Wikipedia talk:Categorization

Soft and hard redirected categories

This guideline clearly indicates redirecting one category page to another must be achieved using the {{category redirect}} template and not via a standard (hard) redirect. What happens, from a technical standpoint, when {{category redirect}} is used in conjunction with a hard redirect, and what is the proper course of action when one encounters such a situation? For an example, see Category:Germany navigation templates. Thanks, -- Black Falcon ^(talk) 02:29, 1 October 2010 (UTC)

Ping. -- Black Falcon ^(talk) 03:15, 13 October 2010 (UTC)

African American category

Continuing the discussion from here but dealing directly with the issue of African Americans. If it is established that the person is American and the accompanying image is of a "Black American" is this sufficient for establishing that an African American category is justified? For an example see the Kevin Durant article, is he an African American basketball player even though there is no source in the article making that specific claim (that I'm aware of)? It's rather difficult to find sources that trace a Black person's ancestry to Africa. Many sources take it as given that someone like Durant (who grew up in a county that is 97% African American) is African American but without supplying evidence for that claim those sources are not useful in this context. I guess what I'm asking is there any room for common sense here? SQGibbon (talk) 19:16, 2 October 2010 (UTC)

In short, no. Maybe Kevin Durant is of Hispanic descent; maybe he is of Scandanavian descent. How do we really know anything on here? Sources. If something cannot be sourced, I don't think it has a place on Wikipedia. Kevin Durant is probably African American, but probably isn't good enough, especially with BLPs. Some editors added categories based on last names, some using pictures and some just their perceptions. In my opinion, all of these are original research; this is why I have been removing uncited ethnicity categories from articles for at least a year.--TM 19:22, 2 October 2010 (UTC)

Just a sidenote, but it is not neccesary to trace one's ancestry back to Africa to be an African American. It is a matter of self identification. Does the person call themselves African American? Do reliable sources? In my opinion, that is enough.--TM 19:31, 2 October 2010 (UTC)

Kevin Durant was born in America (according to the sources in the article) and is African American. This is not based on his name or "just" a perception. He is African American, not probably an African American, is an African American. This is clearly an issue of common sense. If there were even the slightest chance that he (and the other African American basketball players whom you removed from that category) is not African American then we wouldn't be having this discussion. In fact it's so plainly clear that he is African American that the only way to even make your argument you would need to come up with a reliable source that casts doubt on that claim. SQGibbon (talk) 20:44, 2 October 2010 (UTC)

That's not how it works here. He was born in the US and we have sources for that. We have no proof of his heritage. Therefore, we should not include assertions, categories or anything else in articles which cannot be confirmed by sources. There is no assumption of truth on Wikipedia. The only way to include something when challenged is to prove it.--TM 20:52, 2 October 2010 (UTC)

(edit conflict) Are you challenging that he's African American? Do you think it likely that anyone will ever seriously challenge that he's African American? Not every single trivial and clearly obvious fact about a person needs to have a source. If I couldn't find a reliable source claiming that Durant is male would that mean we couldn't make that claim? Anyway, you and I are getting nowhere on this, let's see what anyone else has to say. SQGibbon (talk) 21:00, 2 October 2010 (UTC)

I'll give another example: What if I said Kevin Durant is of Puerto Rican heritage. Now, I can say it is simply common sense and everyone knows it, but you don't think so, so you challenge me to prove it. Should I include the Puerto Rican category in the article? The onus isn't on the editor removing uncited content, it is on the editor trying to include the content to prove it.--TM 20:58, 2 October 2010 (UTC)

I think that you both raise interesting points, but I do believe that a source is necessary. While it is completely true that "not every single trivial and clearly obvious fact" in an article needs to be attributed to a source, such facts still should be attributable.

"All material in Wikipedia articles must be attributable to a reliable published source to show that it is not original research, but in practice not everything need actually be attributed."

–Wikipedia:Verifiability (version)

For living people, in particular, the requirement for justifying categorization with reliable sources is more explicit.

"Category names do not carry disclaimers or modifiers, so the case for each category must be made clear by the article text and its reliable sources."

–Wikipedia:Biographies of living persons, section 'Categories, lists and navigation templates' (version)

So, then, is ethnic or racial descent a "clearly obvious fact"? Some cases may appear more obvious than others (Kevin Durant being a case in point), but I do believe that we cannot reliably infer, from physical appearance alone, racial or ethnic descent, affiliation or self-identification. For example, judging Alex Rodriguez or Halle Berry by physical appearance only would not give a reliable clue as to whether they are of Hispanic, African, Middle Eastern, South Asian, or dark-skinned non-Hispanic White descent. The fact that the same person can have a different appearance depending on lighting conditions, make-up, sun tanning or tanning lotion, and other factors, further complicates the exercise (see, e.g., the following images of Halle Berry, which appear to show a progression of skin tones: 1, 2, 3, 4, 5). -- Black Falcon ^(talk) 21:51, 2 October 2010 (UTC)

I started checking for a source for the Kevin Durant article, and it is in fact difficult to find a source which identifies Durant as African American, perhaps because it does appear to be such an obvious fact. This article from the Bleacher Report was the best one I've found so far, insofar as it identifies his race, but it isn't ideal. -- Black Falcon ^(talk) 22:08, 2 October 2010 (UTC)

African Americans are more than just skin color; not all Brown people are of African descent and not lighter skinned people are of European descent. We just don't know. I'd rather play it safe and not list him (and thousands of others) as an African American than potentially end up with a BLP problem because we just assumed it was common sense. It is really a matter of self identification. Most people in countries such as the United States have multiple racial and ethnic groups in their family heritage. Take for example Malcolm X. It is common sense that he was African American and he identified as such, but he also had significant amounts of Hispanic and European ancestry.--TM 22:16, 2 October 2010 (UTC)

(edit conflict) The question is not if ethnic or racial descent is a clearly obvious fact, it's whether in the particular cases of well-known African American athletes (in this case basketball players) like Kevin Durant if we need a source to attest to the fact that he is African American. This is not a slippery-slope case of wanting to expand this to all ethnicities for all people but only when dealing with the obvious cases like Kevin Durant. Common sense has to have a place in Wikipedia. And yes, if there were even a hint of a whiff that he might not be African American like with A-Rod or Halle Berry then that too would be a different situation. But here (and with other African American basketball players TM removed from that category) surely we can use common sense. Sources show that he is American. He was born in and grew up in an area that is 97% African American. He is often associated with various organizations that support/work with African Americans. Given that he looks African American and that I have not been able to find a single source that even hints that he is not African American I think it is trivially safe to say he African American (oddly I have not been able to find any reliable sources that state specifically he is African American). This does not have to become Wikipedia policy, we can make these judgements on case-by-case basis. SQGibbon (talk) 22:17, 2 October 2010 (UTC)

(edit conflict) I see what you're saying. All of the available information points to him being African American but, frustratingly, falls just short of explicitly confirming it, and I agree that exceptions may be appropriate in a limited number of cases (decided on a case-by-case basis). That being said, I also can understand TM's point about not acting on the basis of commonly-held, but unverified, assumptions. In light of your clarification that the issue does not concern all biographical articles but instead only the article Kevin Durant and some others, I posted a link to this discussion at Wikipedia talk:Biographies of living persons and requested additional opinions on the best way to proceed (perhaps it would be useful to also notify WikiProject Basketball...). -- Black Falcon ^(talk) 06:06, 3 October 2010 (UTC)

Here's the thing: if Kevin Durant weren't black, then there would be ample sources attesting to that fact. The NBA constantly tries to portray itself as ethnically diverse (see here, for example - a gallery of all the Latino players). Plus, journalists are always looking for interesting stories about players' backgrounds (see this article about Nate Robinson's single Filipino great-grandfather). Zagalejo^^^ 05:57, 3 October 2010 (UTC)

I thought the general rule was that we didn't categorize any article by any fact not even mentioned in the article (at a minimum). I don't see a reason to deviate from that. postdlf (talk) 12:55, 3 October 2010 (UTC)

It would be trivial to add "African American" to the lede and such an obvious edit would unlikely fall under any serious challenge. SQGibbon (talk) 12:38, 4 October 2010 (UTC)

What do you think this is? Not a serious challenge? WP:PROVEIT or keep it out of the article and all other articles. It is really simple: if you can't prove it, it isn't necessary.--TM 16:20, 4 October 2010 (UTC)

I apologize, I was under the impression that you were challenging whether the African American basketball player category should be applied to the Kevin Durant (and others) article. I had no idea that you were putting forth a serious challenge to the fact that Durant is African American. SQGibbon (talk) 18:44, 4 October 2010 (UTC)

We aren't discussing what Durant IS but what we can PROVE. We aren't saying he is African American or any other ethnicity because we have no proof. Pictures are not proof. "Common sense" is not proof. Anything without proof (i.e. reliable sources saying so) must be removed from BLPs. This is all basic policy written elsewhere.--TM 18:54, 4 October 2010 (UTC)

This just isn't true, let me quote from WP:BLP "All quotations and any material challenged or likely to be challenged must be attributed to a reliable, published source using an inline citation." Meaning if something is not likely to be challenged it's potentially OK for it to be included in an article (there can be other reasons for not including it). Notice that WP:PROVEIT uses the exact same text. Since there is plenty of evidence that Durant (and the others) is African American, and there is not even a hint of evidence that he's not, and there's nothing at all even slightly controversial about asserting the fact that he is African American then including this claim in his article is not in violation of WP policy. Unless, perhaps, you are putting forth a serious challenge to the claim that he is African American. SQGibbon (talk) 19:08, 4 October 2010 (UTC)

I think you should read further down. As quoted from BlackFalcon above, "Category names do not carry disclaimers or modifiers, so the case for each category must be made clear by the article text and its reliable sources." --TM 19:25, 4 October 2010 (UTC)

Ah crap, that was for me? Your formatting was off by a level (fixed now). (I didn't think your statement here made sense as a response to what postdlf wrote). Anyway, this is my first foray into The Category Wars and I admit there are things about categories about which I was unfamiliar. That said, from the point of view of just the article, I believe adding African American is entirely in line with WP policy. Do you agree? Your position, as I understand it, is that even with the article making the claim that Durant (and the others) is African American is not sufficient for categorizing him as an African American basketball player unless there is a reliable source supporting that claim. Correct? Did you read the source that Black Falcon supplied above (link to the source)? That is a reliable source (I think), is that sufficient for establishing Durant as African American? If not, and assuming that is a reliable source, then what would it take?

