WikiProject Manual of Style	Manual of StyleWikipedia:WikiProject Manual of StyleTemplate:WikiProject Manual of StyleManual of Style articles
v • d • e This page falls within the scope of WikiProject Manual of Style, a drive to identify and address contradictions and redundancies, improve language, and coordinate the pages that form the MoS guidelines.

	WikiProject Categories	CategoriesWikipedia:WikiProject CategoriesTemplate:WikiProject CategoriesCategories articles
v • d • e This page is within the scope of WikiProject Categories, a collaborative effort to improve the coverage of categories on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.

Shortcut: WT:CAT

Archives
Archive 1 Archive 2 Archive 3 Archive 4 Archive 5 Archive 6 Archive 7 Archive 8 Archive 9 Archive 10

Contents 1 Rank free articles 2 Categories named after musicians 3 One-item category? 4 Hard redirects 5 Question 6 Should all members of a distinguished category be a subset of the parent category ? 7 The term "list category" 8 Speedy criterion #6 9 Template:CatDiffuse 10 Category:Pairs 11 People from [Location] categories 12 Eliminate or otherwise adjust speedy rename criterion #6 13 Quick summary edits 14 Attention: Template:This is a redirect has been Nominated for deletion 15 Templates that auto-categorize 16 "Approximate" acyclic directed graph 17 Appropriate categories for a "List of..." article? 18 Cycles 19 Terminology section 20 What is a category? 21 Templates in article categories 22 Exchange of material with Help:Category 23 Category system - description too narrow

[edit] Rank free articles

Rank free articles: I for one feel that the category importance scale should be abandoned. It is subjective and misleading. In my bid for rank free articles in Wikipedia - I give some examples of rank free proposals given in the biological literature, which relates to the way that Wikipedia ranks wildlife pages. Endangered species, for example, are more often than not rated with high importance (I think always). However, there is lots of published research arguing that a conservation focus directed solely on species level of biodiversity is a dangerous premise (e.g., [[1]]). It is true that endangered species are important - but for different reasons a common species could also be important, such as the ecological services it contributes over a broad geographic area. The keystone species concept is another example of this kind of ranking. Some conservationists try to highlight the great importance of one 'keystone' species over another. However, a critical examination of this concept by some of the biggest minds in conservation studies (Mills, L.S.; Soule, M.E.; Doak, D.F. (1993). "The Keystone-Species Concept in Ecology and Conservation". BioScience 43 (4): 219-224.) suggests that it be abandoned because it is not clear what constitutes a keystone species. The review shows that keystone species is used frequently without any data to back up the claim and it is not entirely clear what research could be used to make this kind of determination. This debate is also similar to those in taxonomy - where the 'new' approach is rank free taxonomy (e.g., [[2]]). I have found the importance scale in wikipedia articles a great source of frustration for this reason. There are different reasons why some species are important - but those reasons are not always clear, especially to a wiki editor. Hence, all the literature review on ranking in the biological literature suggests that this type of approach be abandoned. Although it has its supportors - I feel that the place for ranking does not belong in wikipedia. If this ranking scale of high to low importance continues to be used - we will find ourselves in a perpetual loop of debate over what is and what is not of high or low importance and most of the ranking will be irrelevant and totally misleading. External ranking needs to be abandoned so that the reader can make a personal determination of how important the article is. However, I am not opposed to giving some articles a high importance rank if it is more generally applicable to an entire sub-set of articles. Biology, Sustainability, Conservation biology, Ecology, and Evolution are examples of articles that should have high importance - but ranking a species of salamander over a toad as high versus low importance has no place in wikipedia. How do we fix this problem?Thompsma (talk) 22:59, 18 October 2009 (UTC)

[edit] Categories named after musicians

A while ago, there was a CFD in which several categories that shared a name with musicians were deleted. The rationale given for deletion by the original nominator was "overcategorization." There was no clear consensus in the original deletion discussion to keep or delete. But the closing admin stated simply "the result was delete" with no further explanation (a violation of WP:JUSTAVOTE). Following this outcome, a filed a deletion review in which I brought this up. Still, comments were evenly split. I always assumed that when there is a deletion discussion, and consensus cannot be reached, that it should default to keep/restore. I gave several reasons for keeping, most notably from this guideline:

Every Wikipedia article should belong to at least one category. Similarly every category (except Category:Contents, which is the root of the hierarchy) should be placed in at least one parent category.

Given this, in general, every musical artist for whom there are articles on at least several of their albums and at least several of their songs should have a category sharing their own name to serve as a parent category. Who agrees? Who disagrees? Shaliya waya (talk) 16:30, 25 October 2009 (UTC)

• disagree. See WP:OC#EPONYMOUS. Good Ol’factory ^(talk) 20:54, 28 October 2009 (UTC)

[edit] One-item category?

Is there any policy regarding categories containing only one item? I can imagine the potential for expansion plays a role, but having only one applicable item in the meantime doesn't seem to warrant such a navigational tool or look like much else than overcategorization.  Mbinebri  ^{talk ←} 23:30, 25 October 2009 (UTC)

Where this single-item category is part of a wider scheme of categorisation then it is usually fine. The applicable policy, WP:OCAT#Small_with_no_potential_for_growth, will allow for categories which are "...part of a large overall accepted sub-categorization scheme". cf Category:Flags of Nepal which, as part of Category:Flags by country, has survived Cfd a couple of times now.

