetting to editors who think that categories are a rigid classification system. My experience is that if you follow that view you will be extremely frustrated. It also leads to long unproductive battles over category membership. Restore the categoriztion loops and remove the comments about loops. Think of categories as a collection of articles related to the topic and try to avoid thinking of it as a pure taxonomy. -- Samuel Wantman 20:43, 6 March 2006 (UTC)

Okay, thanks. Perhaps JonHarder would like to comment on your suggestion as well. Also, would you be willing to consider a rewrite of Cycles should usually be avoided? That section doesn't really reflect the conversation here and when cycles might be "okay." Thanks. :-) Rfrisbie^talk 21:49, 6 March 2006 (UTC)

I think everyone in this conversation understands the issues here and the difficulties organizing in a way that is both fucntional and avoids cycles. It is disappointing when one can't find a solution that is elegant while also eliminating the cycles. Unless new cycles have been introduced in the past week, this will be the only cycle in the main article space (along with the trivial KLF cycle mentioned above). I don't see a need to make any significant change to the main article; "usually should be avoided" works for me! JonHarder 00:18, 8 March 2006 (UTC)

I've restored the category loops and removed the comments. -- Samuel Wantman 07:05, 8 March 2006 (UTC)

I think the "education" example is an instance of legitimate self-inclusion, but the KLF one (which I removed a couple of days ago without being aware of this discussion) is basically using "categories as links", on which question I have to disagree with Samuel on. I think the reason to be careful about such things is to preserve some sort of meaning to membership in a part of the hierarchy. Inevitably this breaks down after a certain point -- I noticed that Category:Geography of Greenland is at the base of a vertical chain of inclusion 37 (!) deep, featuring among others, it being classified (sorry) under Category:Philosophical concepts, Category:Apes, and Category:Theoretical physics. Clearly, the category system does not at present lend itself to questions such as "what are the articles on theoretical physics"; it'd currently give the answer, "almost all of them". But it would be worthwhile to work towards that as a goal. Alai 00:51, 16 March 2006 (UTC)

Third party opinions requested

There's a dispute over at WikiProject Cricket over the appropriate categorization of Pakistani, Bangladeshi and UAE cricketers. Details here. Third party opinions welcome. --Muchness 14:26, 2 March 2006 (UTC)

Clarifying the section, "Cycles should usually be avoided"

It appears the section on this page, Cycles should usually be avoided is both controversial and ambiguous. While it states they should be avoided, it gives no justification. In addition, the discussion above about Education demonstrates at least one case of a self-referencing system that might be “acceptable,” at least to many editors. Perhaps the more active editors in this area that have weighed in with differing views on the subject, such as SamuelWantman and JonHarder can work together to clarify what types of cycles should be avoided under what circumstances and when it might be okay to use cycles. Using more examples in the "okay" – "not okay" situations would be very helpful as well. Thanks. Rfrisbie^talk 18:48, 8 March 2006 (UTC)

While I am honored to be nominated for a task, you seem to be more concerned about this issue than either I and JoHarder are. There is no reason why you could not come up with some wording and add it to the page. Perhaps a better way to deal with this issue is to take the discussion above and edit into a short sub-page that can be linked to the Cycles should usually be avoided section. You could call it Wikipedia:Categorization/Cycles. A sub-page might be the way to go because in the two years I've been watching the conversations here, this has rarely come up. There was a good deal of discussion about the categorization of John Lennon, which you could also reference as an example. -- Samuel Wantman 08:43, 9 March 2006 (UTC)

Well, I haven't heard from him since you made the change to the category pages, but I got the impression JonHarder thought it was important, at least enough to remove the cycles in the first place. ;-) If nobody else is all that interested in it, I might try something in a little while. I wasn't the one who claimed there was a problem with cycles in the first place, so I still would like to see a consensus from both perspectives in the revision. Rfrisbie^talk 02:12, 11 March 2006 (UTC)

I added something. I'm not going to add a subpage, at least for now. I also didn't add the John Lennon example because it looked to me more like the George W. Bush example used for "THE TOPIC ARTICLE RULE" at Reasons for duplication. Rfrisbie^talk 17:08, 11 March 2006 (UTC)

Thanks Rfrisbie, that's a lot better. I knew there was an example of a useful cycle, I just couldn't think of one. ··gracefool |☺ 12:14, 15 March 2006 (UTC)

CfD reform -- Criteria for deletion based on precedent

I've started a discussion about some reforms for the Categories for Deletion page (WP:CFD). Please take a look here. Thanks. -- Samuel Wantman 10:32, 9 March 2006 (UTC)

Non-lists in Category:Episode lists

There are a whole bunch of sub-categories of this category that are filled not with lists, but with individual episodes of television shows. I perhaps too hastily started removing the Episode lists category from those pages and in several cases it was the only category so I created categories with no categorization. After getting a complaint I stopped, but the question remains what to do with these pages that don't belong.

Besides being illogical, the way it was was confusing because the same show could appear in the Category section and the Pages section so you'd have to look in both places to make sure you didn't miss the show you were looking for. To see how it was, look at the second page (the next 200).

A lot of those shows had their own Category page, like Category:Friends in which case Category:Friends episodes could be a sub-category of that and I don't see the need any other categories. So we could create show-specific categories for the pages that were only sub-categories of Category:Episode lists. Or they could be categorized under Category:CBS network shows or something like that.

But I think a better solution would be to create Category:Episodes by television show which would fit nicely under Category:Television series. Comments? --JeffW 20:44, 10 March 2006 (UTC)

I followed you here from your edit of Category:My Name Is Earl episodes. I created it under Category:Episode lists following the example set by the Simpsons and others. Since category pages display lists, I think you may be taking the name of the "Episode lists" category too literally by kicking out episode subcats, but I don't object to your proposed fixes. A "Category:Episodes by television show" would be more accurate if named "Category:Episodes by television series". —RandallJones 21:34, 10 March 2006 (UTC)

I see your point about Categories being a kind of list, but if you took that thinking to its logical extreme then every Category should go under the Lists category, but that's clearly not what was intended for it. --JeffW 21:46, 10 March 2006 (UTC)

No argument there. —RandallJones 20:23, 11 March 2006 (UTC)

Another reason that these categories are confusing is that there are over 200 pages in Category:Episode lists and when these are sub-categorized, they will then be intermingling with all the Foo episodes categories making them harder to locate. (BTW, does anyone else think that Episode lists should be changed to something like Television series episode lists?). --JeffW 00:13, 11 March 2006 (UTC)

Since it has only television as supercats, yes, it would help avoid mistakes. Our Gang filmography needs to be moved, and the others checked to assure there's nothing from Category: Radio programs. —RandallJones 20:23, 11 March 2006 (UTC)

I just added the cfr to rename Category:Episode lists to Category:Lists of television series episodes. I'll try to find a place for Our Gang and check that there are no other non-television series. --JeffW 19:28, 14 March 2006 (UTC)

I think enough time has passed for discussion and I haven't received any replies requesting that I not do this, so I'm going to create Category:Episodes by television series putting it under Category:Television series and start moving the episode categories there. --JeffW 16:36, 13 March 2006 (UTC)

I accidentally created it as "televion show" instead of "television series". Is it worth doing a CfR? --JeffW 17:40, 13 March 2006 (UTC)

I deleted it (supposing that was what you wanted). BTW good idea to create Category:Episodes by television series, I'm glad these subcategories of Category:Episode lists can now go elsewhere. -- User:Docu

Thanks. I've finished adding the new category to the pages that I had previously stripped of the old category. --JeffW 21:44, 13 March 2006 (UTC)

Sub-categorizing the Lists Category

I recently added and rearranged the sub-categories in the Lists category so there would be categories for the subjects Art, Culture, Geography, History, Mathematics, People, Philosophy, Religion, Science, and Technology. This seems to be working pretty well, but these categories are mixed in with Abbreviations, Books, by Country, by Form (timeline etc.), Reference material, Worst lists, and Year lists which, I think, dilutes their usefulness. I tried sorting the subject cats to the beginning using space followed by the category name, but they came out in two columns. Then someone who was just "tidying up" removed the spaces putting it back the way it was.