Finally, having looked deeper into categories it appears questionable that the African American basketball player category should even exist. Subjects should be categorized by the terminal sub-category, correct? So someone like Wilt Chamberlain would be African American basketball player but not in the American basketball player category? Which means he won't be in the same category as, say, Larry Bird which further appears to me to be a bad thing. And then also reading here seems to indicate that the African American basketball player category should not exist. I know I've introduced a new topic so feel free to separate this into a new section if desired. SQGibbon (talk) 21:59, 4 October 2010 (UTC)

TM - are you familiar with WP:POINT? Do you actually think that Durant is not African-American, or are you simply arguing that he might not be in order to make a point about sourcing? john k (talk) 00:29, 5 October 2010 (UTC)

John K, I'll quote what I wrote above to answer your question "Kevin Durant is probably African American, but probably isn't good enough, especially with BLPs. Some editors added categories based on last names, some using pictures and some just their perceptions. In my opinion, all of these are original research; this is why I have been removing uncited ethnicity categories from articles for at least a year".--TM 02:21, 5 October 2010 (UTC)

To SQ Gibbon, I didn't look at that source until now. Black and African American in the case of the US are interchangable terms. Given this new information, I am fine with including ‹The template Category link is being considered for merging.› Category:African American basketball players to Kevin Durant with a sentence stating he is so with this citation. I am still going to continue removing uncited, unmentioned ethnicity categories and I invite all other editors to do the same.--TM 02:25, 5 October 2010 (UTC)

Why do we need another African American category? --THE FOUNDERS INTENT ^PRAISE _{GOOD WORKS} 18:22, 21 October 2010 (UTC)

Link to Islands of Maine as well as Islands of Waldo County, or only Islands of Waldo County?

A question has come up as to whether articles about Maine islands, like Sears Island should link only to the county-by-county category, in this case the sub-category ‹The template Category link is being considered for merging.› Category:Islands of Waldo County, Maine, or link in addition to the more general category ‹The template Category link is being considered for merging.› Category:Islands of Maine?

Pro: A couple of editors feel that this ought to be one of those "exceptions to the general rule that pages are not placed in both a category and its subcategory." They feel readers would balk at island finding one county at a time. People think of the Maine Islands as a group. Carl Little's book is called Art of the Maine Islands; the Island Institute in Rockland deals with "island communities off the Maine coast", likewise the Maine Seacoat Mission in Bar Harbor deals with "coastal and island communities;" there is the Maine Island trail "a 375-mile chain of over 180 coastal islands" for kayaks, and so forth. People don't think of the islands county by county. And it might seem a daunting task to figure out which county to look in to find the right one, or to get a list of them all (180? county by county?). Not that the more specific county category is without its uses; just that it ought not to be an either/or situation, but rather--both. ElijahBosley ^{(talk ☞)} 21:07, 13 October 2010 (UTC)

Con: Including all islands of Maine in the one category, which would be in the hundreds if all of them were created, is simply counter-productive to the categorization process. A list is a much more useful tool for those looking for all of the Islands. It can include much more information than a category can. Such lists have been created already for towns and other places.--TM 06:26, 15 October 2010 (UTC)

Comments?

Hmm. Nobody seems interested. Well, Namiba maybe you can persuade me. Tell me what a category is for? My sense is a category is a third way to find information on Wikipedia, a kind of generic word search. The primary two ways seek specific information: 1) word searching Google or Wiki's search engine; 2) clicking on a link within an article. Word searches and link clicking are how people normally get to a page. Once there, they may decide they want information not just on this specific thing, this one island, say, but all the islands. If the page links to both the category and subcategory, they have a choice: Islands by County (and maybe they do want all the islands just in one area) or all Maine Islands. As to lists, there are these shortcomings: currently there is no list of Maine Islands. I will be the first to award a barnstar for the hard work of creating one. But nobody has done so, understandably: would a List of Maine Islands include every named rock that pokes its nose out of a harbor, every flooded hummock of trees in every river? A category, unlike a list, is useful even if incomplete. Also lists are hard to find and update. An editor working on one island page at a time, as I did with a minor update to Sears, would need to click away from the page he is working on to search for an appropriate list (is it List:Islands of Maine? or List: Islands of the New England Coast?, or maybe List: US Islands?). Once he has found the right list, or one of them, he checks the long list to see if that island is on it, and finding it missing, opens up the edit page, then updates the list. And then gnaws his nails worried that he has not updated all the other lists the island should be on. A category is much easier. An editor would simply use Hotcat to add ‹The template Category link is being considered for merging.› Category:Islands of Maine, and since categories are seldom exhaustive, not worry about the fact that he is working incrementally one island at a time. Perhaps I misunderstand the function of a category, but isn't it an easy way for editors to group information to assemble an (incomplete) collection of similar things?ElijahBosley ^{(talk ☞)} 13:54, 18 October 2010 (UTC)

There are several reasons why including all islands in Maine in one category should not be: 1, it is a long established principle that diffusing large categories (as we surely agree a category with all of the islands of Maine would be) is common and useful. Lists are more useful because they can include links to articles not yet written. There are dozens if not hundreds of Maine islands without articles, as I am sure you know. Frankly, it is just common usage not to include a subcategory e.g. ‹The template Category link is being considered for merging.› Category:Islands of Waldo County, Maine and a higher category ‹The template Category link is being considered for merging.› Category:Islands of Maine together on the same page. See for yourself at ‹The template Category link is being considered for merging.› Category:Islands of Florida or ‹The template Category link is being considered for merging.› Category:Schools in Rhode Island by county.--TM 17:48, 18 October 2010 (UTC)

Thanks for your post. As noted above, common usage or at least common parlance would group the Maine Islands as "Maine Islands" rather than Islands by County (the Florida Islands do not have their own separate "persona": nobody writes art books called Art of the Florida Islands). Here's the practicality of it for a Wikipedia user: I just tried going from the Peaks Island page and clicking on the "Islands by County" category looking for islands which seceded from the mainland and became their own separate jurisdiction, as Peaks tried to do and failed. I cannot recall the names of all of the islands that seceded but would recognize them if I saw them. It didn't work. No idea what county they are all in, and so no idea where to look. Frye Island on Sebago Lake I know is one of them, but which county is it in? Now, if Peaks and Frye both were listed in a category Maine Islands, one click would have got me where I needed to go. As opposed to the multiple clicks of culling through lists: there is a general List of islands--including all islands, everywhere, impossibly over-broad and under-detailed, then the List of islands of the United States then within that the List of islands of the Northeast United States --and I still have to scroll down through Connecticut to get to Maine. So to reiterate, for practicality's sake, for the sake of the ordinary Wikipedia user who just wants to get where he's going with one click, and for the editor who wants to get stuff done with one Hotcat click, I still think this is one of the "exceptions to the general rule that pages are not placed in both a category and its subcategory." I do wish some other editors would weigh in as I fear neither of us has succeeded in persuading the other. ElijahBosley ^{(talk ☞)} 21:08, 18 October 2010 (UTC)

Are we talking about a limited subset of all islands off Maine, i.e. the 180 or so mentioned in the book Maine Island trail? Or are we simply talking about all islands currently placed in sub-categories of Category:Islands of Maine? If we are talking about a subset of the absolute total islands, then only would I support additional categorization into such a category, e.g. Category:Maine Island trail. Otherwise I do not see the current lack of the standard List of islands of Maine to be a valid argument for making an exception to the quoted general rule for categorizing only into the most detailed sub-category. Actually I would encourage the creation of this list page immediately as well as an article on the Islands of Maine. With the literature you list in your initial post sufficient sources would be ample to start writing such an article. Also, I would think that creating navigational templates to be placed in all island articles would be a todo item in this process. __meco (talk) 22:59, 18 October 2010 (UTC)

Thanks Meco for a thoughtful and insightful comment. Category:Maine Island Trail is a good idea.ElijahBosley ^{(talk ☞)} 12:26, 19 October 2010 (UTC)

List of islands of Maine was started by myself and greatly expanded by Jlm. Adding to it (and creating the red linked islands) is the next step. I'm glad this has been resulted in an addition to Wikipedia.--TM 17:33, 19 October 2010 (UTC)

Ordering categories

Currently there is no rule or recommendation for ordering categories at the bottom of a page. In my opinion, alphabetical order is the easiest and most unbiased way of sorting them. What are other people's thoughts? McLerristarr / Mclay1 16:26, 14 October 2010 (UTC)

I disagree with alpha order. I put then in what I think is order of priority, but that is subjective of course. The category that is subject of the article, which is often the eponymous category, should be first. The birth year, death year and living people cats should be last IMHO. They are of very little use to readers. -- Alan Liefting (talk) - 19:02, 14 October 2010 (UTC)

I agree. When it comes to biographical articles, I put the most important ones first; the ones that are basically guarantees of notability in that they have no non-notable members (presidents of the U.S., etc.). Then after that comes more general career categories, education, place of origin, and last the vital statistics of birth/death year. Alphabetical is just arbitrary and unhelpful for sorting, because category names aren't organized around the best name for alpha sorting, but rather the most clear name. Compare Category:United States Senators and Category:Members of the United States House of Representatives, for example. So I sort them by general importance and by topical groups in articles I create, and occasionally in articles I just edit. I think that's more helpful for readers, particularly given the flood of category tags some articles have. postdlf (talk) 21:14, 14 October 2010 (UTC)

Many users prefer alphabetical ordering and will change the order of categories if they are not alphabetised. Perhaps we need rules on how or how not to order to prevent re-arrangements. McLerristarr / Mclay1 00:54, 15 October 2010 (UTC)

"Many users prefer alphabetical ordering" ^{[citation needed]}. By user do you mean editor or reader? It is what readers want it what is important. -- Alan Liefting (talk) - 06:20, 15 October 2010 (UTC)

Alphabetical only if there's no better way. Normally logical. I.e. start with the eponymous category (if there is one), then the most "essentially" defining characteristics, then the more obscure properties. Stub categories (which I would prefer to be hidden anyway) last. Generally speaking readers won't know exactly what categories are called, so won't be able to look for them alphabetically anyway. However, alphabetical can be used to order very similar categories that have no other distinguishing criterion (e.g. if a river flows through Albania, Bosnia and Croatia, put the corresponding categories in that order).--Kotniski (talk) 07:37, 15 October 2010 (UTC)

By users I mean people with user pages, so editors. I don't need a citation, I've observed it. Some people (including myself) re-arrange the categories into alphabetical order if they are in no logical order. McLerristarr / Mclay1 11:28, 15 October 2010 (UTC)

We know that people do it. But why they do it, and whether readers generally would prefer it that way, is still unexplained. It doesn't accomplish anything helpful in my view. It often doesn't even group related categories together. postdlf (talk) 13:36, 15 October 2010 (UTC)

I have come across pages that have had the cats placed in alpha order. I recall changing the order to place the more important at the top of the list. -- Alan Liefting (talk) - 20:12, 15 October 2010 (UTC)

The Winston Churchill article is an interesting example since it has a large number of categories. It almost follows an alpha order. -- Alan Liefting (talk) - 20:12, 15 October 2010 (UTC)

We should do an RFC on a guideline. My suggestion is:

"If the article is about an eponymous category then that category should be the first one listed. Any other categories should follow in order of importance. If all the categories are of equal importance then an alphabetical order should be used."