Xdamr^talk 23:52, 25 October 2009 (UTC)

I can see why the Flags of Nepal would stay, but what I was referring to isn't part of a larger scheme. To be more specific, I was asking this question in relation to a category created to house an artist's one album. It seems like overcategorization to me to create a category for a topic with nothing to navigate between. I was thinking there might be a policy saying in such an instance the category should wait to be created until there's use for it (i.e., at least two items to include in it).  Mbinebri  ^{talk ←} 02:58, 26 October 2009 (UTC)

Yes There is consensus to retain all "X albums" categories, even if they only have one member. For right now, I guess you'll have to take my word for it, as I do not have a citation immediately available. —Justin (koavf)❤T☮C☺M☯ 21:18, 28 October 2009 (UTC)

[edit] Hard redirects

The RFC on switching to hard redirects is now open; please see WP:Hard category redirects and discuss on that talk page.--Kotniski (talk) 08:47, 26 October 2009 (UTC)

[edit] Question

Koavf and I would like to know if adding Category:Monotheistic religions to Yazidi would be redundant because Category:Yazidi is already a subcategory of Monotheistic religions, or if this would be relevant because of it being a distinct subcategory as by the Wikipedia:Cat#Duplicate_categorization_rule? Debresser (talk) 20:33, 28 October 2009 (UTC)

The same question would be relevant for Christianity and Bahá'í Faith inter alteres (with their respective categories, of course). Debresser (talk) 20:36, 28 October 2009 (UTC)

Wikipedia:Cat#Eponymous categories seems to apply: "an article should not be excluded from any list category on the grounds that its eponymous category is made a 'subcategory' of that category."

What is a "list" category? Debresser (talk) 20:53, 28 October 2009 (UTC)

It says at the start of the guideline. It's a category that lists items of a particular type (so "religions" in this case), as opposed to items related to a particular topic.--Kotniski (talk) 20:54, 28 October 2009 (UTC)

My take It is not clear to me if (e.g.) Christianity belongs in Category:Monotheistic religions, or if Category:Christianity belongs in it, or both. Am I just dense? —Justin (koavf)❤T☮C☺M☯ 21:13, 28 October 2009 (UTC)

My understanding of the guideline is that both the article and the category could be part of Category:Monotheistic religions (provided, of course, that Christianity is considered to be monotheistic, which is the consensus in the article currently). --RL0919 (talk) 21:41, 28 October 2009 (UTC)

Yep. Debresser (talk) 21:46, 28 October 2009 (UTC)

"List category" is an ambiguous name, because we have categories for "List of..." articles. Debresser (talk) 21:47, 28 October 2009 (UTC)

[edit] Should all members of a distinguished category be a subset of the parent category ?

From wp:categorization# Subcategorization it isn't clear if all or just some the page members of a distinguished category are supposed to be a subset of the page members of the parent category. Which is it, all or some? Also, could I have a couple of content category examples of the "perfect" or "ideal" distinguished category? Maybe I can glean the answer from the examples and learn some other things about content categories in general. Pknkly (talk) 05:20, 29 October 2009 (UTC)

In theory, yes. In practice, we're a very, very long way from it. Category:Utah is in Category:States of the United States. Not everything in Category:Utah is a state. This problem could be fixed, by putting the article Utah into Category:States of the United States, and putting Category:Utah into something like Category:United States topics by state. Is it worth the trouble? Hesperian 05:47, 29 October 2009 (UTC)

Again, this is covered by the eponymous categories section (indeed the article on Utah should be put into the category). Since these matters are causing so much confusion, it seems we have to try rewriting the guideline again to make everything finally clear.--Kotniski (talk) 09:03, 29 October 2009 (UTC)

Hear, hear. Debresser (talk) 09:29, 29 October 2009 (UTC)

That's not a "problem", because no one should expect Category:Utah to contain more states, unless it were called something like "Utah states." Obviously Category:Utah contains Utah-related articles and categories. postdlf (talk) 12:07, 29 October 2009 (UTC)

[edit] The term "list category"

We really, really need to do something about the term "list category" in this guideline, as it causes endless confusion with the concept of "category of lists". In the discussions and RFC on eponymous categories this summer, this was very clear; I just noticed that there was a late answer at Talk:Saskatchewan#Category:Provinces and territories of Canada vs. Category:Saskatchewan, the discussion that initiated the RFC in the first place, which even more clearly shows the confusion. I also see here that Debresser commented on the same thing yesterday. The term needs to be changed to avoid further confusion. "Set category" would be most logical, but I doubt most people would think of a set as in mathematics; perhaps "collection category", "item category", "membership category"? I'm sure someone has a better suggestion, but "list category" must go. —JAO • T • C 10:04, 29 October 2009 (UTC)

"Set category" works for me. But then, I'm an old mathematician. What about "group category"? Debresser (talk) 10:42, 29 October 2009 (UTC)