Is there another way to do this? Does anyone else feel that it would be a good idea to ask for something like [[Category:Lists|Art|Subjects]] from the programming crew? (with the second piping meaning to place the link to the sub-category under the heading "Subjects", instead of the default letter headings).

In the meantime would naming the categories something like "Subject: Art lists", "Subject: Culture lists", etc. be reasonable? --JeffW 08:57, 15 March 2006 (UTC)

See Category_talk:Lists#Reorganizing_Subcats_Based_on_Standard_.28topical.29_Schemes_.3F. -- User:Docu

Are you referring to your cat scan tool? If so, I don't see how it helps in figuring out how to sub-categorize and format a category in a way to make navigation easy for a normal user, who presumably won't be using cat scan. --JeffW 19:52, 15 March 2006 (UTC)

Divider between categories in category listing at bottom of a page

Not sure if this is the best place to ask – please redirect me if necessary – but would anyone else prefer the "middle dot" rather than the "vertical line" as the divider between category names in the list at the bottom of a page, i.e.:

Categories: This category · That category · Another category · etc.
rather than
Categories: This category | That category | Another category | etc...?

When glancing, I find it slightly easier to distinguish the categories listed using the dot rather than the line. Thanks for your thoughts, David Kernow 05:13, 15 March 2006 (UTC)

I don't know if they're easier to distinguish, but they look cooler. --JeffW 19:12, 15 March 2006 (UTC)

"Related categories"?

I know we have examples of "related categories" being created manually, such as WTC-9/11, but is there any serious thought to making them an automated feature? That sure would make it a lot easier to create meaningful broader/narrower term hiercharies and avoid those pesky loops! >;-o Rfrisbie^talk 03:17, 16 March 2006 (UTC)

Automation is in, as part of the MediaWiki software, or, via by bot, etc? I'm not sure I immediately see how this might be done. Alai 13:40, 16 March 2006 (UTC)

I wasn't very clear, sorry. Say I type in "[[Relatedcategory:Foo]]" on a category page. Then that shows up in its own section at the top of the page, just under the subcategories section. That way, "See also" comments can be avoided, subcategories are more likely to be true subsets, while related categories can be overlapping sets. Gyrations around loops should be much simpler too. Rfrisbie^talk 20:56, 16 March 2006 (UTC)

This suggestion has been floating around for a while. I think I made a similar suggestion about a year ago on this page. I was thinking of using a related category classsification to handle the situation when one hierarcy of categories is a subset of another hierarchy, such as Category:Bridges in the United States and Category:Toll bridges in the United States. But after more than a year of thinking about this issue, I wonder if having a section of related categories is any better than what we are doing now. I wonder if it would be any clearer and easily understood. Having a comment that says "See also" seems to be a clear way of saying "This doesn't belong in this category but you might be looking for it here". Trying to create "true subsets" might be a path that creates more editing conflicts. Part of the beauty of the categorization scheme is that it is not a rigid taxonomy. -- Samuel Wantman 22:18, 16 March 2006 (UTC)

I agree suggesting "related categories" isn't a new idea. They're a staple of any good taxonomy. I also agree the "See also" kludge is conceptually the same. And I agree a network is more useful than a hierarchy. On the other hand, a "manual" approach loses any meaningful capability for systematic database management of these category relationships. I fully expect that the easier they are to create and view in a consistent way, the more they will be used. In my opinion, that's the tipper for doing it through a related categories code. Rfrisbie^talk 22:36, 16 March 2006 (UTC)

Top-Sorting

I've noticed that it has become common practice to sort the main article of a category under * so it will appear at the top of the page list. How does this interact with the catmore template? Is one preferred or should both be used?

Also, in many cases it's appropriate to sort a list (especially when the all the pages in a category are instances) to the top of the page list. If there is already a main article for that category, my gut says lists should appear after the main article. Should they be sorted as ** so they appear after the main article but in the same block? Or perhaps + should be used so they appear as a second block. Or should the main article be sorted with a space instead of * and the * used for the the lists? --JeffW 19:07, 16 March 2006 (UTC)

This is currently being done several different ways. The guidelines suggest the space for main articles, but are not clear about the star. I think that eponymous articles should be sorted with a space and all others should use the *. --Samuel Wantman 22:27, 16 March 2006 (UTC)

I've noticed this very "diversity" in the sorting of sub-categories of stub types, which possibly most often use a * as a prefix, sometimes a space, and sometimes no key prefix at all (or indeed, no key). In such cases I think it makes sense to use something, as often these are largeish categories with multiple "pages", and it's useful if all the subcats appear on the first one. Some standardisation would probably be nice (said more in faint hope than strong expectation). Alai 01:21, 18 March 2006 (UTC)

Subcategory intro wording

I recently noticed awkwardness in the wording of the automatically-generated text that precedes a category’s subcategories. The awkward “There are 62 subcategories to this category.” would be better worded as “There are 62 subcategories within this category.” —optikos 00:47, 20 March 2006 (UTC)

Is it appropriate to try regulate guideline/policy via help: namespace?

Eponymous categories

There are many occasions when having articles and categories with the same name is not a good idea. Eponymous categories should be avoided unless there are more than a dozen or so related articles. If all the articles are cross-linked to the main article in a prominent way there is no need for the category. Another very important consideration is the effect of having eponymous categories for all the siblings of the article. For example, curently RENT has the eponymous Category:RENT (nominated by me at WP:CFD). If every broadway play had an eponymous category it would fill up Category:Broadway musicals with duplicate entries in the subcategory and article sections with little or no added utility. If Category:Celebrities appearing in South Park (also at CFD) exists, then every TV series could have similar eponymous categories. This would lead to dozens of categories of little value being added to the articles of hundreds of celebrities. I think the standard should be: "Eponymous categories should be avoided unless there are more than a dozen links in the See also section." I'd like to add a guideline about this. Any comments? -- Samuel Wantman 08:21, 30 March 2006 (UTC)

Categories are simply sets of related articles and subcategories belonging to one or more parent categories. They have a short intro ideally explaining what's in and what's out, and that's almost all you can say about categories. Make sure that they make sense for a long period, moving all elements of a set to a renamed category could be a PITA. Unless the category is mainly filled by un-subst'ed template(s), then modifying the template(s) would do the trick. Of course some categories are bogus, like some articles, templates, lists, guidelines, or redirects. In other words, I didn't get your point. But I dissolved a bogus category with only two articles as child of another category with only two articles some hours ago, if that's related to your comment (?). -- Omniplex 10:14, 30 March 2006 (UTC)

My point is not about how a category is named, but whether it should exist in any form. It is you last example, a bogus category with only a handful of articles, all crosslinked and with the potential of being copied thousands of time to similar small collections of articles (like Broadway plays). -- Samuel Wantman 21:44, 30 March 2006 (UTC)

I do admit it seems a little odd when the same name appears in the subcategory and the article section of the category page. Couldn't we deal with that problem with just making it a policy that when a category and an article share the same or nearly the same name that they shouldn't both be listed in the same categories? --JeffW 06:03, 2 April 2006 (UTC)

I agree in principle, but I'm not sure about having a rule. The guidelines state that each should be categorized by their own merits in categories with members who are siblings. Sometimes there are cases where the eponymous category differs slightly from the article (usually with the letter "s") such as Suspension bridge and Category:Suspension bridges. In these cases it makes sense to have both categorized in Category:Bridges. If you look at the categories found in Category:Category by person, most of these belong in very few categories, while the eponymous article belongs in many categories. This leads to having a structure that would bring the user to an article from many categories, and then from the category, they would procede to the eponymous category. This seems like a natural flow that I think would be just fine in most situations. I can imagine that there are eponymous categories that might be suitable for inclusion as a subcategory in other categories, so I don't think this should be a hard and fast rule. Common practice nowadays is to put the eponymous categories into way too many other categories. I think this is a result of the old system of JUST putting the article in the eponymous category and no where else. --Samuel Wantman 10:50, 2 April 2006 (UTC)