That seems to address the points raised in the discussion above. -- Alan Liefting (talk) - 20:12, 15 October 2010 (UTC)

One side note. The MoS calls for the stub templates to be located following the categories. This will place the stub categories at the end of the list automatically. Vegaswikian (talk) 21:13, 15 October 2010 (UTC)

This does not affect the order of the categories that we are discussing here. Stub categories should of course be at the end. -- Alan Liefting (talk) - 23:25, 15 October 2010 (UTC)

Alan, I'd further state that categories should be grouped thematically, and those groups in turn should be ordered by importance. On Winston Churchill, for example, he has two death-related categories: Category:Cardiovascular disease deaths in England and Category:Deaths from stroke. Currently those are separated by six unrelated categories; those should be grouped together and placed at the end of the category tags, next to birth/death years. I would also group together those relating to his military service, those relating to his MP years...and so forth. postdlf (talk) 21:17, 15 October 2010 (UTC)

Ok, the order should be priority, thematically and then alphabetically. -- Alan Liefting (talk) - 23:25, 15 October 2010 (UTC)

We should not forget the advantages of listing categories alphabetically. When one is seeking a category for a particular intersection or topic, it is much easier to find the desired category when categories are ordered alphabetically. In addition, different people will have different opinions about the relative importance or relevance of various categories, and choosing to order alphabetically avoids that problem. I do think that exceptions should be made when there is a need (e.g. in the case of eponymous categories); however, in general, I consider alphabetical ordering to be more intuitive than trying to guess the priority rating criteria used by another editor. -- Black Falcon ^(talk) 22:00, 16 October 2010 (UTC)

From what I have seen the vast majority of WP articles have only one to five categories. The Winston Churchill example I gave above is an exception rather than a rule. Therefore, category order is not a big issue really. Up to five categories only need a glance to be checked. -- Alan Liefting (talk) - 02:06, 18 October 2010 (UTC)

I don't buy that alphabetizing helps anyone find categories, because the naming varies too widely even within the same topic, and the tags can't be piped within articles like articles can within categories. Category:Judges of the United States Court of Appeals for the First Circuit, but Category:United States Supreme Court justices. Category:United States Senators, but Category:Members of the United States House of Representatives. Category:Ohio politicians, but Category:Writers from Ohio. The starting letter is meaningless and does not even group like categories together.

The concern about differing opinions on what categories are more important is really a red herring. It really doesn't matter whether Category:Presidents of the United States or Category:Chief Justices of the United States goes first on William Howard Taft, so long as they both go before Category:Cardiovascular disease deaths in Washington, D.C. I'm more concerned with grouping categories meaningfully rather than some kind of strict ordering of importance: all education categories together, all military service categories together...something that actually makes sense of the jumble of tags rather than the randomness of whether a federal judge category starts with a J or a U. postdlf (talk) 02:40, 18 October 2010 (UTC)

The naming of the categories should also be standardised then to avoid differences between, for example, Category:XXX deaths and Category:Deaths from XXX. McLerristarr / Mclay1 05:15, 19 October 2010 (UTC)

In fact, these long lists of categories we get on some articles just say to me that the category system is overloaded and being used for things it was never originally intended to be used for - and that there ought to be developed a properly functional and useful tool for making use of all this structured data.--Kotniski (talk) 05:42, 18 October 2010 (UTC)

Eponymous categories of people

In an attempt to sort out categories requesting photographs of people (see User:People-n-photo-bot) I have come across a difficulty related to categorization. On running a program down from Category:American politicians I was getting a list of people who are or were not American politicians; for example military personnel and terrorists. This has been traced to categories such as Category:George W. Bush. Am I correct in saying that Category:21st-century presidents of the United States should only contain articles of presidents an not the Eponymous categories? --Traveler100 (talk) 08:50, 17 October 2010 (UTC)

This has been the subject of many discussions. The current situation means you are correct in what you are saying. McLerristarr / Mclay1 01:53, 18 October 2010 (UTC)

On a similar subject, what about the sub-categories of Category:Presidents of the United States, most of these are not presidents? --Traveler100 (talk) 06:02, 18 October 2010 (UTC)

It's difficult to categorise things like that. ‹The template Category link is being considered for merging.› Category:Children of Presidents of the United States are not Presidents of the United States but there is no other way of categorising them. They are related to the subject so I would say it is fair enough to keep it a sub-category. McLerristarr / Mclay1 08:49, 23 October 2010 (UTC)

Category:Children of Presidents of the United States should instead be a subcategory of Category:Presidency of the United States in my opinion, the same as Category:First Ladies of the United States. Epbr123 (talk) 09:35, 23 October 2010 (UTC)

The only reason these things are hard to categorize is that we have this stupid system where sets and topics are co-mingled in the "Category" namespace and it's not always convenient to follow the naming convention that plural names represent sets and singular names represent topics. There should be some kind of more explicit way of distinguishing between sets and topics. —Coder Dan (talk) 10:07, 23 October 2010 (UTC)

I didn't realise there was a Category:Presidency of the United States. In that case, I agree. We need indeed need to make sure we have distinguish between topic categories and list categories. I will fix the US presidents categories. McLerristarr | Mclay1 10:57, 23 October 2010 (UTC)

I agree entirely that Category:Presidents of the United States should include George W Bush and not ‹The template Category link is being considered for merging.› Category:George W Bush (which contains a multitude of non-presidents). Occuli (talk) 19:32, 21 October 2010 (UTC)

categories - real and fictional

Is there any convention on sub-categories of people (and other subjects)? I have been scanning down from Category:People by occupation to create lists of people by groups but have been getting unexpected results. For example film actors under the category Criminals. Reason turns out to be fictional people sub-categories being in the sub-categories of real people. Is this an accepted practice or should this be corrected?--Traveler100 (talk) 16:19, 21 October 2010 (UTC)

WP:Categorization of people. But how exactly is Category:Film actors nested within Category:Criminals? I don't see it. postdlf (talk) 16:24, 21 October 2010 (UTC)

thanks for the link. How it can occur: going down from criminal via assassins you will eventually get to Ninja which includes Ninja in Fiction. Similar for serial killers and a number of other types of criminals.--Traveler100 (talk) 16:30, 21 October 2010 (UTC)

Right, so Category:Fictional Foo is included in Category:Foo, but how does that nest Category:Film actors within Category:Criminals? postdlf (talk) 16:32, 21 October 2010 (UTC)

I'm still not sure which category situation you refer to. Is "Criminals" directly claimed on actors in existing categories or is the problem a result of you scanning different categories and combining the results into lists? Can you give a specific example with category and actor names? PrimeHunter (talk) 16:33, 21 October 2010 (UTC)

Looking at it more closely what I am getting is films and characters in those films (so not real people). For example under serial killer or ninja. Have checked a few of the actors in the list they are actually also convicted criminals. I have not however check all individual entries.--Traveler100 (talk) 16:53, 21 October 2010 (UTC)

The category system isn't a strict hierarchy because its subjects often aren't just one type of thing. Category:Foos may contain both individual Foos and articles that are about Foo as a topic. postdlf (talk) 17:24, 21 October 2010 (UTC)

The problem Traveler100 has noticed might be linked to the tendency of editors to confuse articles and categories (see presidents topic above). I have always thought that Category:Foos should include Category:Fictional foos (and this has been the convention in most cases). Some actors will be criminals ... not quite sure how the 'fictional' question comes into this. Occuli (talk) 19:38, 21 October 2010 (UTC)

Category:Ninja should not be a people category at all (as it contains non-people) and so its parents are incorrect. Occuli (talk) 19:49, 21 October 2010 (UTC)

Douglas Sarine is at present in 'Criminals' via subcats. However this is because he is in a host of incorrect categories: eg he is not a 'fictional ninja', or a film, or a series, etc. Occuli (talk) 20:23, 21 October 2010 (UTC)

Category cycles

I have created a list of all category cycles. If anyone is interested in fixing them, see Wikipedia:Dump reports/Category cycles. Svick (talk) 16:59, 23 October 2010 (UTC)

Proposing new weak security categories.

I welcome input and assistance with my proposals: [1]. --Elvey (talk) 21:22, 24 October 2010 (UTC)

Sortkey

Currently, WP:SORT states "By convention, the first letter of each word in a sort key is capitalized, and other letters are lower case." This has led to the addition of such defaultsorts by bots and other AWB edits. I would like to question this rule. I can imagine that in some cases, having a defaultsort or other sortkey with this formula may be beneficial. However, I don't see why it has to be the rule.

• Disadvantages compared to having no enforced/automatically added/prescribed defaultsort:

1. Page moves will check the new page title for sorting, instead of the defaultsort which is based on the previous name. Looking at the recent move log, I see that e.g. Dhanwar language was moved to Dhanwar Rai language. If an automated defaultsort had been added, this page would still be sorted at Dhanwar Language. Because it has no defautltsort, it is correctly sorted at Dhanwar Rai language. Ouachita people stood for over a year at Ouachita poeple. Actual example: in this edit, AWB added the defaultsort Selce E Poshtme to the article Selcë e Poshtme. However, the article is now at Lower Selcë, with the old defaultsort kept. No auto-added defaultsort would have prevented this (or at least made it a less common problem).
2. As long as not all pages have the prescribed defaultsort, sorting in cats is incorrect. E.g. on Category:Employment law terms, "Employment testing" is given before "Employment contract", because the first was given a defaultsort by Yobot, and the second is not so blessed yet. Similarly, List of logic systems sorts before List of first-order theories because the former has been edited by Smackbot. However, as soon as all the pages have this kind of defaultsort, if ever, these pages will again have the sorting order they had without defaultsorts.