Yes, either works for me. Another term I've seen used in this sense is "index category". (We could also do with better terms for "distinguished" and "non-distinguished" subcategories - would "non-diffusing" and "diffusing" be any better? They seem slightly more meaningful.)--Kotniski (talk) 11:12, 29 October 2009 (UTC)

That is even more confusing. Debresser (talk) 11:49, 29 October 2009 (UTC)

I like "index category" as a replacement for "list category". If we make a change like that, we should be sure to catch it everywhere it appears, or it will cause even greater confusion. --RL0919 (talk) 13:15, 29 October 2009 (UTC)

What about "series"/"group" category, to indicate that it contains multiples of similar things—members of a series or group, as opposed to a "topic" category, which contains subtopics related to a singular thing. The distinction should be clear from the category's name: plural versus singular titling. Category:European countries is a group of countries; Category:Europe is a topic with subtopics. Subcategory members of a group category may in turn be further subgroupings, or topics in their own right; and subcategories of a topic category may also be subtopics that are singular topics or groups themselves. postdlf (talk) 13:53, 29 October 2009 (UTC)

Ok. 3-0 in favor of "group category". Now the relevant Wikipedia pages should be updated with the proper definitions and the right term in the right place. Debresser (talk) 14:31, 29 October 2009 (UTC)

Actually I said I liked "index category". But I'm fine with "group category" as well. As to pages to update, it appears that in addition to WP:CAT, updates would be needed for WP:FAQ/Categorization and WP:FAQ/Categories (and possibly others, since I have not done an exhaustive review). --RL0919 (talk) 14:46, 29 October 2009 (UTC)

I like both "index" and "group" as well. Great development on this! Thanks all, —JAO • T • C 14:54, 29 October 2009 (UTC)

I think it is a reasonable change -- but I don't think less than a day of discussion and agreement by a small number of participants is a warrant to implement the change yet. The confusion has been there for a long time, there's no need to rush. older ≠ wiser 14:57, 29 October 2009 (UTC)

I dunno. "group" is too generic. Techincally all categories group articles. I think what is trying to be conveyed is that it's a group of nouns which are of type x, rather than a group of nouns which fall under a certain topic. It would seem a rather fine distinction, depending on how it's phrased.

This is a good point. Categories by function "group" and "list". So, we may be jumping from one bad reference, "list", to another, "group". "Group" may be just as difficult to grasp as is "list" for the same reason - categories already "group". Pknkly (talk) 17:18, 30 October 2009 (UTC)

How about "grouped by topic" and "grouped by sub-type"? - jc37 15:27, 29 October 2009 (UTC)

Well, how about adapting those for brevity: "topic category" (as now) vs. "type category" (instead of list category). Thoughts?--Kotniski (talk) 10:39, 30 October 2009 (UTC)

That conveys precisely the opposite "all topic-related articles". I stick to "group category", and do not agree with jc37 who thinks that is too generic. Debresser (talk) 12:20, 30 October 2009 (UTC)

I think "group" is too generic (see above)).Pknkly (talk) 17:18, 30 October 2009 (UTC)

Suggest Subject category - I was just looking at list categories where it seems to say that "list categories" are "subject categories". So, why not make them a "subject category". I liked the advise given above by one of the editors about using nouns. Stay away from verbs or words that may be interpreted as verbs (e.g., group, index, list), especially verbs that describe what a category already does. If we use "subject categories" it would compliment "topic categories". Pknkly (talk) 17:18, 30 October 2009 (UTC)

Unconvinced - I think for most people "subject" and "topic" are synonyms, so having those two terms in opposition would be confusing.--Kotniski (talk) 17:52, 30 October 2009 (UTC)

I like "set category" - I don't think that's too mathematical, IMHO. "Index category" seems a bit vague (isn't any WP category an index of sorts?) and "group category" seems a bit nebulous too. Letdorf (talk) 19:33, 3 December 2009 (UTC).

[edit] Speedy criterion #6

Where did speedy criterion #6 come from? Debresser (talk) 11:48, 29 October 2009 (UTC)

If I recall correctly, it grew out of some repeatedly repetitious full discussions on if and when to disambiguate "Georgia" in category names. After probably dozens of discussions, I think someone proposed a general rule here or on the CfD talk page, which was supported by consensus, that disambiguations should be consistent throughout a parent–subcategory heirarchy regardless of the context of the rest of the category name. In part it was thought to be a good thing in that it would bring predictability and consistency to the naming conventions for categories. It's probably at least a year old now; I can't remember exactly when it was adopted. Good Ol’factory ^(talk) 21:05, 29 October 2009 (UTC)

Thanks. I didn't notice it till after the edits two days ago. If so, then much of the present discussions at Cfd is redundant. Actually, I've been wanting to say for a while that now that we have more participants, we also have more nominations on a daily basis. It is a good thing we should start using #4 and #6 more extensively. Debresser (talk) 21:59, 29 October 2009 (UTC)

Which discussions are redundant? We can probably just process them after two days if they have attracted no opposition. (But note that it only applies if the eponymous parent category is disambiguated; it does not apply if only the corresponding article is disambiguated but not a parent category.) Good Ol’factory ^(talk) 07:35, 30 October 2009 (UTC)