I've been seeing a situation while working with the Lists category that bears on this. When someone decides to break a list out of an article into their own article the list often doesn't doesn't really fit into any of the parent categories. Now, it won't be a category orphan because it will fit into one of the Lists subcategories, but it seems to me that every list ought to fit into the non-list category bush somehow. So it would be nice to create an eponymous category so the list has a category to belong to, but I think most people would object to a category with only two items. I don't have any reasonable solutions, but its something to think about. --JeffW 06:10, 2 April 2006 (UTC)

Can you give an example? Could the list be in one of the parent categories of the (unwanted) eponymous category? Is there a possible new subcategory that could contain this list and similar lists related to the siblings of the topic? -- Samuel Wantman 10:50, 2 April 2006 (UTC)

How about Timeline of Northfield history. It wouldn't really fit in Category:Cities in Minnesota and it doesn't really belong in Category:History by city where it is now. --JeffW 14:58, 2 April 2006 (UTC)

Or take List of teams sponsored by Adidas. It would fit great in an Adidas category if there was one. But there isn't and it doesn't really fit in any of the categories that Adidas is in, so I just put it in the Sports category and a Lists category. --JeffW 13:27, 4 April 2006 (UTC)

Redirect

Should WP:category redirect here? Afonso Silva 20:08, 30 March 2006 (UTC)

WP:CAT goes to Wikipedia:Category with links to Wikipedia:Categorization. Often I found the most reliable technical info on Meta. -- Omniplex 02:12, 3 April 2006 (UTC)

Oh, sorry, I forgot to capitalize the "C", so, I thought there was no redirect, that seemed really strange, so, I asked this. Just forget it. And thanks for the meta tip! Afonso Silva 21:59, 13 May 2006 (UTC)

Advice on Categorization?

It's been a while (September 2005) since I worked on Categories, and I was wondering how much things have moved on since then? Is there any summary available of the discussion that hasn't made it onto the guidelines page yet? I also had some questions that I was hoping to get some advice on:

(1) I recently worked on a page that seems to touch a lot of different areas: Ptolemy (name) - it originated from a disambiguation page (Ptolemy (disambiguation), and I wasn't sure whether to leave the disambiguation tags and categories in place. Eventually I did remove them and added some different categories (Etymology, Names, Ancient Greece, Ancient Egypt, Cultural history) - can anyone think of any more, or more suitable, categories?

(2) Would I be right to say that the categorisation scheme is flexible enough to allow slightly different methods to apply in different subject areas? I wrote something here about categorisation in one particular subject area, and was wondering if it sounds OK in terms of the overall categorisation guidelines?

At the same location, I've been considering using the Category Scan tool to get an overall picture of the categorisation scheme for a particular, fairly small, area. Is there an easy way to use the Category Scan tool, or something similar, to generate a "tree-like" snapshot as something that could be posted on a Wikipedia page, either as a text-based "tree", or a screenshot? Carcharoth 19:58, 2 April 2006 (UTC)

Some more thoughts on Categorization

I've been reading through this talk page, and I came across the following comment: "I noticed that Category:Geography of Greenland is at the base of a vertical chain of inclusion 37 (!) deep, featuring among others, it being classified (sorry) under Cat:Philosophical concepts, Cat:Apes, and Cat:Theoretical physics."

I thought I'd follow this up and see why this was the case, trying to find the routes from those three categories down to the Geography of Greenland category. I swiftly gave up, or assumed that the connections had been severed, but I did come across one thing I found strange.

Why is Category:Geography a subcategory of Category:Earth sciences? This seems to me to be a clear case where some of the articles and subcategories of Geography belong in Earth Sciences, but some don't (eg. Political Geography). Lumping all the Geography cat/articles under Earth Sciences has created false links on the "tree". This is a clear example of bad (or lazy) categorisation in a high level category.

Maybe a guideline needs to be written (unless it already exists) that when considering what parent categories to put a category in, consider whether the contents of the category could also be split among different parent categories, either with or without creating new categories to make the categorisation easier.

In the Geography example, Geography would be removed from the Earth sciences category and all the 'science' geography items (for example) would be placed in the Earth Sciences category as well as the Geography category, but the 'non-science' geography items would be placed in the Geography but not the Earth Sciences category. Thus the Geography category, while not itself being a member of the Earth Sciences category, would have children that were members.

Overlapping sets and links skipping levels on the "tree", or reaching over to a distant part of the "tree". This is how I see the Categorisation system, and some of the subtleties of categorisation. But I thought it best to see if I'm getting this right before I start being bold and changing things where I think the categorisation is not quite right. I'm also still looking for a way to show other people small areas of the expanded category "tree". I'm only aware of the Category Scan tool, but how do I get the output from that into a form that is easily posted on a talk page? Carcharoth 01:35, 3 April 2006 (UTC)

Differentiate between categorizing articles and categorizing categories

Something I've realised recently is that it is probably useful to differentiate between categorizing articles and categorizing categories. The guidelines here almost exclusively talk about how to categorize articles. I feel that something about when and how to categorise a category would be a useful addition to the guidelines.

I see "categorizing categories" as creating links on the category "tree", making links between different areas, as opposed to populating the tree (adding articles to categories). Before doing this, it is important to realise the nature of the category and where you are in the category tree: ie. how many levels downwards does the category reach that you are trying to "graft" onto the "tree", and how many articles are present inside the category and its subcategories? If the number is small and in a tightly focused area, then you may have a pretty good idea of suitable categories that should be linked to. On the other hand, if the subject is broad (like the Geography category) then a simnple graft is not the right way to proceed (you have to surgically tweak out branches from the geography category and graft them onto separate areas of the "tree").

Another way to look at categorising categories might be the following. When considering what categories to put an article in, consider whether, instead of putting the article in a candidate category, you can link the article to that category by putting one of the article's categories (parent or other ancestor) in the candidate category. The link, and browsing route, is thus created without having an excessive amount of similar categories for that article. In other words: categorize an article on where you want it "listed" with similar articles, but rely on the process of 'categorising the categories' to create the "browsing links" that allow people to browse the whole structure to find a particular category.

Would something like this be useful for the guidelines? Carcharoth 02:03, 3 April 2006 (UTC)

Extremely long category titles with arcane industry codes

A new editor is trying to build categories that mimic the industry codes in an economics journal. When I objected, the editor stated he is willing to accept suggestions, but as I have not worked either in his field, or with the categorization project much, I could use some assistance in helping him to achieve his goals. Here is an example of the type of category he is creating: Category:Business Administration and Business Economics; Marketing; Accounting JEL:M See Talk:JEL classification codes. --Blainster 06:52, 3 April 2006 (UTC)

I'm the editor in question. Following the advice of Blainster, I've come here to see whether people think the category structure I'm proposing is a good idea, and if so how I can improve its implementation. The background is that the Category:Economics structure has been a mess for a long time, as you can see from the talk page. Before I started work it had about 50 top-level subcategories, but still went nowhere near covering the field. So I suggested on the talk page that we should follow the classification system from the Journal of Economic Literature which is standard within the economics profession. It has about 20 top-level categories, and then two levels of subcategories. I got a limited but generally positive response, checked with the American Economic Association that they were OK with the use of their system (subject to the use of the form JEL:X) and went ahead. There were some problems with this. In particular, while most of the categories are fairly sensible like Category:Public Economics JEL:H, there are a couple of monsters like the one Blainster cites.

I think use of a sensible classification system, the product of a lot of thought, has great potential benefits for the Economics section of Wikipedia. In particular, a look at the categories and subcategories reveals lots of gaps that ought to be filled, even bearing in mind that we're aiming at a general audience and not trying to duplicate Palgrave (the main econ dictionary/encyclopedia). To take just one glaring example, we don't have an article on Production economics, though there are a number on topics that appear superficially related.