I would like a discussion about which advantages are so strong that they outweigh these disadvantages in such a way that these problems and errors are acceptable collateral damage. I currently don't see them, and would like to change this guideline (and all related ones, and AWB) so that this unnecessary and in many cases harmful rule disappears completely. Fram (talk) 18:12, 26 October 2010 (UTC)

I don't see that the rule is wrong, though it should perhaps be pointed out that there's no point in applying the rule to just a few pages in a category - if you do that, and leave the others as they are, then it will indeed (at least for a limited time) make things worse rather than better.--Kotniski (talk) 18:33, 26 October 2010 (UTC)

But if you do it to some pages in a category, you are doing it to other categories as well (since the defaultsort works for every cat of a page, and page 1 in a cat will have different other cats than page 2, and so on). Not really a workable solution. And it doesn't answer the question: assuming it worked like this anyway, how would it even make things better? Fram (talk) 18:54, 26 October 2010 (UTC)

The case sensitivity of sorting means the rule is valid to me. A bot can readily add the missing DEFAULTSORTs (they're already being done gradually), and I don't think established pages are moved often at all (though a bot could also do those ones). Rjwilmsi 19:02, 26 October 2010 (UTC)

Could you give examples where it is actually better? Your response is purely theoretical, while I have taken the time to show how it works in reality. Fram (talk) 19:22, 26 October 2010 (UTC)

Better than what? Clearly having a rule is better than having no rule (or rather: in any given category, a rule is better than no rule; and it makes things much simpler if we have the same rule for practically all categories). Are you suggesting a different rule? It could perhaps be refined (to make it more compatible with a larger number of pages for which no sort key has been defined), but at the cost of increased complication.--Kotniski (talk) 19:34, 26 October 2010 (UTC)

"Clearly"? "Clearly it is much better to look at each page individually? "Clearly" it is much better to keep the page title as the default sortkey? It seems to me to be the simplest if the rule was: leave things as they are, only add a defaultsort when you actually have a reason for it, not just adding a defaultsort because a rule is better than no rule? My "rule" decreases complication, instead of increasing it. Fram (talk) 19:58, 26 October 2010 (UTC)

As for "established pages aren't moved often at all": my examples at the start were from the latest hour at that time. Looking at the current move log, the defaultsorts would be incorrect at this move, this one, and this, and a fourth and a fifth in just over one hour time. Let's say that this is a typical hour (no idea actually), then this is 100 pages a day that need a bot edit to correct a previous bot edit (of unproven use), while without the first bot edit no second edit would be needed. Fram (talk) 19:31, 26 October 2010 (UTC)

The benefit of the sortkey is that it keeps pages like Rabbit Hole and rabbit hole grouped together in the hypothetical category of pages beginning with "Rabbit". This is because all lowercase letters in ASCII have a higher value than all uppercase letters. Without a capitalized sortkey, the category would be sorted "Rabbit...Rabbit Creek...Rabbit Hole...Rabbit Transit...Rabbit breed...Rabbit hole..." . Without changing the way categories sort, both approaches lead to items in unexpected positions. If pages really are to be sorted as if their names are capitalized, having a case-/accented-insensitive sort would be the ideal solution; unless that is implemented, capitalizing DEFAULTSORTs works to force the behavior, with the downsides of necessary cleanup and flawed sorts due to incomprehensive use. —Ost (talk) 20:58, 26 October 2010 (UTC)

I have no opposition to people adding a single sort (piped after the cat) on pages and cats where this "really" is a problem. Considering that e.g. Rabbit Hole and rabbit hole have at first glance no cats in common, the defaultsort would have no positive effect here. I know the theoretical reason for this, but have yet to see one actual example, never mind the whole series of them that would really make such a rule to be implemented everywhere worthwhile, instead of occasional, single)cat on single-page added sorts. For the moment, we have actual downsides and theoretical upsides, so that is not really a good reason for a rule that imposes this. To have a rule, there has to be a sufficiently large group that would profit from this, to counter all the work it actually produces. Fram (talk) 21:04, 26 October 2010 (UTC)

Some perspective here: Everyone else in my house uses Wikipedia as a reader, but none of them even knew what categories are, without confusing them with references. Admittedly I've used categories as an editor to find other articles that need a similar edit, but in that case the sort order doesn't matter. If someone besides editors does use categories, the "Rabbit hole" example would only matter if the list was long enough to include all those entries, perhaps hundreds of thousands. In that case, why would someone read the whole list one at a time, or am I missing something? Art LaPella (talk) 21:59, 26 October 2010 (UTC)

(edit conflict)I disagree. The "theoretical" downside is trivially obvious because of ASCII; just because I'm not as invested in the debate and I didn't take the time to find a category with the example does not mean that one does not exist. In fact, it should be harder to find examples of this instance particularly because many pages already have DEFAULTSORTs and I wouldn't want to change them to prove a point. Additionally, my example was more specific than it needed to be as any pages with categorized with mixed case would have conflated sorting; "Rabbit Transit" being sorted before "Rabbit breed" is also out of order. In short, both systems have downsides and adding DEFAULTSORTS is at least attempting to fix the problem, though it may never finish; I think the approach attempting to remedy the problem is better, though changing the sort function to meet user expectations is ideal. —Ost (talk) 22:29, 26 October 2010 (UTC)

Yes, I really can't understand why the developers haven't fixed this major software bug after all these years (you'd think getting alphabetical order right would be fairly fundamental for a reference work site, but apparently they have better things to do). But as long as the software behaves as it does, we have to work round it, and I really don't see what improvement can be suggested to the present "rule". Having no system at all would mean that pages get sorted unalphabetically. Admittedly it's quite hard, clicking at random, to find examples in practice where this would happen, but the sort of thing we mean is here, where if it wasn't for the rule, Tom and Jerry: The Movie would come after Tom Sweep.--Kotniski (talk) 05:33, 27 October 2010 (UTC)

Thank you, finally we have one example. I don't doubt that there are some more examples of this, but I debate whether we should impose a rule on 'all articles and categories, to solve a problem which only exists on a very limited number of articles, in a very limited number of categories (i.e. not all categories of these pages, but just one). Tye fact that it is easy to find examples of where the current situation creates problems, and that it is hard to find situations wheer it has solved any, indicates to me that the cure is worse than the problem. You are all trying to "fix a problem" or even to "fix a major software bug", but no one has been willing or able to demonstrate that there actually is a major problem. Why is it better to change all articles to help a few, when these few could be dealt with on a case by case basis? When someone actually notices that in that category, those two movies should be in another order, go to one of the articles, and add a sortkey after the category. Isn't that much, much, much simpler than the current situation? Note that due to the defautlsort on the Tom and Jerry Movie, Category:Animated comedy films is now incorrectly sorted. So adding a defautlsort to that one article, means that it should also be added to all these other ones, and so on, and so on. We are creating a bigger problem in trying to fix a smaller one. Fram (talk) 06:55, 27 October 2010 (UTC)

Fram I admire your drive to sort out this inconsistency. I am willing to help progress a bot request to complete missing DEFAULTSORTs (Yobot would probably be the best candidate to extend existing tasks), and if nobody else is willing I can make a request myself to deal with incorrect DEFAULTSORTs following page moves (monitor Page Moves log, check if DEFAULTSORT corresponds to that generated by old page name, if yes, set DEFAULTSORT to that generated from new page name). Depending on how the API works for the page log I could run that bot every few days or weeks. Rjwilmsi 08:14, 27 October 2010 (UTC)

Still, that would mean thousands and thousands of edits (probably millions, most pages don't have a defaultsort yet) for what? So much effort for so little benefit? Why are people trying everything to keep this rule, instead of arguing why this rule is so desperately needed? There are a few cases where it is beneficial, yes, but nothing that warrants such zeal (including it in AWB, proposing to write a bot, ...). Would dropping this rule (and dropping it from AWB a a result) make Wikipedia any worse? No. The current situation as a net result makes Wikipedia worse (some sorts would be better, more are worse, and moves are a problem). Creating a bot would improve this, but at the cost of many, many, many edits. Removing the rule and the AWB edits would prevent the problems as well, and the few cases that then aren't automatically helped by the proposed bot can be done manually, when someone comes across it and has a problem with it. Can someone explain why this isn't acceptable? Fram (talk) 08:30, 27 October 2010 (UTC)

I see your point - there seems to be a conflict here between the ideal situation we'd like to see and what produces the best results in the short term. As an ideal, every page would be sorted right in every category, and every time a page was moved its sort keys would be updated. In practice, many pages are categorized without any thought for sort keys, and people (including me, I admit) don't generally check the sort keys when they rename a page. I don't think we're ever going to get this entirely right, but having some kind of automated process to check for missorts (perhaps more intelligent than the AWB algorithm) seems like the best way of minimizing error.--Kotniski (talk) 09:45, 27 October 2010 (UTC)

A bot that checks the page move log, and changes instances where the defaultsort (if any) matches the old name to the new name, seems to me as such a useful bot. Fram (talk) 09:54, 27 October 2010 (UTC)

Would you like me to investigate further and request approval for my bot? Rjwilmsi 10:40, 27 October 2010 (UTC)

Depends on what you are discussing. A bot to add defaultsorts? Please no, don't. A bot as outlined by Kotniski and supported by me, to change defaultsorts which are incorrect after a page move? Fine, absolutely. Having this bot makes some of the most acute problems of the current situation obsolete, but that doesn't mean that there suddenly is a good reason for mass- or default additions of defaultsorts through AWB or bots, or for the mandatory format of them as expressed in this guideline. A bot that solves problems is not a reason to continue the creation of said problems in the first place, barring a convincing display of a serious problem that is solved by said creation at the same time. Fram (talk) 10:52, 27 October 2010 (UTC)

If I were to create a bot it would get the list of moved pages since last run, ignore pages without a DEFAULTSORT then if the DEFAULTSORT matched the AWB-generated DEFAULTSORT of the old page title, change it to be the AWB-generated DEFAULTSORT of the new page title. That would mean editor-chosen DEFAULTSORTs would be left alone. For biography articles I'd log them and manually review the bot's proposed change prior to doing it, since DEFAULTSORT generation rules struggle to meet all the subltleties of WP:NAMESORT. That means DEFAULTSORTs would be in Title Case, per current rule and my opinion of the correct rule. If you don't agree then the deal's off I suppose. Rjwilmsi 11:24, 27 October 2010 (UTC)

Oh, I agree, with the caveat that if the current rule would change, the bot may have to adapt as well (if e.g. the rule would change to "only add defaultsorts in cases where they are necessary to correctly sort multiple categories, then in the case of the above mentioned move of Selcë e Poshtme to Lower Selcë, the bot wouldn't change the defaultsort, but would remove it as unnecesary). Fram (talk) 11:39, 27 October 2010 (UTC)

← I agree that it seems silly to have to add and maintain DEFAULTSORT when a software tweak could fix most of the pages at once. Is there a bugzilla on this I can vote for? If the software treated uppercase and lowercase equally, and treated letters with accents as their non-accented counterparts, DEFAULTSORT would only be needed for people (to sort by lastname first) and a few other limited cases. –xeno^talk 13:46, 27 October 2010 (UTC)

There's a very old bug on bugzilla with lots and lots of votes (particularly from non-English users, as the current system totally messes up their alphabetical orderings) - last I heard some devs were thinking about possibly thinking about starting to solve it.--Kotniski (talk) 14:55, 27 October 2010 (UTC)

Very old is an understatement! Template:Bug voted for. –xeno^talk 14:57, 27 October 2010 (UTC)

I have a collapsed thread on my talk page which is about sorting out out of orders on a category by category basis. This would only make needed changes. it would have been done by now but I have been rather busy. Rich Farmbrough, 15:36, 27 October 2010 (UTC).