Ah, so there's a reason for the strange naming of Category:Georgia (country) at the Olympics, on which I once commented on its talk page. It still feels very funny though... —JAO • T • C 07:29, 30 October 2009 (UTC)

All those song and band names, no? Debresser (talk) 10:28, 30 October 2009 (UTC)

The ones I nominated were eponymous categories themselves, so it wouldn't apply to those. If those are renamed, though, then their subcategories for songs, albums, etc. would be eligible. Good Ol’factory ^(talk) 00:36, 2 November 2009 (UTC)

I see. It is all very confusing. Debresser (talk) 03:55, 2 November 2009 (UTC)

Couldn't No.6 now be regarded as being redundant to No.4? This would after all be a "...rename bringing a category or categories into line with established naming conventions for that category tree."? (a standard disambiguator could after all be considered a 'convention', even if that is not the traditional way we have interpreted No.4) Or is this being a little too expansive?

Xdamr^talk 04:00, 2 November 2009 (UTC)

In my opinion that would be too freely an interpretation of #4. Even I never intended such. Debresser (talk) 04:04, 2 November 2009 (UTC)

Probably best. --Xdamr^talk 04:14, 2 November 2009 (UTC)

[edit] Template:CatDiffuse

{{CatDiffuse}} has been nominated for deletion. See WP:TFD at Wikipedia:Templates for discussion/Log/2009 October 28

70.29.209.91 (talk) 19:12, 29 October 2009 (UTC)

The result of the deletion discussion was keep.

--David Göthberg (talk) 03:40, 20 November 2009 (UTC)

[edit] Category:Pairs

The just-created Category:Pairs contains Crime and Punishment, Fire and Ice, Religion and politics, and more. Is this a helpful category? Is this the right place to ask, or am I supposed to make up my own mind and take it to WP:CFD? Johnuniq (talk) 08:59, 3 November 2009 (UTC)

Probably best to take it to CFD. Doesn't sound a very helpful category to me, but some people might have different opinions.--Kotniski (talk) 09:21, 3 November 2009 (UTC)

I agree that it should be discussed at CfD. (In the meantime, I had to add pork and beans: scandalously not included.) Good Ol’factory ^(talk) 09:54, 3 November 2009 (UTC)

As can be seen by the red link above, the user who created the category asked for it to be deleted, and it's all cleaned up. Johnuniq (talk) 21:00, 3 November 2009 (UTC)

[edit] People from [Location] categories

Inclusion criteria What are the criteria by which a category of the sort (e.g.) Category:People from Chicago, Illinois may be added to an article? That article reads "The following people were born in, residents of, or otherwise closely associated with the city of Chicago and its surrounding metropolitan area." Is this normative for other such categories? Can someone point me to a policy/guideline/consensus about this matter? Thanks. —Justin (koavf)❤T☮C☺M☯ 02:54, 5 November 2009 (UTC)

Though "from" is such a wishy-washy word that just about any connection merits inclusion in practice, there should still be accuracy. An easy solution is that you remove any location category from an article that doesn't bother to mention it.

But the problem is that when the location categories are cut so finely as a single city, it's likely that they are going to be "from" multiple intersecting places given the patchwork of separate municipalities that make up most metropolitan areas. Particularly when we're dealing with notable, accomplished people, people's lives don't limit themselves to municipal boundaries (and there are even a smattering of neighborhood categories!). Where was I "from" during the two years that I lived in Arlington, VA, went to school in D.C., hung out in Alexandria, VA, and dated someone in Silver Springs, MD? All in a day's commute. If I merited an article, I'd have ten "from" categories just for where I've lived in my not-so-long lifetime. Plus probably another for my birthplace, where I never even lived. Which is one of the main reasons I really detest subnational people categories and wish they'd all be listified, but I'm unfortunately alone in that understanding. One step towards controlling it would be to categorize at no finer a level than a metropolitan area, which would address your problem here too. postdlf (talk) 18:55, 5 November 2009 (UTC)

If a normative category doesn't exist then a template for creating content category pages entitled "People from (city),(state)" should be created. It could include narrative as required by wp:Categorization#Category description. The template would automatically use narrative similar to the one used by Chicago, but it would stay clear of vague terms or phrases such as "otherwise closely associated". It should use examples and let people know what not to include within in. A clear narrative about what should be included and what should be exuded would support Postdlf's suggestion of pulling the category from articles that do not clearly mention, with a good citation, the criteria given within the narrative of the category. Pknkly (talk) 07:48, 6 November 2009 (UTC)

In my experience, the only truly effective way to change or narrow a category's inclusion criteria is by renaming it. Category tags can be added by people who have never even bothered to look at the category's description page, so if the limitation is not obvious from the name it will generally get ignored. postdlf (talk) 16:41, 6 November 2009 (UTC)

[edit] Eliminate or otherwise adjust speedy rename criterion #6

There is an ongoing village pump discussion about this here. Good Ol’factory ^(talk) 21:27, 5 November 2009 (UTC)

[edit] Quick summary edits

An edit summary for paragraphs 2/3 and 3/3 of the Quick summary section:

1. Moved the sentence A category page is a page in the Category: namespace. to the 3rd paragraph to tie in with the idea of the :colon prefix. That also gives us a chance to re-iterate that the links do not appear, which is worthy of repetition because it is a unique characteristic of the topic and needs to be learned by the reader.
2. Removed the sentence The target of this link is a page called "Category:Name", and to retain audience level, I created a link to what wikilinking is, by substituting the word "Including" (paragraph one, word one) with "Linking".
3. Included the concepts of what links here and rendering in the second and third paragraphs.