I think the main problem is with the use of the industry codes in the category names, and the fact that they will distract the majority of people. Also, some of the "names" are just listings of the names of what would effectively be subcategories. I suggest that the actual names used in the category tags (ie. the names that appear at the bottom of the article page and on category pages) are simplified versions that don't have the code, but still communicate the essence of the system. To satisfy the request of the AEA to attribute as JEL:10 etc, maybe you should ask them if it is acceptable to just put the full name and code on the blurb at the top of the category page, and saying that the system is "based" upon theirs? Though I suspect that wouldn't be acceptable to them - they do realise that Wikipedia is likely to mangle and merge their system with ours and others, right?

Anyway, if modified names and full attribution on the category page is acceptable to them and us, you could then post a list of, or link to, the AEA/JEL names and codes (on the talk page concerned), and a consensus could be reached over what we should call these categories. Though having looked at the JEL system here, I see that it is rather a long list...

On reflection, my feeling is that using a system rigidly is not flexible enough. Borrowing ideas from it is great, but I think Wikipedia should organically develop its own categorisation systems while using ideas from many other systems. Looking at their system to identify gaps in Wikipedia's coverage is a great idea, for example. Carcharoth 10:27, 3 April 2006 (UTC)

There are several issues. Using the AEA system is a possiblility, but the codes should not be part of the category name. If the category name is the same as a AEA classification, it would be ok to mention that in the text of the category, but I suspect that even this is not the best way to handle it. Carcharoth is correct, Wikipedia's classification scheme developes organically, and it is unrealistic to expect everyone to adhere to a rigid classification system. It would be much better to have an list that had all the classifications and linked to the appropriate articles. This would be much more manageable, and useful to those who are looking for it. Categorization and classification are not the same thing. Trying to make them the same only leads to frustration and conflict. -- Samuel Wantman 11:11, 3 April 2006 (UTC)

It seems consistent with my interpretation of AEA's views that we could leave the codes out of the category names, and include the code in the category pages. JQ 11:52, 3 April 2006 (UTC)

Don't agree, see below: wikipedians objected to using SOC tags and codes in category descriptions, while that would make the system rigid and singleminded; also there's no reason why wikipedia should have a category named Category:Business Administration and Business Economics; Marketing: such name is at least too long, not conforming to general wikipedia:naming conventions (categories) ("Short [...] names are preferred for categories."); and probably over-all not suitable for wikipedia (wouldn't survive a CfD I presume). --Francis Schonken 14:17, 3 April 2006 (UTC)

More than a year ago there was some discussion whether SOC should be adopted as a wikipedia categorization scheme. Even DDC was proposed at some point. Some remnants of such discussions can be found here: Wikipedia_talk:Category_schemes#Preliminaries_and_related_topics, here: Wikipedia:Categories_for_deletion/unresolved#Non-wikipedia_classification_systems, and here: Category talk:Occupations... After these discussions had died out, I didn't think anyone was still going to try a "non-wikipedia classification system"... JEL doesn't look any better than those other abandoned ideas afaik. --Francis Schonken 11:35, 3 April 2006 (UTC)

PS: see also Wikipedia talk:Dewey Decimal System#Categories - all redlinks, no? --Francis Schonken 11:42, 3 April 2006 (UTC)

On this point, it appears to me that the outcome of the discussion over SOC has been that this scheme is in fact used as the basis of Wikipedia categories, and that attempts to change this were unsuccessful. JQ 11:52, 3 April 2006 (UTC)

True, I think SOC was pretty much the foundational system of category:occupations subcategories, but not in a rigid way, and I have no idea how compatible wikipedia's current categorization scheme for occupations would still be with SOC, anyway the reference to SOC was removed from the "occupations" category definition a long time ago, as a principle also while US governement-centred; JEL (industry-centred) would be as singleminded I suppose: any such system, if adapted in wikipedia, ends up to be a "mixed" system in the end unavoidably, so specifically the external categorization scheme's tags and codes are avoided. --Francis Schonken 12:03, 3 April 2006 (UTC)

I think a reasonable starting point would be to remove the JEL codes from the category names and also add (Economics) where this is not already in the name. I'll add the code on the page for the category itself, and check with AEA that they have no problems with the implementation. However, there's no obvious way for a non-admin to change a category name. Can someone give me a hand on this. —This unsigned comment was added by John Quiggin (talk • contribs) .

This is now posted at WP:CFD. Discussion should continue there. -- Samuel Wantman 00:00, 4 April 2006 (UTC)

People by language

I want to create a category for sorting people as per their language knowledge. So, L.L. Zamenhof should be listed in Category:Esperanto speakers, Category:Polish speakers, etc. or Kimi Räikkönen in Category:Finnish speakers, Category:English speakers. I think this will be useful, but I want to have consensus to start this work. Best, Mxcatania 14:47, 3 April 2006 (UTC)

I'd say if they are categorised by nationality, and that is the same as the language, then that is enough (you could make "English people" a subcategory of English speakers if you wanted). If they know a language other than the one associated with their nationality, then they can be usefully added to such categories. One more thing, I would only do this if there is a reliable reference, in the article about the person, saying that the person speaks this other language. The important thing here is to decide whether the category will be _useful_. Does it duplicate too much of what the nationality categories do? Ditto for ethnicity categories (though I think there's been a lot of debate about that already). Language is not among the current reccommended criteria given here. Should it be? Carcharoth 16:05, 3 April 2006 (UTC)

As for suitable Category names, I'd suggest Category:People by language (if you want to be comprehensive), or Category:People with a second language (and obviously third and more languages). I thought there was a Category:Multilingual people category in the People category tree, but I can't find it. Maybe the link I've just written will be blue? Carcharoth 16:10, 3 April 2006 (UTC)

Found it!! :-) Category:Polyglots was hidden away in Category:Linguists. I rescued it and added it to Category:People. Anyway, the Polyglots category could be a useful starting point, or it could be that that is all that people really want from a classification system. I think there might be a demand for people who aren't polyglots, but who speak more than one language. But classifying all the people who speak one language seems a bit excessive. Carcharoth 16:22, 3 April 2006 (UTC)

Carcharoth, Thanks for your feedback with this issue. I'd like to know whether I need a consensus to do this, or not. I think Category:People by language would be the best name. The Category:Polyglots would fit as a subcategory of that one I'm proposing. Again, thank you for your words. Best, Mxcatania 17:04, 3 April 2006 (UTC)

Based on the above, I'd say create Category:People by language, move Category:Polyglots to be a subcategory of that category, and consider creating Category:Multilingual people as another subcategory. One problem is that many people outside the English-speaking countries will have English as a second language. So this might all end up a bit unwieldy. I think subdividing by language would be unhelpful, though if you went that way, it would be Category:Speakers of English as a second language or something similar. Carcharoth 16:35, 3 April 2006 (UTC)

PS. Category:Esperantists already covers speakers of Esperanto. Carcharoth 16:39, 3 April 2006 (UTC)

PPS. I rescued Category:Esperantists and put it in Category:People, so you now have two categories (that one and the Polyglots one) to start a People by languages category. Carcharoth 16:54, 3 April 2006 (UTC)

I really don't see the point. Is it noteworthy that a German speaks German? Are you commiting your time to going through all the thousands of existing People articles and deciding which language they speak? --JeffW 17:05, 3 April 2006 (UTC)

I think that people navigating from 'People' -> 'People by language' should encounter a note telling them that the category only covers people noteworthy for their uses of languages (as in the two categories mentioned above: Polyglots and Esperantists). Other subcategories can be added to this 'People by language' category as needed and as appropriate. The note should also say, up front, that people whose mother tongue is the same as their nationality are listed under their nationality. That should make the task manageable. Carcharoth 17:21, 3 April 2006 (UTC)

I think what Carcharoth said was perfect. I agree with that and, on the other hand, yes, it's challenging to tag people as per each language they know, but we can do it! It would be nice, I mean. Mxcatania 17:26, 3 April 2006 (UTC)

I'd like to know what subcategories besides the two mentioned that you are planning to add. I don't think that being bilingual is notable enough, for example. --JeffW 18:47, 3 April 2006 (UTC)

I suggest this:

For people from USA, Britain, Australia, etc, to create a category Category:English speakers so Category:American people would be subcategory of that I'm proposing.
For people from a region inside a country (e.g. Category:Catalan people) to add them to Category:People by language since they speak the region language.
For people talking an Auxiliary/Artificial language (e.g. Category:Esperantists), to add them to Category:People by language) since they speak that artificial language.
To include category Category:People by language as a subcategory of Category:People.
To find a solution for sorting people speaking a language and not being of the native region that language is from (e.g. Fernando Alonso, who speaks English and he is not from Britain).