800k pages are about persons and another large number has special characters. So, maybe adding DEFAULTSORT to the remaining pages (excluding those that are only one word without any internal capitalisation) isn't a big task afterall. I can fill a BRFA if we conclude here. Rjw, maybe you can give us some statistics of how big this task would be. -- Magioladitis (talk) 07:25, 28 October 2010 (UTC)

No real statistics, but some 50,000 or so "List of" articles, thousands and thousands of plants (e.g. nearly two thousand articles in Category:Plant pathogens and diseases, plus 23 subcategories, but also 135 in things like Category:Anthurium), numerous articles in anatomy (e.g. 130 solely in Category:Arteries of the head and neck. We have seven articles starting with "intraocular". None of them have a capitalized second word. Some of them (at least Intraocular pressure) has an auto-added defaultsort, putting it out of order in categories. "Correcting" the other six is a pointless method of solving a problem: when one is wrong, the other ones shouldn't confirm to it, the wrong one should be corrected. People seem to be running towards a bot to tag them all, without even thinking why. Fram (talk) 07:47, 28 October 2010 (UTC)

My summary so far:

• The best solution is making sorting case insensitive, but this has been requested a long time ago, and it will probably not happen in the near future.
• The most obvious problem with the current system, incorrect defaultsorts after page moves, can possible be solved by a bot that Rjwilmsi has kindly offered to create and run (under certain conditions).
• Apart from this, we have categories where an added defautlsort (according to the current sort rule) solves sorting problems, and we have categories where it created such problems. Solutions (apart from my first point) are either to add a defaultsort to all pages (which would mean many thousands of edits, most of them not needed at all), or no longer adding defaultsorts by default and as a rule, but only manually in categories where they are needed (i.e. either as a defaultsort, or as a category-targeted sort). The auto-adding is the easiest but also in many cases the most pointless: the second requires more thinking and manual work, but will reduce the number of unnecessary edits (and make AWB and bots with the same functionality a bit lighter, as one rule/task less has to be done!).

If people consider this a fair summary, how do we proceed? Fram (talk) 07:33, 28 October 2010 (UTC)

Statistics: I gave the figures elsewhere for the number of correctly DEFAULTSORTED items, according to the capitalisation rules only (about 1.8 million I think?). I don't have the figures for the number that naively need or do not need a DEFAULTSORT, according to the capitalisation rules, but I think the first is hundreds of thousands only. The number that have an incorrect DEFAULTSORT, again according to capitalisation rules only is the low tens of thousands. Rich Farmbrough, 19:51, 30 October 2010 (UTC).

Category:Works by date as introductions

It seems to me that this category hierarchy should be connected to the Category:Introductions by year hierarchy. I just suggested this at Category talk:Works by date, but I guess the chances of getting any feedback on that thought is considerably greater here. __meco (talk) 11:54, 27 October 2010 (UTC)

User:Xrayburst1/Stellar Explosion

Can someone help me with User:Xrayburst1/Stellar Explosion ? I tried to decategorize it, per WP:USERNOCAT, but I was cited for vandalism when I did that. 76.66.203.138 (talk) 07:30, 28 October 2010 (UTC)

 Done Thanks Fram. 76.66.203.138 (talk) 07:57, 28 October 2010 (UTC)

RfC on Sortkey issue

Currently, WP:SORT states "By convention, the first letter of each word in a sort key is capitalized, and other letters are lower case." This has led to the addition of such defaultsorts by bots and other AWB edits. The topic of this RfC is whether this should be a hard rule, implemented on all articles, or a suggestion, only implemented on those articles where it is really an improvement. Fram (talk) 10:45, 2 November 2010 (UTC)

Discussion

This issue has been discussed and described at the section "Sortkey" on this talkpage. Basically, it boils down to this:

1. Some pages are incorrectly sorted when they don't have a sortkey added, since sorting is case sensitive and we want it often to be case insensitive. This is the case when you have multiple pages in the same category that start with the same word, but where some have a following word starting with an uppercase, and some with a lowercase.
2. Some pages are incrrectly sorted when the defaultsort is added to it, and no defautlsort is added to another page in the same category with a name starting with the same word. Before any defaultsort was added, these were sorted correctly.
3. Adding defaultsorts to all pages would solve issue 2, at the cost of hundreds of thousands of edits which would not improve anything in most cases
4. Defaultsorts are a problem when a page is moved: without the defaultsort, the title is automatically the sortkey, meaning that a moved page gets sorted under its new name. With a defaultsort, such a page would stil be sorted under its old name. A bot has been suggested to solve this. Fram (talk) 10:55, 2 November 2010 (UTC)

Examples for the above issues are given in the mentioned "sortkey" section, more examples can be provided if wanted.

My suggested solution is:

1. Change the rule to a suggestion, that if a page actually gets incorrectly sorted, add a sortkey to the actual category of the page that is a problem, like this: Sort With Upper Case.
2. Leave all other pages alone: if it isn't broken, don't fix it.
3. Following this, leave out the addition of such defautlsorts from AWB and other automated or semi-automated edits. Fram (talk) 10:55, 2 November 2010 (UTC)

Addition (after comments below made it clear that the proposal wasn't sufficiently explicit): defaultsorts for biographical articles should not be affected by this proposal, and should remain: these are in general very useful. Fram (talk) 14:21, 2 November 2010 (UTC)

Comment on the assumptions above.

3. No one is suggesting adding DEFAULTSORT to all pages. 86% of titles in main-space would not need DEFAULTSORTS for case insensitive sorting (excluding diacritics).
4. Very few pages with a default-sort are moved, less than 100 per day, of those a significant number either

• do not need a change (http://en.wikipedia.org/w/index.php?title=Carrier_sense_multiple_access_with_collision_avoidance&action=historysubmit&diff=393975613&oldid=383690234 ),
• are still better sorted than if they didn't have a DEFAULTSORT (e.g. http://en.wikipedia.org/w/index.php?title=London_Film_Critics_Circle_Award_for_Director_of_the_Year&action=historysubmit&diff=394354616&oldid=351097286 )
• were equally incorrect before the move (e.g. http://en.wikipedia.org/w/index.php?title=Carl_Eduard_Hellmayr&action=historysubmit&diff=394353036&oldid=394352770 )
• or are fixed by the mover (same example).

Rich Farmbrough, 13:41, 2 November 2010 (UTC).

User:Rich Farmbrough/temp111 contains a list for 6.2 days if anyone wishes to analyse it. Rich Farmbrough, 14:06, 2 November 2010 (UTC).

E.g. Kidnapping of Jalal Sharafi, clearly incorrect defaultsort after move. Black-lored Parrot had a defaultsort added through AWB: while the page move hasn't really changed this, it is one of three black-x birds sorted out of order at Category:Birds of Indonesia, where 25 others are not sorted by this rule. Not having a defaultsort would have been better in this case as well. The same goes for different other moved bird articles from the list as well. Ulnar collateral ligament of thumb was sorted before Ulnar carpal collateral ligament before the move, and still is afterwards.

Basically, looking through that move log, and ignoring articles where this RfC wouldn't make a difference (mainly articles on persons), I see multiple articles where not having a defaultsort would be better, and none where the defaultsort actually made an improvement. Fram (talk) 14:21, 2 November 2010 (UTC)

Oh, and I have reverted your AWB change here, since you changed a defaultsort that editors had added specifically to be different from the article title, to one matching the article title, thereby ruining the prupose of their defaultsort. Please, if the pre-move defaultsort didn't match the article title, there is no need to make it so post-move, as one may think thta the difference was done deliberately... Fram (talk) 14:37, 2 November 2010 (UTC)

I analysed over 100 and only 4 had defaultsorts that would have been put on by AWB before the move, and were wrong afterwards. Rich Farmbrough, 01:02, 4 November 2010 (UTC).

And how many had defaultsorts that would have been put on by AWB, and were not wrong afterwards? Not 96%, obviously, as these incldue many biographies... And of those that would have been put on by AWB, how many made an actual positive difference to the sorting of the article in even one category? More than those 4%?Fram (talk) 12:58, 4 November 2010 (UTC)

Not all pages, but many hundreds of thousands of pages. 14% of the titles still means close to 500,000 pages.

About your page move examples: thanks for providing such a nice example. The Category:London Film Critics Circle Awards lists four specific awards. If none of them had a defautlsort (my preferred state), all four would sort correctly. Currently, due to changes you made to them with AWB[2], they have a defaultsort which hasn't improved any sorts. However, if someone would (quite understandable) change the defaultsort for London Film Critics Circle Award for Actress of the Year from the current, pre-move "DEFAULTSORT:London Film Critics Circle Award For Best Actress" to the correct "DEFAULTSORT:London Film Critics Circle Award For Actress Of The Year", it would no longer sort correctly. So this is, contrary to your claims, a series of articles where no defaultsort would have been better; they would all have sorted correctly, and no changes were necessary after the move either.

The first one you provide, "Carrier sense" etcetera, is currently, thanks to an unnecessary defaultsort, incorrectly sorted at Category:Channel access methods: the two articles that received a defaultsort through Smackbot edits are now sorted before the main article, which doesn't have a defaultsort. If Smackbot hadn't added the two defaultsorts, these articles would sort correctly, thanks to the page move, which now indeed had no effect on the sorting.

Your examples three and four are about a person, where there is no objection whatsoever to adding a defaultsort in nearly all cases.

As was discussed in the previous section on this: can you provide some insight into why this should be a hard rule for every article that matches this title description, and not a soft rule to be implemented only when it is really improving things? Fram (talk) 14:08, 2 November 2010 (UTC)

• If Template:Bug were fulfilled, a great deal of of the DEFAULTSORT would become unnecessary. We should try and have the bug fixed and eliminate the need for these edits. –xeno^talk 13:59, 2 November 2010 (UTC)
  • Yes please !!vote for bug 164. Rich Farmbrough, 14:06, 2 November 2010 (UTC).
    