The minor changes are suitable for correction without the need to revert the multi-paragraph copy edit. — CpiralCpiral 00:51, 19 November 2009 (UTC)

[edit] Attention: Template:This is a redirect has been Nominated for deletion

Please see: Wikipedia:Templates_for_discussion/Log/2009_November_15#Template:This_is_a_redirect for a Tfd discussion related to this WikiProject. (related to the above notification)

Not positive I have the terminology correct, but believe this is whats called a 'meta-template', a template used to create other templates. The template: Template:This is a redirect was created by User:Lenoxus, a member of the Wiki_Redirect project. The template allows easily creating a specific template used to put TYPE redirects into a specific redirect category - "redirect of TYPE", e.g. "schools" and "hospitals", or "from alternative spelling" and "from alternative name". Categorizing redirects may also serve to facilitate disambiguation.

Template:This is a redirect directly supports intentionally entering the disambiguation categories indicated on the disambiguation guideline, which as a new editor here, I find straightforward. This approach both prevents overcategorization, but does not auto-categorize. -MornMore (talk) 02:54, 20 November 2009 (UTC)

[edit] Templates that auto-categorize

Feedback wanted from anyone interested in template programming and categorization.

Some templates add categories to the pages they are used on. This unfortunately also means they usually add categories to pages where they are only tested or demonstrated. So we have a how-to guide about how to avoid this problem. And some templates also need to use different categories in different namespaces, which can be tricky to do. So now we have made a meta-template that will make it much easier to handle categorization in templates. Here's the relevant links:

• {{cat handler}} – The new template. We'd like some feedback on this template and its documentation before we deploy it. Comments and questions are very welcome on its talk page.
• Wikipedia:Category suppression – The how-to guide. There's a discussion on its talk page about which parameters we should standardise for category suppression, more comments would be welcome there. By the way, some of the methods in that guide will probably become outdated once we deploy the above template.
• Wikipedia:WikiProject Category Suppression – Personally I think a WikiProject for only category suppression is a bit overkill, but the project is going to widen its scope so perhaps it is okay. See its talk page.

Oh, and please don't start a discussion here, instead use the talk pages of the above pages.

--David Göthberg (talk) 03:43, 20 November 2009 (UTC)

[edit] "Approximate" acyclic directed graph

I'm curious: How can a graph contain "approximately" no cycles? I thought "The category system" is intended to describe the desired state of affairs, rather than the "impurities" due manual labor. Spill paint over the hood of a Ferrari, and you still have a Ferrari. BTW, finding cycles looks like a job for CatBot. Regards, Paradoctor (talk) 20:58, 20 November 2009 (UTC)

I agree. The category system is acyclic and directed. Cycles indicate errors, or at least suboptimally muddy thinking. Here, therefore, we should described it as directed and acyclic. Hesperian 05:47, 21 November 2009 (UTC)

Well, I agree that's how it should be, but we can't claim that it is. Many people put categories into other categories just because there's some relation between them, not because one's actually a subcategory of the other (eponymous categories being the most obvious example). If we want to ensure there are no cycles, we must make a rule that the subcategory relationship is really a subcategory relationship, and that will upset a lot of the categorization schemes currently in use (and so will doubtless run into opposition).--Kotniski (talk) 11:42, 21 November 2009 (UTC)

We can't claim that it is, no, but I think we can safely claim that it is best practice; which is surely the purpose of this page. Hesperian 11:50, 21 November 2009 (UTC)

Well, all right, try it and see if people object. (Many people don't object explicitly but would nevertheless object to the consequences, because they don't really get it.)--Kotniski (talk) 12:12, 21 November 2009 (UTC)

I'm not sure I understand what you mean. Are you saying that some categorization schemes currently in use contain cycles by design?

I finally found out what has been bugging me: "Acyclic directed graph" is too general, the category system has the same structure as a DAWG. Paradoctor (talk) 13:20, 21 November 2009 (UTC)

Actually, it is a concept lattice. Hesperian 13:40, 21 November 2009 (UTC)

For what the guideline used to say some time ago, see the old version of the guideline (the section "Cycles should usually be avoided"). When I rewrote the page I took the passage out because it seemed to be a lot of words saying nothing (it first says that we don't like cycles, but then goes on to say that some cycles - no particular clue as to which - are acceptable). I mentioned it at talk at the time and no-one commented, but if anyone wants to belatedly reinsert that passage now, so we can perhaps work on changing it to say something more useful, then that would be fine.--Kotniski (talk) 15:09, 21 November 2009 (UTC)

Question: Would Category:Categories that do not contain themselves contain itself?