This would be an attempt to sort people by language. If this proposal is accepted, then we don't need to categorize every article, just subcategorize some categories. Best, Mxcatania 20:28, 3 April 2006 (UTC)

I find this to be an interesting discussion. I've been wondering about language categorizations for a while, and thought I'd wait until I was clear in my own opinions before saying anything. As the conversation has started, I'll add my two cents:

It is important to identify where subcategorization by language will be a valuable addition to a category. The categories that got me thinking about this were all the literature categories. These are broken down by nationality, but it makes as much or more sense to group writers by language.
If there are multiple subcategorization methods for a category (language, nationality, sub-genre, etc...) the categories should be populated at the higher level.
Since this would be an additional level of subcategorizations for many categories, how does this level intersect with the existing categories. For example, if there is Category:English speakers how does it intersect with Category:American writers? Will there be a subcategories called Category:American writers in the English language and Category:American writers in the Spanish language?
Discussions about subcategorization methods for a topic should happen here and/or at a WikiProject, and there should be notifications on many talk pages so that the decisions have community approval. Perhaps we need a page, similar to Categories for deletion that lists proposed reorganizations. It could be called Wikipedia:Proposed category reorganizations. The process could be two stage. First stage a week(or more) to discuss and brainstorm a proposal, than a week to reach consensus on the approval. The first week could be skipped if the proposal was first discussed on a Wikiproject. So in effect, the Wikipedia:Proposed category reorganizations would also serve as a WikiProject for subjects that don't yet have a WikiProject, or the larger scale category structure.

-- Samuel Wantman 21:33, 4 April 2006 (UTC)

So, by your #2 above, if people are categorized by language and nationality then all people should be listed in the People category? I wouldn't find that a good thing. --JeffW 02:30, 5 April 2006 (UTC)

That's not what i meant. I meant if a category like actors, directors, writers, singers, etc... gets subcategorized. -- Samuel Wantman 06:01, 5 April 2006 (UTC)

My proposal:

Category:People

Category:People by language
  Category:Catalan people (1)
  Category:English speakers
   Category:American people
   Category:Australian people
   Category:British people
   Shakira (3)
  Category:Esperantists
  Category:Idists
  Category:Spanish speakers
   Category:Argentine people
   Category:Chilean people
   Category:Colombian people
     Shakira (2)
   Category:Spanish people
   Category:Venezuelan people
  Category:Welsh people
   ...

(1) They are part of Category:Spanish people, then they both speak Spanish and Catalan as per this categorization.

(2) We assume people do speak the native language of the region he are from. Shakira is supposed to speak Spanish because Colombia is a Spanish-speaker country.

(3) When he/she speaks other language (and it's worthy mentioning), then we include it in other language category. In this case, Shakira being Colombian (Spanish spaker) sings in English, then we include her into English speaker category.

I think this would be the best way. Mxcatania 14:03, 5 April 2006 (UTC)

This probably won't fly. The problem is that you end up putting people categorized by nationality into categories by language. Multi-lingual countries like the US, get put into English, so then all the Spanish speaking Americans are put in a subcategory of English speakers. This is not how categorization is supposed to work. To make these categories work, you need to have subcategories that are subsets of the larger categories, not ones in which only a majority are members. Category membership is for articles in which the categorization is an important attribute in the article, which is discussed in the article. Even if you discount that argument and say that everyone in the US category speaks some english, how do you put the Spanish speaking Americans in the Spanish speakers' category?-- Samuel Wantman 08:46, 6 April 2006 (UTC)

Samuel, you have a point there. But the Spanish speaking Americans - if notable - applies to (2) same as Shakira applies to English speakers. I mean, if someone in the US speaks Spanish and it's worthy mentioning (maybe Christina Aguilera), then you tag him/her as Spanish speaker as an exception. We will do exceptions for cases like this (Category:Mexican American writers will apply to Spanish speakers being tagged ad hoc and English speakers being American people). Let me know your comments. Best, Mxcatania 13:05, 6 April 2006 (UTC)

I'd dispense with the nationalities. It is reasonable to have categorization by language for those categories where language is important and notable. So instead of adding nationality subcats to Spanish speakers, I'd add subcats like Spanish language writers, Spanish language actors, Spanish language orators, Spanish language singers, Spanish language comedians, etc...' All these categories make important distinctions about language, nationality does not. -- Samuel Wantman 05:40, 7 April 2006 (UTC)

Totally agree with you, Samuel. Your proposal is better. Do we need some consensus to go on? Or may we begin? Best, Mxcatania 12:58, 7 April 2006 (UTC)

If you are just adding these categories on top of the existing structure without removing anyone from any of the categories in which they now reside, no more discussion is necessary. Typically these categories are created and populated withoug prior discussion. The extra discussion here means that the new categories are more likely to be found acceptable. Once created, anyone may list these categories for deletion or renaming at WP:CFD. Make certain you follow the current guidelines for categorizing people, and everything is correctly capitalized. I'm wondering about alternatives to "Fooish language fooers", like "Fooish speaking fooers" and "Fooers who foo in Fooish". I don't think these are better, but perhaps someone else has an idea of a better format. This can also be easily changed later at WP:CFD so there really is no impediment to beginning. -- Samuel Wantman 06:39, 9 April 2006 (UTC)

The more I think about this, the more important and valuable it seems. I'm hoping these categories can be widely applied. I'm especially looking forward to seeing the subcategories of Category:Film directors by language, which is much more useful than Category:Film directors by nationality.--Samuel Wantman 07:54, 10 April 2006 (UTC)

A Radical Proposal

I've been thinking about Samuel Wantman's proposal to limit the eponymous categories unless there are at least a dozen related articles. But what if we went the other way?

What if every article in the main space was also a category? Then if I had an article that is related to Adidas, I could just add Adidas as its category without worrying about creating the category. I have two ideas for the user interface. In the first there would be buttons at the top and bottom of every article to take you to the category side of the article. The second would be to put the panes for subcategories and included pages at the bottom of the page, so that you would have the article text, followed by the categories the article belongs in, followed by the subcategories of the article/category, followed by the articles that have this article as their category.

Now, how to do it technically is a different matter and I'm not familiar with the the technicals of Wikimedia and I realize that the programmers have a big backlog as it is. But its something to think about. --JeffW 13:41, 4 April 2006 (UTC)

Your idea has merit. Unfortunately, it seems near impossible to make such a big change to Wikipedia. I can recall similar suggestions to yours in the past. When portals were beginning to show up I suggested that they be moved into the article space, and the topic articles be moved to ''Subject'' (main article). Like your suggestion above, there was hardly a comment responding. Wikis don't seem to undergo major revolutions, but rather evolve slowly. Trying to get them to change radically is very frustrating. -- Samuel Wantman 07:08, 9 April 2006 (UTC)

This suggestion seems to add an unnecessary extra layer to categories. It is nice to have a few links at the bottom of an article that you can use to find related articles. I think there is too much of a drive to turn it into a hierarchy. Cacophony 07:40, 9 April 2006 (UTC)

Hm. But that is what categorization is. Is it a waste of time to try to make a hierarchy where an article may be found by following links down from a major category? Putting a few sideways links doesn't hurt too much but where do you stop? It seems like what you want are See Also links. Perhaps these could be made bi-directional so that if I add a link in the See Also section that the system automatically adds the back link? --JeffW 14:08, 9 April 2006 (UTC)

At the most simplistic level, this suggestion doesn't seem much different from what "What Links Here" already accomplishes. For more esoteric topics, a category really isn't necessary, and major topics almost all have one or more categories of their own already. -Sean Curtin 04:34, 11 April 2006 (UTC)

"Proposed Category Reorganisations" page

I think Samuel Wantman's idea above for a "Proposed category reorganisations" page is a great idea. As well as the proposed "People by language" organisation, I would like to add the "Geography/Earth sciences" reorganisation I proposed above. That is not so much a reorganisation, but a tidying up and correction. I come across quite a lot of these big tangled messes in my perusal of the category system, and it would be nice if there was a central place to mention them and get people interested in sorting them out. Obviously the Wikiproject, if active or interested in categorisation, is a good place to start, but I'm never sure whether to bring these things up on the project pages or here. I guess "both" is the answer.