• I see no reason why every page should not have a sortkey by default. Debresser (talk) 14:29, 2 November 2010 (UTC)
  • What would be the benefits of this? Disadvantages plenty: first a run through all 500000 pages or so that would need one with this rule: then new bots or new maintenance (backlog) cats, with "pages missing a defaultsort" and so on, since new pages will not automatically get a defaultsort, making it necessary to go through these pages every day: all this for what prupose exactly? What is the ratio of pages improved by this, vs. pages needlessly edited by this? Fram (talk) 14:37, 2 November 2010 (UTC)
    • > 0.6 Rich Farmbrough, 12:57, 4 November 2010 (UTC).
      
      • Thanks, I think... How did you get this figure? Sample? Thin air? Fram (talk) 08:31, 5 November 2010 (UTC)
        • No I compared the number of pages that naively need a DEFAULTSORT with the number of OOO's (out of orders). Incidentally I don't support a full DEFAULTSORT run, although I am not against the idea, it seems unnecessary at this time, despite the massive advantage of simplicity. Rich Farmbrough, 08:44, 12 November 2010 (UTC).
          
• Going through every category to find pages that are sorted incorrectly would be harder than using a bot to go through every page and add an automatic sort key. Problem 4 isn't a problem with the sort keys, it's a problem with editors. Editors should just change the sort key when the page is moved. Simple. McLerristarr | Mclay1 08:10, 5 November 2010 (UTC)
• Yes it's harder but I have code to do it on dumps already. This might help those who don't want to see a single un-necessary edit. Rich Farmbrough, 08:44, 12 November 2010 (UTC).
  
• I propose that we add DEFAULTSORT to every single page. 1/3 of Wikipedia pages are biographical and a big percentage uses special characters. Adding DEFAULTOSRT to the rest won't be a big problem as soon as we have a good estimate of how many pages we are talking about. After we finish adding DEFAULTSORT to all pages we can just have a bot to check if pages moved daily have the correct sortkey. -- Magioladitis (talk) 12:58, 5 November 2010 (UTC)
  • ...and a second bot to add defaultsort to every new page (and every page moved from another namespace to the main namespace) as well (not intended as criticism of your opinion, just an indication of what is needed if we indeed go this way instead of my way (which is rather lonely so far, sob sob ;-)).

There are two BRFA's in already from me and Rjwilmsi, to deal with moves that need changing. Adding new pages would be a cinch I think. Rich Farmbrough, 08:44, 12 November 2010 (UTC).

Moving forward

Seeing this has stalled for 2 weeks now and two BRFAs by an editor involved are waiting approval, I will attempt to at least partially move this issue forward.

So: Should DEFAULTSORT be added to articles where it does not impact the actual sorting of the article? Hypothetical example: add DEFAULTSORT to ReD as "Red" when it is already sorted correctly between Rack and Ruth.

• Oppose. Don't fix what's not broken. A bot/script/AWB can (be modified to) determine when adding a sortkey will actually change the sorting order. —  HELLKNOWZ  ▎TALK 13:18, 28 November 2010 (UTC)

Village pump: "Is an unsourced catagarisation of a person as belonging to a particular ethnic group violation of WP:BLP policy?"

Discussion in progress here. postdlf (talk) 18:03, 7 November 2010 (UTC)

Category:Actors by religion

I think that, since we have the Category:Jewish actors, it's only fair we create (or re-create) categories for actors of other specific religions:

• Category:Christian actors
  • Category:Roman Catholic actors
• Category:Atheistic actors
• Category:Hindu actors
• Category:Muslim actors

Also Category:Comedians by religion:

• Category:Christian comedians
  • Category:Roman Catholic comedians
• Category:Atheistic comedians
• Category:Jewish comedians (Already have that one)

...Not sure if there are enough Hindu or Muslim comedians out there to qualify...

If Jews are specified by their occupations, why can't other religious people? Carol Anne Freeling (talk) 01:26, 14 November 2010 (UTC)

Because Jews see themselves as a race rather than just a religion. We could also have Buddhist and Scientologist but I think Roman Catholic is getting a bit too specific. McLerristarr | Mclay1 04:31, 14 November 2010 (UTC)

However, the creation of these categories must meet WP:Overcategorization#Non-notable intersections by ethnicity, religion, or sexual orientation. McLerristarr | Mclay1 04:39, 14 November 2010 (UTC)

Please read the deletion discussions before recreating deleted pages. Category:Christian actors should be deleted as it's pointless and Category:Christian comedians should only be kept for comedians whose religion has something to do with their comedy, but I can't think of an example (anti-religion comedy is much more common). McLerristarr | Mclay1 04:44, 14 November 2010 (UTC)

See Wikipedia:Categories for discussion/Log/2007 January 12#Subcategories of Actors by religion and Wikipedia:Categories for discussion/Log/2009 June 14#Category:Jews by occupation. PrimeHunter (talk) 05:06, 14 November 2010 (UTC)

These categories should all be deleted as completely unnecessary invitations to breaking BLP, and certainly none should be recreated per WP:OTHERSTUFF. Hans Adler 08:03, 14 November 2010 (UTC)

I must concur. Actors' religion is an extremely non-defining characteristic for being an actor. __meco (talk) 08:29, 14 November 2010 (UTC)

• As for the Jewish categories, the correct reason is given above. But as to Chrisians and other, I don't think we should have them. Then again, since we have "Christian films", why shouldn't we have "Christian actors" as well? :) Debresser (talk) 10:27, 14 November 2010 (UTC)
  • It would have to be Category:Actors in Christian films for it to be related, but I certainly do not think that category is necessary. McLerristarr | Mclay1 11:28, 14 November 2010 (UTC)

That's what I meant. :) Just joking. Debresser (talk) 11:38, 14 November 2010 (UTC)

I agree with the above reasons for deleting these; was there no reason for recreating them other than "Jews have their own categories"? That argument was well countered. I've speedy deleted those that were recreations of previously deleted categories, and just listed (the improperly named) Category:Atheistic actors for deletion. The comedian by religion categories should also all be listed (none of those were recreations, right?). postdlf (talk) 19:19, 14 November 2010 (UTC)

Categories applying to musical works

Two weeks ago I created various sonata categories and applied them to dozens of pages that detailed relevant sonatas. This morning, most of my changes were wholesale reverted (see edits on the 18 November here—most with an edit comment similar to "cat is for a specific musical work, not for its composer)"). Some background to this issue:
I believe there to be three levels of musical work pages:

• Level-1: works listed on a composers' page (e.g. Stanley Hollingsworth)
• Level-2: works listed on a dedicated "list of composer's works" page (e.g. List of compositions by Carl Philipp Emanuel Bach)
• Level-3: specific pages devoted to individual works by a composer (e.g. Oboe sonata in C minor (HWV 366)—a page I created for a work by Handel)

In my experience, the more "notable" a composer, the more likely they are to have level-2 and certainly level-3 pages. For the rest of this discussion, I'll stick to the example of "oboe sonata" (and Category:Oboe sonatas), but the underlying principle at stake here applies to other instruments (recorder, violin, and flute in the above reversions), and indeed other genres (e.g. symphonies). I believe that it is okay to apply the Category:Oboe sonatas tag to the lowest level available for a composer. For example, in the case of Hollingsworth above, it would have to go at his level-1 page (where "Sonata for oboe and piano (1949)" is mentioned in a list) because there just isn't a level-2 or level-3 page. In the case of CPE Bach above, the category would have to go at the level-2 page (where, for example, "H 549 \ Oboe Sonata in G minor (Wq 135)") is mentioned because there isn't a level-3 page available for the work. In the case of Handel above, the category can go at the level-3 page.
The reasons I believe it okay to put the Category:Oboe sonatas at all three levels (on a case-by-case basis) include:

1. As they stand, pages such as Stanley Hollingsworth are not just a composer page: they are a composer plus list-of-works page.
2. Listing (say) the "Stanley Hollingsworth" page on the Category:Oboe sonatas page helps to drive readers towards the location of an oboe sonata (something which the readers who have found their way to the category page are obviously interested). No matter what state articles are in, we have to focus on improving the experience for the only people who matter: our readers.
3. It is obviously desirable that we have level-3 pages for all the oboe sonata works, however I think we all have to be realistic and assume that that isn't going to happen (either ever, or for a very long time for the less notable composers). In addition, I'm certain that there would be "notability" issues if level-3 stub pages such as Sonata for oboe and piano (Hollingsworth) were created (which is the only avenue possible for applying the Category:Oboe sonatas according to the above reverter).
4. There is no time limit at WP. By that I mean that it is fine to apply the categories to level-1 and level-2 pages now as they can be removed when appropriate later. When all the level-3 oboe sonata pages are created for a composer, then the aim would be to remove the category from the level-2 page. Similarly, if a level-2 page were to be created for a composer, then it would be appropriate to remove the Oboe sonatas category from the level-1 page. What's the hurry?
5. Having the categories on the level-1 and level-2 pages means it is more likely to attract the attention of editors—who might then create the level-3 pages.
6. Categorization is never going to be an exact science. I maintain that it is better to include level-1 and level-2 pages in categories than it is to not categorize musical works (without the categorisations I created, the oboe sonatas would be much harder to find). In other words, it doesn't have to be neat—just functional.
7. Having the Category:Oboe sonatas page at the bottom of the Stanley Hollingsworth page is not a distraction (to the readers of the "Stanley Hollingsworth" page).
8. In the two weeks since I applied the categories, there has been no complaints (save the sudden reversions—without discussion—this morning). Note that the lack of complaint or reversion by any other editor on any of the dozens of pages I edited includes notable pages such as List of compositions by Antonio Vivaldi and List of compositions by Carl Philipp Emanuel Bach.
9. If someone arrives at Category:Oboe sonatas and sees a link to Stanley Hollingsworth are they going to be put off by thinking that a mistake has happened, or are they likely to assume that there is a mention of an oboe sonata at the destination page? What's the real harm? Remember that there is no other mention of his oboe sonata anywhere else on WP.
10. There are no guidelines (that I can find) that say that something like Category:Oboe sonatas can't be placed on pages such as Stanley Hollingsworth or List of compositions by Carl Philipp Emanuel Bach.