Seriously, let's declare cycles to be the epitomy of evil. Either the text will stand, or someone will come up with a cogent counter-example. Win-win. Hesperian 11:32, 22 November 2009 (UTC)

Erm, what are "meta-fields"? Judging from the text, the only acceptable loops seem to be trivial self-references of a category to itself. I'm not sure I understand the use of that. Anyway, since cycles are generally considered a bad thing, this should be explicitly mentioned on the page among the "Things to avoid". There is a problem, though: I don't think there is a way for editors to determine whether a bad cycle exists, other than by manually checking. That's tedious at best, error-prone, and not necessary. As sugggested in my initial remark, there should be a tool that detects cycles, so contributors can diagnose such a problem, and identify affected categories. This is important, as creation of cycles is the result of uncoordinated collaboration. Regards, Paradoctor (talk) 15:52, 21 November 2009 (UTC)

_{edit conflicts are fun |-()}Oy, I noticed two more: "form a hierarchical structure, consisting in effect of overlapping trees". In a hierarchy, all elements are comparable wrt level. Here's a subgraph, which can (and presumably does) exist in the category graph: A>B>C & A>D>E>C. Is B on the same "level" as D, E, both, or none? Furthermore, "overlapping trees" does not specify how the trees overlap, allowing the possibility of different roots.

I suggest we reformulate the intro thusly:

"Wikipedia's categories resemble a rooted tree, the difference being that subtrees may overlap.",

and drop the technical stuff about DAGs/DAWGs/CLs/whatever in the second sentence, or stuff it into a footnote. This is a guideline, not a course in graph theory. ;) Paradoctor (talk) 15:52, 21 November 2009 (UTC)

It does not in any way resemble a rooted tree. It is a lattice, which is completely different. Saying it is like a tree will only confuse people.

And even if we were going to go down a "They resemble a tree except..." line, it would be must easier to understand if you phrased the exception as the logically equivalent "the difference being that categories may have more than one parent category". Hesperian 11:32, 22 November 2009 (UTC)

I don't see where the second extremal element required for a lattice comes from, there is no "bottom" category corresponding to the other extremal element required for a lattice. Neither is it a lattice graph, except in the trivial sense of choosing a sufficiently high-dimensional lattice.

As regards "logically equivalent", I agree. But if it is equivalent, what makes it better? This is an expositional text, which is read mostly by people who do not share our specialist knowledge, and are mostly interested in how-to instructions. Actually, the tree metaphor is slightly misleading, maybe "bush whose branches may rejoin" conjures a more appropriate image. A diagram might be helpful here. Regards, Paradoctor (talk) 13:46, 22 November 2009 (UTC)

Well, I'm afraid I'm out of my depth now as far as the graph theory goes... "Directed acyclic graph" I could understand - can you explain why that's not the appropriate technical description? And isn't something like "set of overlapping trees" an appropriate description for the intelligent non-specialist?--Kotniski (talk) 15:50, 22 November 2009 (UTC)

To say it's a DAG is not wrong, it's just too general. It's like describing your car by saying "it's a Volkswagen". Ok, but is it a Lupo or a Constellation? With a general DAG, clicking the parent categories of a category repeatedly will lead you to different roots depending on your choices along the way. With our system, you'll always end at the same node, category:contents in our case. That's because every category must be in another category (except the root, of course). I have not yet found a proper name within graph theory for this type of graph, so rooted tree with overlapping subtrees is the best description I can come up with. The problem with the "overlapping trees" is the same as with DAGs, it is not clear that there is a common root. HTH, Paradoctor (talk) 17:19, 22 November 2009 (UTC)

Right, I see. So what about "rooted directed acyclic graph"? And "set of overlapping trees with a common root" for the second audience (though there it's not necessary to be so precise).--Kotniski (talk) 17:27, 22 November 2009 (UTC)

Errm, what you call the "second" audience is actually the one this guideline is written for. Imagine an art historian reading "rooted directed acyclic graph". She/He might understanded the sense of "rooted" used here, and can probably manage to clear "directed", but I think "acyclic" is where the fast forwarding will begin. As I said, as a footnote I have no problem with whatever precision you care for, but the "second" audience is the one we work for. If you'd like a few numbers, check out the traffic stats for the article page and the talk page, you'll see that the traffic differs a factor of ~20. This means that hundreds of people view the article each day, of which the vast majority doesn't know and care less about DAGs, rooted or not. ;) Paradoctor (talk) 18:11, 22 November 2009 (UTC)

Well yes, I'd agree with that... So to rephrase, is "rooted directed acyclic graph" OK as the description we give for the specialist (minority) audience, and is "set of overlapping trees with a common root" a good enough description for the majority audience?--Kotniski (talk) 18:22, 22 November 2009 (UTC)

Can't speak for anybody else, but my vote is o|‾|_ <utter, complete and total submission> Paradoctor (talk) 20:07, 22 November 2009 (UTC)

The people we are writing this for aren't going to care about the global structure. The number of root nodes is irrelevant because it does not affect them. It doesn't even affect me, and I've been working in categories for years. The people we are writing this for are going to care about local structure: how the category system looks close up. All that matters is that any given category can contain multiple categories, and can itself be contained within multiple categories. At this scale, the structure of our category system looks nothing like a tree, and it is only after deep consideration that ones recognise "set of overlapping trees" as an accurate description. Hesperian 23:45, 22 November 2009 (UTC)