Another thing I want to ask about is the stability of the categories. I often find that if you leave a neat and tidy area of the category tree for a few months, it can have degenerated into an untidy mess by the time you come back. Is there any way to address this? I find that people tend to ignore or not read clear instructions in the blurb at the top of a category page, and are even less likely to look at the talk page for a category. Carcharoth 09:51, 5 April 2006 (UTC)

Yes, "both" is the answer. Sometimes I try to think of where the "audience" for the discussion might be watching and post a short description with a link to where the discussion will be. Perhaps these should also be posted at WP:CFD because that has a very dedicated audience. The first step to making something like this happen is to write up a description at Wikipedia:Proposed category reorganizations and create a template for posting on the categories. I would be happy to work on this next week (after I finish my taxes).

I am also often frustrated to see what has become of categories that were just fine when I last looked. I'm still upset about the depopulating of Category:Film directors into dozens of little subcategories. Perhaps the best way to deal with these degenerated categories is to use WP:AWB. -- Samuel Wantman 06:57, 9 April 2006 (UTC)

Top-sorting, again

I've had a number of recent discussions about top-sorting of subcategories with people I've ended up working at cross-purposes with, through no fault of any of us. Would there be any objection to a) saying something about top-sorting being a good idea, and b) if so (and if not, if it happens to be done anyway) which character to use? I'll put my 2 eurocents in for "yes", and for "*", respectively, but I'd just appreciate some clarity, one way or another. Alai 22:19, 5 April 2006 (UTC)

What kind of top sorting? I've been going by Main articles should be top-sorted with a space and other things like lists should be top-sorted with an asterisk. Although I haven't been automatically top-sorting lists without looking to see what's been done with that category already. --JeffW 22:56, 5 April 2006 (UTC)

As I say, of subcategories. Stub (sub-)categories as it turns out, but I assume a more general principle applies. Alai 23:41, 5 April 2006 (UTC)

Fwiw, I like the idea of using | foo (pipe+space+sub-cat title) to sort the main article for a category to the very top, then using |*foo (pipe+asterisk+sub-cat title) to sort all the other sub-cats to the top. Anyone using the category then doesn't have to guess whether a sub-cat is worded in such a way as to fall under varying letters (e.g. Images of Cthulhu might be under I or C, depending on who sorted it). Her Pegship 20:37, 6 April 2006 (UTC)

Articles that cover many topics briefly - how to categorise?

I have come across several articles that give an overview of an area, and thus touch on many subject areas that it might be possible to categorise by. What is the guideline for this sort of article? Categorise as widely as possible, or try and restrict it? An example is Rationing, which I recently categorised under several other categories that are relevant for the sections of the article. In effect, I am categorising based on the sections. Is this done in general? Are there better ways to do this? How briefly or extensive does a section have to be to make the article deserve/not deserve a category based on that section? Carcharoth 13:02, 7 April 2006 (UTC)

Following on from the above, is there a category for "topic disambiguation" pages? The example I gave of rationing seems to be disambiguating the term between different subject areas. I suppose if articles on the individual sections were ever created or spun off the article, then the relevant categories would go with them, and the only ones left behind, eventually, would be the "topic disambiguation" category, and the "general topic categories" (here, probably administration and economics). Carcharoth 13:12, 7 April 2006 (UTC)

It's not a disambiguation page, it's a multi-stub page. See Wikipedia:Disambiguation#Summary or multi-stub pages. Multi-stubs do not get the disambiguation notice (and hence, should not get the category either). --TreyHarris 14:22, 9 April 2006 (UTC)

Brilliant. Thanks. Carcharoth 17:41, 9 April 2006 (UTC)

PS. What is the verdict on how to categorise multi-stub pages? I don't mean categories that say it is a multi-stub page, but do they need categories for each of the "stubs"? Eg. Medicine, Military, for the rationing multi-stub page? Carcharoth 17:42, 9 April 2006 (UTC)

From a category perspective I'd rather see the article broken up into it's component stub pages. Absent that I think you have to add categories for each stub within the page so the information can be found by following categories. (This could also be solved by allowing category links to apply to sections instead of the page as a whole, but I'm done with offering pie-in-the-sky technical fixes for problems). --JeffW 16:39, 11 April 2006 (UTC)

Categorising "Categories by name" categories

I've recently been looking at Category:Categories by person and was horrified to see that many of these eponymous categories are categorised as if they were the article about the person!!! This strikes me as just plain wrong.

Looking at two examples of these eponymous categories, such as Category:Bill Gates and Category:Beethoven, the categories typically are placed in categories like Category:Billionaires and Category:German pianists. This then leads to the ridiculous situation where articles in the original "Bill Gates" and "Beethoven" categories, such as The_Road_Ahead and Scots_Musical_Museum appearing under "Billionaires" and "German pianists" respectively.

In my opinion, people will arrive at an eponymous category by clicking on the link at the bottom of an article in that category, and they will be pleased to see this lead to a wide-ranging selection of articles on that topic.

The alternative route for arriving at the category is through browsing downwards through the category "tree". Taking "Beethoven" as an example, people arrive at (say) "German pianists", and see that there is a "Beethoven" category, and click on it to get to the article and associated categories and articles. In my opinion, the correct route is rearrange things so that people arrive in the category "German pianists" and click on the article "Ludwig van Beethoven" and then click on the Category link at the bottom of that page. If that seems too arduous, it is possible to create what I call a "soft" link to the category from (in this case) the "German pianist" category, by inserting the following as the category blurb: "This category is for German pianists. For a collection of articles about Beethoven and his works, see Category:Beethoven."

[Of course, there should be a way to distinguish between sideways links and up and down links, but the Wikipedia categorisation system isn't advanced enough to do that.]

In other words, these eponymous categories should draw articles together from a wide range of different areas, linked only by the person (or topic - see next comment below), but these categories should be at the apex of a tree, with only one parent category - in all these cases Category:Categories by person. For an example of this, see Category:Hillary Rodham Clinton.

So I think that all 58 of these eponymous categories should be stripped of all their categories except the Category:Categories by person tag, and these tags added back to the "root article". This will make the categorisation more logical. It will still be possible to hop from the grouped articles up to the category, but accessing the "by name" category from the category tree will be less easy, though the links can be added back as "soft" links.