Thanks for taking the time to read this. I'll be guided by whatever consensus arises.  GFHandel.   02:37, 18 November 2010 (UTC)

• If you want to achieve the objective (1) you stated, you should create a list of oboe sonatas, a la Oboe concerto. You may then list them according to style or composer, and include that page in Category:Oboe sonatas. I feel using the category on Stanley Hollingsworth would definitely be a misuse of the categorisation system. By the same token, Michael Jackson should not be tagged with Category:Pop music, even though he may have been known as 'the King of Pop', however, including him in the sub-category Category:Pop musicians is acceptable. Thus, fewer would object if you were to create Category:Sonata composers or somesuch, and included Mr Hollingsworth. Tagging per (2) is still not terribly proper use of categories because only part of the list is ever likely to belong to the category, but I think fewer are likely to object. The absence of specific guidelines only means we don't mean to be overly bureaucratic; it doesn't absolve us from using our common sense. --Ohconfucius ^¡digame! 02:53, 18 November 2010 (UTC)

---

As an example of what I did, here is the state of the Category:Oboe sonatas page after I put the "Oboe sonatas" category on the List of compositions by Carl Philipp Emanuel Bach page:

Sub-category:

• Oboe sonatas by George Frideric Handel (2 P)

Pages:

• Oboe Sonata (Poulenc)
• Oboe Sonata (Saint-Saëns)
• List of compositions by Carl Philipp Emanuel Bach

Isn't that completely obvious that CPE Bach wrote oboe sonatas, and as a bonus gives our readers a one-click mechanism to go and find them? Obviously I added other pages (including composer pages where they contained a list of works), but the principle is the same. Surely that is a useful service for our readers, and an encouragement for our editors that there is work to be done in creating individual sonata pages?
 GFHandel.   04:44, 18 November 2010 (UTC)

---

If Stanley Hollingsworth's Oboe Sonata is so obscure and non-notable that it's never going to merit its own article, or not for a very long time, then why is it so important that readers be alerted to its existence at all, particularly through unorthodox methods? If you're unwilling to create a stub, then you can take OhConfucius's suggestion and create a List of oboe sonatas and include it there. You're obviously up to speed on your oboe sonatas, so you have enough knowledge to start it off, and others will undoubtedly contribute as well. That List can be categorised with Category:Oboe sonatas, and then you've get your result. Readers will know about Hollingsworth and all the other composers of oboe sonatas. That's the big picture outcome that's sadly lacking at the moment. You could make it a sub-list of List of sonatas. Same deal for flute sonatas, xylophone sonatas, hecklophone sonatas or whatever. Jack of Oz ... speak! ... 06:15, 18 November 2010 (UTC)

"then why is it so important that readers be alerted to its existence at all"—please don't presume to judge what is important for the readers who do want to research oboe sonatas. I don't see the point of compounding the problem with an article called "List of oboe sonatas" when we have a category to handle just such information (just one more list that will suffer from problems with maintenance). Please also don't presume that I have time for the oboe sonata work—I have my own interests (and do a significant amount of research and work for WP in those areas), and just happened to stumble into the sonatas area while doing other work. My solution (detailed above) provides immediate benefit to readers and editors and stimulates the linkages between articles. I'm now worried that responses that don't address my points above are falling into the category of I just don't like it. I'm still waiting for someone to point out the guidelines that indicate that "List of compositions by Carl Philipp Emanuel Bach" (which contains oboe sonata information) can't be placed in Category:Oboe sonatas. I've seen much stranger entries on Category pages than this (something that genuinely can benefit our readers).  GFHandel.   06:37, 18 November 2010 (UTC)

It's got nothing to do with compounding the problem. The "problem" was of your making, and for the most part I've undone those edits of yours. I stopped when you objected. There are many lists on Wikipedia; by being not limited to entries that have their own articles, they can contain far more information than their corresponding categories can contain. For example, the majority of entries in List of sonatas do not have their own separate articles - but that doesn't stop us from listing them all. There is exactly the same opportunity with a List of oboe sonatas. It can be categorised in Category:Oboe sonatas; readers will find the list there, click on it, and hey presto they've got the full information on every major composer and many minor ones who wrote oboe sonatas, all in one place. That's far better than having to click on dozens of different articles to find out the details. You seem intent on fighting tooth and nail for your uniquely inappropriate solution when a far more elegant and effective one is available right now. ALL articles on Wikipedia have the potential issue that they won't be maintained properly. That's ALL articles. Why would list articles have a particular issue in this regard? -- Jack of Oz ... speak! ... 08:45, 18 November 2010 (UTC)

Regarding the suggestion of List of oboe sonatas being created (based on the principles of the List of sonatas article), I can do no better than quote someone at the Talk:List of sonatas page when he wrote in response to criticism of the page:

The article is quite incomplete and it seems too broadly scoped to try and complete it. What's the goal of the list? Every sonata ever written?

I'm a little mystified too. We list 11 of Haydn's piano sonatas, but he wrote 62! We list only a selection of Mozart's. What point is it making with these selected sonatas?

and who wrote that response? Why it was you JackofOz! So you are promoting the use of such a list in this case, but previously have stated "What point is it making with these selected sonatas"? Based on other examples, you must realise that List of oboe sonatas is always going to be incomplete, so by your own observation, it would be a bad idea.

Come on Jack, my solution is a very good one with the limited resources we have at the moment. As stated, my solution will adapt into the neatness you want over time, but in the meantime it offers a great service to our editors, but more importantly to our readers. To create more artificial constructs that will simply contain more incompleteness is not the answer (and I will support your observation above to the death to prove that).

As a compromise, how about I undertake not to use level-1 articles? For example, I undertake to first create an article List of compostions of Stanley Hollingsworth (a level-2 article) and then put the Category:Oboe sonatas on that? Obviously the aim would be to remove the category to level-3 articles (when they become available).

 GFHandel.  

OK, that was a bad example. It just shows that I have fingers in too many pies around here and my memory does have its limits. But the fact that a particular list is in a poor state right now does not mean we should abandon all lists on Wikipedia. The very point of raising the issue on that talk page was to bring it to the attention of other readers to see if it could be improved. It hasn't had much reaction to date, but I live in hope. There's still nothing preventing you or anyone from creating a list of oboe sonatas. You could take your cue from the following: Bassoon sonata, Cello sonata, Clarinet sonata, Flute sonata, Piano sonata, Viola sonata, List of violin sonatas. Not all of these are comprehensive at present but they provide at least the core repertory. I have no particular problem with you categorising level-2 articles in the way you propose (as long as you don't spell compositions as "compostions" :). -- Jack of Oz ... speak! ... 19:35, 18 November 2010 (UTC)

Thanks (and I give the undertaking to check my spelling more thoroughly). We should wait a while to see what others might say in this neck of the woods. Cheers.  GFHandel.   19:42, 18 November 2010 (UTC)

Can someone please explain why what GFHandel is trying to do is not better for our readership? It seems more convenient to me. What am I missing? Greg L (talk) 20:24, 18 November 2010 (UTC)

My basic objection was, and remains, that GFH was categorising a composer as an oboe sonata. But a human being is not any kind of sonata. Or any other kind of musical work. A novelist is not a novel. A sculptor is not a statue. Etc. This started out at User talk:JackofOz#Removal of sonata categories, which you may wish to read. -- Jack of Oz ... speak! ... 20:30, 18 November 2010 (UTC)

I do agree with JackofOz that "a human being is not any kind of sonata" (and I probably did push the boundaries there—hence my compromise proposal above), however note that I did attempt to do something sensible (given the state of a large number of our composer articles) by slapping the category on a page that was not just about a human being, but instead about a human being plus a list of that human being's works (e.g. the level-1 article Stanley Hollingsworth). The problem here has arisen because of a difference in belief about what should appear on a page like Category:Oboe sonatas. If I may, JackofOz and Ohconfucius believe that entries on that page should only point to articles specifically about individual oboe sonatas, whereas I believe that an extension should permit links to be placed on the category page that direct our editors and readers to other articles that contain information about oboe sonatas (but only when more specific articles don't exist, e.g. the level-2 article List of compositions by Carl Philipp Emanuel Bach which mentions "oboe sonata", but for which there is no level-3 article). In a perfect wiki-world, JackofOz is quite correct (but could someone please give me a yell in about 2050 when we arrive at that perfect world?).  GFHandel.   21:05, 18 November 2010 (UTC)

I support GFHandel's points, since they appear to be more useful for readers. Tony (talk) 06:19, 19 November 2010 (UTC)

WP:DIFFUSE

I've taken the liberty of creating the shortcut WP:DIFFUSE. The diffusion of very large categories -- and in particular, the difficulty new users have in understanding the purpose of our category trees -- is a constant issue here and I felt this shortcut would help point users to the appropriate section more easily. I feel this is a fairly non-controversial addition to this page, however, if others object and wish to discuss first, please feel free to undo or rollback my edit. best, Shawn in Montreal (talk) 16:52, 26 November 2010 (UTC)

• Sure enough, I see that an editor has chosen to undo my edit. Can I ask this editor, and others, to discuss this suggestion here? I'd like to see if there is consensus to re-add this, as a useful shortcut. Thanks, Shawn in Montreal (talk) 17:29, 26 November 2010 (UTC)

"Diffuse" is both a misleading term to describe the subdivision of categories and an ambiguous name for something related to categories. The shortcut should at least be called something like "WP:CAT DIFFUSE", and my preference would be to get rid of the word "diffuse" entirely. —Coder Dan (talk) 17:38, 26 November 2010 (UTC)

Why is it misleading? It's the term Wikipedia uses, and I don't know of any other use of "DIFFUSE" on Wikipedia, so the shorter shortcut seems entirely reasonable. Do you have a better term you'd like to try and bring into use?--Kotniski (talk) 09:32, 27 November 2010 (UTC)

> Why is it misleading?

Diffusion is the process of distributing multiple objects over some area. That has nothing to with creating subcategories.

> I don't know of any other use of "DIFFUSE" on Wikipedia

Not now maybe, but who knows what use might arise in the future. It's harder to change bad names later than to select good names in the first place. "It seems OK now" is short-sighted thinking. Shortcut names exist in one global namespace, so they should be chosen carefully.

> Do you have a better term

"Subdivide". I would replace this text:

• a large category will often be broken down ("diffused") into smaller, more specific subcategories

with this:

• a large category will often be subdivided into smaller, more specific subcategories.

The shortcut could be called something like "WP:CATDIVIDE".

—Coder Dan (talk) 16:29, 27 November 2010 (UTC)

I don't see that either term is better or worse, for a concept that doesn't quite fit either of them (it kind of is "distributing multiple objects over some area" - the objects being the pages, and the area being the space of subcategories), but then, I don't see why both shortcuts can't co-exist.--Kotniski (talk) 18:35, 27 November 2010 (UTC)

> "distributing multiple objects over some area" - the objects being the pages, and the area being the space of subcategories

That's not what the section says. It says "A category may be diffused", not "the articles may be diffused". It also says "some members are placed in subcategories, while others remain in the main category". All members of each subcategory are also members of their ancestor categories, but "others remain in the main category" implies that articles in subcategories are not members of the main category, which is incorrect.