Well, for me it seems quite intuitive given the way people often approach categories - you have an overall category (say "Villages"), this is broken down by country, then by region, then perhaps by district, forming what is obviously a tree. But the trees are seen to overlap when one category fits into more than one parent category ("Villages in X" is under "Geography of X" as well as being under "Villages").--Kotniski (talk) 08:38, 23 November 2009 (UTC)

My sentiments exactly. Also, the multiple vs. single root issue does affect users in the form of the commandment "Every Category Shalt Have A Parent". We could of course just lay down the rules and be done with it, but that's not going to work in a community of peers. There will always be some with bright ideas, or those who just don't like rules they don't understand. Paradoctor (talk) 11:01, 23 November 2009 (UTC)

[edit] Appropriate categories for a "List of..." article?

There has been some edit-warring behaviour over the last month regarding the categorization of the List of Volkswagen Group factories article. One editor (78.32.143.113 (talk)) is adamant that the article belongs in categories such as Category:Motor vehicle manufacturers of Germany, Category:Truck manufacturers, Category:Bus manufacturers and Category:Motor vehicle engine manufacturers and has reverted multiple attempts by myself and another editor to rationalize this article's categories, despite attempts to discuss the matter. My opinion, based on the existing membership of these categories and similar articles such as List of Ford factories, is that the appropriate article to include in these categories would be the one about the company itself (i.e. Volkswagen Group or Volkswagen Commercial Vehicles), and not this related "List of..." article. The only WP guideline regarding categories for "List of..." articles I can find is You can add a suitable subcategory of Category:Lists at the bottom of the page (WP:LIST), which doesn't really provide much guidance. Anybody care to comment? Letdorf (talk) 13:36, 3 December 2009 (UTC).

Well, I agree with you. A list of factories is not a manufacturer of anything, nor is it even a list of manufacturers - it clearly doesn't belong in those categories.--Kotniski (talk) 14:22, 3 December 2009 (UTC)

That depends entirely on the intended use of the category. See list-and-topic ctegories. So, check whether the categories state their intended use. If they do, you'll know. If they don't, start discussion to get consensus about it. HTH, Paradoctor (talk) 15:24, 3 December 2009 (UTC)

Of the six categories in question, three are described on the category page as list-type categories, and three have no description. However, judging by the current member articles, the implicit consensus seems to be that they are list-type categories too. Letdorf (talk) 19:28, 3 December 2009 (UTC).

Don't look at me. It's 78.32.143.113 (talk) that's cruising for fishsticks. ;) Paradoctor (talk) 20:47, 3 December 2009 (UTC)

[edit] Cycles

I found a cycle in the Category:Universe subcategories tree: Category:Universe --> Category:Cosmology --> Category:Physical cosmology --> Category:Universe. I read somewhere that cycles should usually be avoided, but cannot find this advice in the guideline. What is the best course of action? --M4gnum0n (talk) 15:50, 9 December 2009 (UTC)

I call these nested categories, and it's a big problem in articles were people places multiple categories that are related (i.e., one being a sub of another). Articles should be placed in the lowest level category possible. I usually remove the higher level category. Multiple subcategories within a category are fine if the fork is appropriately related to the article. People sometimes search for articles or subject in different ways. I should have stated that this does not mean that Wikipedia:Overcategorization should not be considered. --THE FOUNDERS INTENT ^PRAISE 16:01, 9 December 2009 (UTC)

[edit] Terminology section

I added a terminology section. It was necessary in order to keep the "Quick summary" short, and in order to understand the terminology in "The category system" section. Happy editing! — CpiralCpiral 05:18, 14 December 2009 (UTC)

Er ... it's not exactly clear. I consider myself reasonably familiar with categories and categorization, but I didn't find it at all easy to understand. Good Ol’factory ^(talk) 06:08, 14 December 2009 (UTC)

Familiarity is the problem. As a newcomer I do not find the article easy to understand. — CpiralCpiral 17:00, 14 December 2009 (UTC)

I never claimed the article was easy to understand, but I honestly didn't find this to be an improvement. Good Ol’factory ^(talk) 21:07, 14 December 2009 (UTC)

Then it was not an improvement, but the article might yet be made easier understand. — CpiralCpiral 21:05, 16 December 2009 (UTC)

Um, yeah—that was my point. It was not an improvement. Perhaps I was not blunt enough, but I didn't want to hurt anyone's feelings. Good Ol’factory ^(talk) 22:11, 21 December 2009 (UTC)

I also found these clarifications closer to obfuscations. I revert these edits for the mean time. Let's discuss it here first. Debresser (talk) 13:16, 14 December 2009 (UTC)

Here's the unclear section, with salient sentences broken-out with numbers for discussion purposes.

(I changed the last sentence to remove the word salient.)

Here are some sentences from the first real section "The category system":

1. Categories are of two basic types, topic categories, and list categories.
2. Wikipedia's categories form a hierarchical structure.
3. If the articles of one category logically also belong to another category...