What do people think? Carcharoth 20:54, 10 April 2006 (UTC)

Is this just about eponymous people categories or does it apply to eponymous company categories and eponymous topic categories? --JeffW 23:58, 10 April 2006 (UTC)

Um. I would apply it to eponymous topic categories as well, and, thinking about it, to eponymous company categories as well. I'm trying to find the relevant categories. But the main thing to remember is NOT to categorise a category as if it was an article. It should be categorised on the basis of what articles and subcategories it contains! This should be emphasised MUCH more strongly in the guidelines. Categorising a category is different from categorising an article. The trap is particularly easy to fall into when the category name looks like an article name. Carcharoth 07:58, 11 April 2006 (UTC)

I've been pushing for this sort of thing for quite some time (over a year). When the new guidelines were put in place a few months ago, it formalized a transition from the opposite situation which used to be the norm. The previous norm was that the article was ONLY put into the eponymous category, and that category was made a subcategory wherever the article belonged. Thankfully, we finally reached consensus to categorize each by its own merits, and that logically leads to the suggestion put forward by Carcharoth. What I find interesting in all this is that when the same thing was proposed just a few months ago it was very strongly resisted. What was needed was a smaller step (the modified guidelines), which once implemented makes this step more obvious (although I'm sure there are many who will not agree). To make a big change at Wikipedia, you have to plan a series of small steps to get there and boldly start implementing and discussing them. -- Samuel Wantman 07:26, 11 April 2006 (UTC)

I'm still trying to think this through, but so far I agree with this. I find it jarring when I come across a person, company, or some other singular item in the subcategory list of a category page, and it's particularly surprising when the same name shows up in both sections. So I agree that only the article should appear in most categories, and the eponymous categories should be categorized separately from the normal system. However, I think Category:Categories by person is misnamed and miscategorized (see my comment below). A better name would be Category:Categories of individuals and it would be a subcategory of Category:Categories of item. --JeffW 17:42, 11 April 2006 (UTC)

Picking up the thread from further up, I looked at Category:Categories by topic, and there are currently 11 subcategories.

This organisation in this "by topic" category, strikes me as being a bit haphazard at the moment. I arrived there expecting to see a large alphabetical list of subcategories in "by topic", possibly subcategorised by "categories by fooian topic". But instead there is this rather strange listing. I guess I should start discussion on this at the (currently non-existant talk page for the "by topic" category page), but for now, I've put comments below:

These all appear to be geographical and should, IMO, be in something like Category:Categories by location, which would then subsume back under the "Geography" category and the "Categories" category.

Category:Categories by company

This currently only contains computer companies, though I can see how the concept could be expanded, and it does look like a very useful category. Looking at how the company categories have themselves been categorised, my comments above about eponymous categories would also apply.

From looking at all these categories, I'm still not sure what a "topic" is! I'm not sure that language, nationality and people are topics, but rather they are something else. Language and nationality are cultural labels, while the people one should explain itself. I think that any "Category by Foo" would fit well in the "Categories" category and the "Foo" category, but shoving them in this "by topic" category feels a bit artificial.

I even tried to click on topic to find out what it means, but ran into four layers of disambiguation pages and still failed to find a suitable Wikipedia page on what a "topic" is: topic -> subject -> content -> information?

Overall, my impression from these wide-ranging categories found in Category:Categories by topic is that they are trying to bring together loosely related articles in much the same way that a portal does. Indeed, the portals are only one level up to Category:Categories, and then one level down to Category:Portals. So Category:Portals and Category:Categories by topic do seem to be brother super-categories, of sorts!

My argument would be that, since there is no Portal:Topics, all the "by topic" categories could simply link into something similar to Category:Browse, which looks rather odd at the moment. Maybe I am thinking of Portal:Browse?

Anyway, what I would expect to see under "categories by topic" is a large list of categories named for their topics. Either "Foo" or "Foo topics". For example, Category:War (different from Category:Wars) or Category:War topics. Ultimately, this reminds me of the distinction between "topic" categories like Category:Opera and Category:Disaster and the corresponding "event/object" categories like Category:Operas and Category:Disasters.

I would expect to see the "topical" categories link up to the "by topic" top level eventually. The path from "Opera" is via "Genres", but the path from "Disaster" or "War" is less clear. The route from "War" seems to stop at "Behaviour", which then leads to various "Top 10" or "Fundamental" categories, though, IMO, a link to "by topic" could also be inserted here.

Maybe the simplest way to think of this is to say that any category that has a corresponding portal should go in Category:Categories by topic, and maybe even to say that only portal-like categories should go in Category:Categories by topic? Does that sound reasonable? To have the "by topic" category be an analogue of the Portal category? Carcharoth 07:58, 11 April 2006 (UTC)

This all has me rather confused but meaning of Category:Categories by topic seems to be any category that starts with "Category by...". But what does Category by Foo mean? Category:Categories by person contains all the eponymous categories named after people. But Category:Categories by city just contains other categories all of the form "Foo by City" and it specifically does not contain eponymous categories named after cities. So there are two vastly different concepts here masquerading under the same name. --JeffW 16:55, 11 April 2006 (UTC)

To follow up I decided to investigate how a city with an eponymous category is categorized. I picked Chicago, whose only relevant categories to this discussion are its eponymous category and cat:Communities on US Highway 66. The cat:Chicago is linked in several ways, but one path in particular leads to cat:Cities in Illinois, cat:Cities in US, cat:Cities by country, cat:Categories by country, and finally to cat:Categories by topic. Carcharoth, this isn't your ideal categorization because the Chicago article is linked almost exclusively to its eponymous category and most linkages are through the category. Would you just turn that around and remove all the categories from cat:Chicago and transfer them to the article Chicago? What should the parent of the Chicago category be then?

My response here would be that the Chicago category is acting like a portal for a wide range of Chicago articles. As such it should be a natural endpoint, and should have few parents, or even (gasp) none. I don't see any reason why it can't be seen as a natural dead-end for a specialised purpose (gathering together "Chicago" articles), with only one route in (from any of the articles in the category). Then, to get back out into the "tree", you have to exit via the way you came in, or via a different article. I see absolutely no reason why every category has to have a parent category - there are plenty of routes away from the category - the fact that they all point downwards into article space (or subcategories), rather than upwards (into parent categories), is neither here nor there. Carcharoth 15:32, 12 April 2006 (UTC)

And should we apply this model to people? Chicago isn't linked, even indirectly, to cat:Categories by city but instead to cat:Categories by country. By this scheme, cat:People by occupation should be in cat:Categories by occupation. Other things in this category might be cat:Journals by occupation or cat:Tools by occupation. And eponymous people categories shouldn't be in cat:Categories by people, but the answer of where to put them depends on your answer to where to put the eponymous Chicago category. --JeffW 21:17, 11 April 2006 (UTC)

I'm still very confused at this point by the naming conventions being used in this area of the "tree". One thing that I think is confusing me, and maybe others, is the distinction between "Categories by Foo" and "Foo by category". I think the semantics of the English language means these constructions can mean similar things to some people and in some contexts. I agree with your "Category of item" suggestion further up the page, and I'm now trying to figure out how using "of" makes things different from "by"... Carcharoth 15:32, 12 April 2006 (UTC)

It gets more confusing. Category:Lists by country has in a sense the opposite meaning of Category:Categories by country. I think the meaning of the lists category is more natural. --JeffW 13:30, 15 April 2006 (UTC)

To take the specific example that JeffW raised above, Category:Chicago, Illinois is well-integrated into a geographical hierarchy, and as such, articles about the geography of Chicago fit well in that category. Tracing back up the category tree, we see that the slip in language came at the point where "Geography of the United States" led to "Cities of the United States" - the correct formulation would have been "Geography of the cities of the United States". The alternative wording has led people to treat it as a general all-purpose category for any articles remotely related to US cities. The wording of intermediate subcategories gets rather long, but the point should be that these are only used to categorise categories, so it doesn't matter - they are purely category-category links. Then, as you get near the article-inhabited part of the "tree", the wording narrows again, so you end up with "Geography of Chicago", which is in both "Geography" and "Geography of cities of the United States", etc. and also in "Chicago". The Chicago category would contain all the articles concerning Chicago, but starting from the apex category "Chicago", rather than reaching it through the apex category "Geography". Other subcategories might be "History of Chicago", which would be in both "History" and "History of cities of the United States". It all flows naturally from careful consideration of the names of the categories. Which is probably why there is so much activity in deleting and renaming categories. People intuitively realise that misleading naming can cause chaos in a categorisation system. Carcharoth 16:19, 12 April 2006 (UTC)

Categories by name and Categories by topic

I'm confused over whether a category of articles linked by a person should be grouped under Category:Categories by person or Category:Categories by topic? I think that the "by person" category should be used to group articles about the person (biographical, what the person did in their lifetime, etc), and that this should be clearly distinguished from a higher level category that groups a wider range of articles linked by the topic of that person and his works and related materials. The "by name" categories are already in a subcategory of the "by topic" category, so the logic works well here.