The whole concept of diffusion is based on a fundamental misunderstanding of what the category hierarchy represents. The idea of diffusing articles treats subcategories as being completely separate from their ancestors, as though members of subcategories are no longer members of the parents. But of course they are, so nothing really gets diffused in the sense of being moved out of the parent. The articles may not be listed in the parent, but delisting articles doesn't mean they're not members of the parent. Articles can be listed in multiple category pages, so listings aren't really objects that can be diffused. —Coder Dan (talk) 21:49, 27 November 2010 (UTC)

Well, technically they are not members of the parent (and indeed in some cases - though not in the "diffusing" case - members of subcategories don't even logically belong to the parent), so that view is not so much a misunderstanding as an alternative understanding. I don't have a problem with changing the language to use "subdivide" instead of "diffuse", but just to help Wikipedians understand each other, we should note that when people talk about diffusion of categories (as they sometimes do) this is what they mean.--Kotniski (talk) 10:25, 28 November 2010 (UTC)

> technically they are not members of the parent

Kotniski, you really need to learn something about categories before trying to contribute to this page. You apparently know nothing about them at all. A category is a set, and a subcategory is a subset. Technically speaking, all members of all subcategories are also members of their parent and other ancestor categories. —Coder Dan (talk) 17:28, 28 November 2010 (UTC)

Please be civil. Words can have more than one meaning, particularly on Wikipedia - here, a category is a particular type of software object, which doesn't necessarily have the same properties as the "categories" you know of from the outside world. And here, members of subcategories are not automatically members of their parents, either technically or (necessarily) logically. --Kotniski (talk) 19:06, 28 November 2010 (UTC)

> Please be civil.

I am being civil. If I were being uncivil, I would refer to you and your comments with words like "idiot", "troll", and "bullshit".

> Words can have more than one meaning.

This is English Wikipedia, so any words used in it need to be either common English or clearly and explicitly defined. There's no language in this guideline to suggest that your nonstandard definitions of "category" and "subcategory" apply here.

> here, a category is a particular type of software object

A category is a category, and a category page is a web page. There's nothing in the guideline about software objects, and it would be stupid to define a user-oriented navigation system in terms of them.

> members of subcategories are not automatically members of their parents

As usual, you don't have a shred of reasoning to back up your comments. All you do is repeat and reword the same nonsense and gibberish. The guideline doesn't exactly contradict you, but it strongly suggest the opposite, and it has no other text to suggest otherwise:

• If logical membership of one category implies logical membership of a second, then the first category should be made a subcategory (directly or indirectly) of the second.

Anyway, I seem to be in the minority here, so I'll just repeat my comments from last year: There's not a single word of support in the guideline for Kotniski's bizarre definitions, so even if there really is some kind of consensus in favor of them, then WP:Categorization#The category system is yet another example of the confusing, incompetent writing on this page, and User:Kotniski has relentlessly resisted my attempts to fix it.

—Coder Dan (talk) 19:58, 28 November 2010 (UTC)

Sorry, but it's you who's talking nonsense. The Mediawiki software developers have provided a certain type of functionality using a set of objects that they, the software and everyone else refers to as "categories". The functionality is not particularly great, and the way Wikipedians have come to use it is a bit weird in places, but this is the reality that this page attempts to describe. If you can improve the wording, then go ahead, but what you write must reflect what actually happens and (preferably) the terminology that is actually used.--Kotniski (talk) 08:15, 29 November 2010 (UTC)

> Sorry, but it's you who's talking nonsense.

No, it only seems that way because you're just as confused as everyone else. The software objects represent nodes in the category hierarchy, not entire categories of articles. What actually happens in practice is exactly what I've described: Generally speaking, articles in subcategories are also members of their ancestor categories in the normal English sense of the word. For example, romance novels are novels, and romance novels are romantic fiction, even though they're listed in ‹The template Cat is being considered for merging.› Category:Romance novels and not ‹The template Cat is being considered for merging.› Category:Novels or ‹The template Cat is being considered for merging.› Category:Romantic fiction. The word "diffuse" incorrectly suggests that listing articles in subcategories somehow moves them "out" of the parent. There's nothing "hyper-technical" about using normal English, and there's no reason a more sane guideline couldn't work for all wikipedians. Anyway, I'm sick of arguing with you and Good Ol’factory about this, so I've already removed this page from my watchlist. I'll respond on this thread as long as you keep making absurd arguments, but other than that I've given up on trying to talk sense with you. —Coder Dan (talk) 18:08, 29 November 2010 (UTC)

I've no strong opinion about diffuse vs. subdivide, but Coder Dan's opinion represents only one side of a longstanding ambiguity in the actual use of categories that Kotniski alludes to. Some categories represent typological qualities -- the members are all of the same type as identified by the category. Other categories represent relatedness -- the members are all related in some way to the topic of the category. Category:Paris is a member of Category:Capitals in Europe and contains as a sub-category Category:People from Paris. Quite obviously, the members of Category:People from Paris are not also members Category:Capitals in Europe, which is what is implied by what you write. older ≠ wiser 18:34, 29 November 2010 (UTC)

> People from Paris are not Capitals in Europe

The ambiguity is in the category naming system. I've suggested a couple times in the past that it would be nice to have separate Topic and Set namespaces for topic categories and set categories. Category:Capitals in Europe is named like a set (plural), but it's really a topic. A more descriptive name would be "Things related to capitals in Europe", and people from Paris do fit that description. —Coder Dan (talk) 21:29, 29 November 2010 (UTC)

That was surely not its original intention though (are all people and all events supposed to be members of the Category:Geography just because they were born in some place or occurred in some place)? A more fundamental ambiguity is what we mean by "member" - do we mean it in the technical sense (a page that has been assigned to a given Mediawiki category) or in the logical sense (a topic that logically belongs to the class that the category is intended to represent)? (Similarly "subcategory" - do we mean a category whose page has been placed in a given category, or do we mean a category that represents a subclass of the class represented by a given category?) The software rather imposes the technical definitions on us, though I believe we can and do talk about "logical members" and "true subcategories" if we want to refer to the logical versions of the concepts. (As usual, suggestions on how to improve the terminology are welcome - but I would think that any attempt to redefine the basic term "category" to mean anything other than the objects listed at Special:Categories is probably a non-starter.)--Kotniski (talk) 09:06, 30 November 2010 (UTC)

> That was surely not its original intention
Nonsense. The category system is explicitly referred to as a hierarchy.
> are all people and all events members of Category:Geography
Sure, why not? Topics are broadly defined, so they include a lot of articles. Even if that is a tenuous relationship, the intention is obviously for the category system to approximate a hierarchy of subsets and supersets. There's a long string of categories between ‹The template Cat is being considered for merging.› Category:Geography and individual people, so it's not surprising that those would be exceptions to the rule.
> what we mean by "member"
All words used in English Wikipedia should be either (a) standard English or (b) clearly defined in Wikipedia documentation. If "member" is standard English, which is obviously the case, then you're doing Wikipedia a disservice by arguing about the definition. If it's used in some other bizarre sense, which is true only in your imagination, then you've already done Wikipedia a disservice by defending such a poorly written guideline.
> The software imposes the technical definitions on us
Rubbish. The only thing the software describes is where articles are listed and which categories are related to each other. It has absolutely nothing to say about the nature of those relationships. Only a troll or an incompetent software developer would confuse those two things. Everything you've said here about software objects is sheer fantasy.
> to redefine the basic term "category" to mean anything other than the objects listed at Special:Categories is probably a non-starter
More nonsense. WP:FAQ/Categories#How_are_categories_organized? already defines parent categories as "more general groups of articles". The only non-starter on this page is trying to talk sensibly with User:Kotniski.
—Coder Dan (talk) 15:08, 30 November 2010 (UTC)

You're so aggressive and unpleasant that it's pretty much impossible to engage in any constructive discourse with you. You have some valid insights, but...well, as I say. --Kotniski (talk) 16:42, 30 November 2010 (UTC)

Engaging in constructive discourse with me is trivially easy. All you have to do is communicate sensibly and substantively. The only reason I'm aggressive and unpleasant is because you're stupid and trollish. The structure of the Wikipedia category software says nothing about what inter-category links actually mean, and all of the existing documentation clearly indicates that the Wikipedia category system is intended to approximate a hierarchy of categories in the ordinary, everyday English sense of the word. —Coder Dan (talk) 17:12, 30 November 2010 (UTC)

Did anyone say it isn't? Apart from a few word-choices, I don't think we actually disagree about anything substantial. But you must watch your language (and probably your blood pressure) - calling other editors stupid and trollish is generally not looked on kindly around here.--Kotniski (talk) 17:26, 30 November 2010 (UTC)

> I don't think we actually disagree

You said "members of subcategories are not members of the parent" and "a category is a particular type of software object", which directly contradict the existing documentation and what I've been posting. As I've said before, even if your absurd comments are just some kind of strange misunderstanding, then your communication skills are so bad that you have no business contributing to Wikipedia guidelines. As for Wikipedia's civility guidelines, it's too bad that they make it easier to be a troll than to fight them. —Coder Dan (talk) 18:24, 30 November 2010 (UTC)

• I support the continued use of "diffuse" to refer to what it has always meant with respect to categories and think a shortcut as proposed by ShawninMontreal would be useful. It's possible to worry too much about the possible meanings of how things are worded, but I think we need to avoid the hyper-technical. Not everyone reads the same information in the same way and generally we just need to go with what works for the majority of users. For years users have been misusing the word "deprecated", but it's not that big of a deal. Good Ol’factory ^(talk) 08:21, 29 November 2010 (UTC)

Not alphabetized correctly

Anyone know why Puma (AFV) is alphabetized as an "I" word on Category:Infantry_fighting_vehicles. I thought maybe a non-printing character got inserted in the title when it was created but it's alphabetized correctly at Category:Armoured fighting vehicles of Italy. Marcus Qwertyus 02:31, 29 November 2010 (UTC)

It's fixed now. It had the wrong sort key. —Coder Dan (talk) 04:32, 29 November 2010 (UTC)

{{Container category}} & {{Wikipedia container category}} & {{Wikipedia category}} proposed to be merged together

FYI {{Container category}} & {{Wikipedia container category}} have been proposed to be merged into {{Wikipedia category}} ... see Wikipedia:Templates for discussion/Log/2010 November 30 . -- 76.66.202.72 (talk) 05:05, 1 December 2010 (UTC)