Essentially, I would expect to be able to find a group of articles about person X by finding the article on person X, and then clicking on a category link at the bottom to a category of articles about the person, and then clicking on a link in that category to get to a grouping of articles about the subject concerning that person and wider issues such as study of that person and the influence of that person.

Does this description of "by topic" and "by person" sound right? Carcharoth 21:07, 10 April 2006 (UTC)

Category sorting in "Top 10" category

In Category:Top_10, the subcategories and articles are sorted using the "space" trick after the "bar". I like the way this looks, though I didn't start it. However, everytime I check back, someone will have spotted this "strange" space and removed it, saying they like the categories appearing under those big "ABC" letters. Of course, the reason for the pipe sorting in this category, which by definition has a limited number of subcategories, is to avoid this and just have them list in alphabetical order without the lettering layout. But how can you stop people changing things back? I can bet that if I come back 6 months later, it will be a mixture again. Is this a good case for adding comments only visible to editors? I've done it on Technology as an example. Carcharoth 21:20, 10 April 2006 (UTC)

Using the "space trick" in this context is simply subverting the usual MediaWiki category sorting mechanism for no good reason (simple aesthetics, which will differ from person to person). That's why it keeps getting changed back. I would change it myself right now if I weren't in the middle of something else. There are good reasons to change the sorting in certain cases, as described in this article — this isn't one of them. - dcljr (talk) 23:03, 10 April 2006 (UTC)

Its not just simple aesthetics. When there are only ever going to be 10 subcategories having those letters breaking up the list just makes it harder to quickly read through the list; its just not good ergonomics. --JeffW 23:56, 10 April 2006 (UTC)

And sometimes you get two or more entries under one letter, and the list is broken up into uneven chunks. I don't personally agree with the concept of a Top 10 category, but it is a good example of a category that, by definition, will never expand beyond its current contents, as you say. Unfortunately, because category pages are generated dynamically, you have to use screenshots to show differences in the layout. OTOH, worrying about category page layout seems a bit silly, as we have so little control over that anyway.

BTW, in response to dcljr, this trick does not subvert the usual MediaWiki category sorting mechanism. The entries are still sort alphabetically as long as you put something after the space. In this case "....| Technology" (note the space after the bar).

I've now seen "*" and " " used as 'placeholders' to start different lists in front of the "ABC..." listing. Are there any other characters that can be used, or, more importantly, characters that should be avoided? Could you use "#" or "$" or "&" or other symbols? Carcharoth 06:16, 11 April 2006 (UTC)

Perhaps I should have said, "subverting the default MediaWiki category display mechanism", or some such. I didn't mean that it doesn't still sort alphabetically; I was referring to supressing the letter headers (I said "sorting" because the stuff after the "|" character is usually called the sortkey). Anyway... another popular sortkey character is "μ" for stubs (example). I would guess that trying to use "]" or "|" wouldn't work, since those are part of wikilink syntax. Apart from that, there probably aren't any other characters you can't use; but there are probably quite a few you shouldn't use, because they would only confuse viewers of the category ("why are these listed under '$'?"). - dcljr (talk) 06:15, 13 April 2006 (UTC)

Categorizing an article by a different name?

Is there a way to categorize an article by a name other than its title? For example, say philosopher Foo has an article. Foo is most famous for his Bar theory, and much of the article is concerned with Bar. Can I have a link titled "Bar" in Category:Philosophical theories which is piped to the article about Foo, or would Bar need to have its own page? - Rynne 22:44, 10 April 2006 (UTC)

No, you can sort it by something other than its title, but it will always display on the category page with the title of the article. In this case, since most of the article is about Bar, perhaps you should change the article to be about Bar and move the page to Bar. This will leave a redirect at Foo pointing at Bar...which makes me think of something else.

You could add a redirect from Foo to Bar and add Foo to the category. I don't know if that would be considered a valid use of a redirect. --JeffW 23:53, 10 April 2006 (UTC)

Actually, if Bar is a redirect to Foo, it is possible to add the category to Bar, which is what will appear in the category list. The category must be on the same line as the redirect test or else it is ignored. I use this fairly often, although I also see folks removing the category from redirects -- I'm not aware of any guidelines regarding this. older ≠ wiser 00:47, 11 April 2006 (UTC)

There are Bar --> Foo redirects, so I'll give that a shot. Thank you. - Rynne 13:50, 11 April 2006 (UTC)

category research

Hi, I compared MediaWiki's category system and its usage in Wikipedia with traditional classification and collaborative tagging. People tend to confuse a lot in categories so maybe my paper can help a bit (or confuse you even more - it's a scientific research paper ;-)

the paper, short intro

Greetings -- Nichtich 09:22, 11 April 2006 (UTC)

Proper hierarchy for naval categories?

What is the proper hierarchy for including naval subcategories, articles or subcategories? For example, looking at Category:Destroyer classes, almost all of the ships classes have both an article and a subcategory (such as the Arleigh Burke class destroyer article and Category:Arleigh Burke class destroyers), however in some cases the article is a member of Category:Destroyer classes while in other cases it is the subcategory. Thanks, --Kralizec! (talk) 15:49, 12 April 2006 (UTC)

Sanity check

Would it be ok if I were to tag chinese ctities under Category:Taiwan as the goverment of Taiwan (Republic of China) claims to rule mainland china? --Cool Cat^Talk|@ 09:19, 12 April 2006 (UTC)

Seems like a question you could easily answer for yourself... if not, maybe after reading Wikipedia:NPOV tutorial#Categorisation, and other guidance linked from there.
If still not being able to answer that question for yourself after reading all that, maybe ask your question at wikipedia talk:categorization, or start an RfC (but I think you may assume that the outcome of such RfC would be pretty much predictable - only encouraging you to try to find a sensible answer to your question yourself - if you'd try to find it yourself, I'm convinced the eventual answer will stick better)

--Francis Schonken 09:49, 12 April 2006 (UTC)

You are right the outcome is easy to guess. The result would be that what I am suggesting is nonsense right? I for one would oppose tagging of Beijing under a taiwanieese category. I'll copy this chain to all those pages you mentioned. --Cool Cat^Talk|@ 08:50, 15 April 2006 (UTC)

From Wikipedia talk:Neutral point of view --Cool Cat^Talk|@ 08:50, 15 April 2006 (UTC)

Any comments? --Cool Cat^Talk|@ 08:10, 17 April 2006 (UTC)

Categorizing Multi-Page Articles

When an article is split up into multiple pages, how should the "sub"-pages be categorized? If it is a list that is broken out then it can be put in a Lists sub-category, but what should be done in the general case? For a specific example, see Facebook features. This article is closely related to and is linked to by Facebook but I don't see the point of putting the child article into the same categories as the parent. That would just cause clutter in those categories.

I'm tempted to create a Facebook category to hold all the related pages, but I'm sure some would object to that. On the other hand, maybe it should be ok for a page to not be in a category, but then it will keep showing up on lists of uncategorized pages. Any suggestions? --JeffW 21:26, 17 April 2006 (UTC)

I notice that Facebook's use in investigations is categorized in some of the same categories as Facebook and I see nothing wrong with that. Often sub-articles get categorized differently from the main article, and it is good to consider what categories might be different. For example I notice that this article is in Category:Surveillance and Facebook is not.

There is no simple answer for deciding when to make the eponymous category. If there are numerous related articles, it might be usefull to make a new category. Personally, I think they should be avoided until it is certain that the category will add needed utility for those browsing the subject. If there are just a few articles and they are all already cross-linked as "main articles" or under "see also", there isn't much of anything to be gained by making the category. -- Samuel Wantman 08:25, 24 April 2006 (UTC)

Well, maybe you could put Facebook features into Category:Software features but I think that would be a rather pointless category. And while it may be ok in some cases to list the subsidiary page in the same categories as its parent, I think in general it just adds clutter. Maybe the solution is to suggest to the author that the features of a software application are too central to it and shouldn't be broken off into a separate page. --JeffW 19:05, 24 April 2006 (UTC)