Sentence 1 has reference 2. Sentence 2 has reference 3. Sentence 3 has reference 1.

May our opinions meet no urgency in the answering, but find the importance of consideration nevertheless. Thank you very much for your consideration of these matters, and I look forward to a leisurely discussion concerning the wording of this article, and how that pertains to it's accessibility, to future categorizers, and to Wikipedia.

— CpiralCpiral 17:00, 14 December 2009 (UTC)

@Debresser "referent" is correct, check out the lede of reference. The problem with that "clarification" is that it attempts way too precise language, and fails at it. In a terminology section, I expect to see a list of terms and their definitions, discussions only when a term is difficult. E. g.

• category: a collection of Wikipedia pages
• category page: a page listing all members of a category
• member: any and only those pages that contain a link of the form [[category:X]] are members of category X
• subcategory and parent: a category X may be a member of another category Y, in that case we call X a subcategory of Y, and Y a parent of X
• ... and so on

Paradoctor (talk) 17:46, 14 December 2009 (UTC)

Thank you. It's certainly easier to have category have the one referent, the category namespace referent. Now if we could only take that consistency and clarity to the article, I'd no longer own my proposal, with the aim of improving the wording of the project or adding a terminology section to match the wording of the article. — CpiralCpiral 21:05, 16 December 2009 (UTC)

Indeed, an attempt to be overly precise is often contra-productive. But I welcome your initiative, so let's see what we can incorporate in the project page. Debresser (talk) 18:24, 14 December 2009 (UTC)

I might have digested and honored the tough article if category was one, new concept with the occasional use of it's borrowed sense (from it's home namespace). — CpiralCpiral 21:05, 16 December 2009 (UTC)

[edit] What is a category?

In the category namespace category conveys the concept "a collection of pages". But according to the wording of the article,

Category:music contains a list of categories (i.e. member pages)

a category is also a member page and a subcategory. I don't think category should be used as both the container and the contained, unless we say so in a terminology section. Rather, if we adopt Paradoctor's simple terminology for the term category —this subject is a branch off the previous subject [WT:categorization#Terminology section]]— we could then be consistent (to a reasonable extent, Debrresser) with that term in the article.

This article needs lucidity concerning the word category because it is an article about categorization, and because categorization is an important craft to teach and use. The new concept for first time readers of Categorization is category. Really and sincerely, in your opinion must category mean both container and contained, both category and subcategory?.

Appropriately for a talk-page I must say category currently has four referents: (1) the category namespace one "collection of pages", (2) the category namespace one "member page", and (3) the category namespace one "subcategory", (as in If the articles of one category logically also belong to another category, then the first category is made a subcategory or as in such eponymous categories are considered distinguished subcategories (in fact they are not true subcategories at all) (4) the main namespace one wikt:category. If two, three, and four are the same, we should explain that in a terminology section in a way I, posing as a newcomer, could understand.

The very first sentence: Wikipedia's categories form a hierarchical structure could be read incorrectly as "A collection of pages forms a structure." I'm sure you would agree that if you were struggling with a new concept, that it might just as easily be written "The set of category pages forms a ... structure." I predict this will all take some time, and I am in no hurry on this. I want to volunteer to help improve this article.

Peaceful editing! — CpiralCpiral 21:00, 16 December 2009 (UTC)

<soapbox> One of the things I really like about Wikipedia is the range of interesting people I meet. :) </soapbox>

I think you're asking a question that needs not be answered here. You can use a watch without having the slightest clue as to what makes it tick. Consistency can comfortably take a backseat to what works. A version of the page that is clear to you will confuse the heck out of the majority of readers, IMHO. Exhibit #1: Good Ol’factory's and Debresser's reactions to your first edit. Exhibit #2: I must read your stuff carefully to understand it, and I occasionally edit on wikis where I can't even read the alphabet!

You might like the Harmonious Editing Club. Regards, Paradoctor (talk) 22:59, 16 December 2009 (UTC)

Thanks, Pardoctor. I'm gonna use extra pixels to be extra clear, and I'm gonna stick with your definition pf category, and Debresser's warning about "terminal" precision, and I'm sure Good Ol'factory's nose will be nearby for quality assurance. It all seems very clear and simple to me now. Happy editing! — CpiralCpiral 05:02, 17 December 2009 (UTC)

[edit] Templates in article categories

The issue of having templates as an entry in a category that is otherwise for articles is still an ongoing problem. There is no explicit guideline on it. Convention, which is occasionally ignored, is to place all templates in Category:Wikipedia templates and its subcats. Categorising templates is an administration task and they are not needed in content categories. For previous discussions see:

• Wikipedia_talk:Categorization/Archive_10#Templates_in_encylopedic_categories
• Wikipedia_talk:Categorization/Archive_10#Templates_in_article_space

I would like to have this guideline added:

"User pages and templates should not be in categories that contain encyclopeadia (content) articles."

Or suchlike. -- Alan Liefting (talk) - 19:23, 19 December 2009 (UTC)

That wording would exclude templates from "administrative" categories that contain content pages. What about making a clear distinction between "user" categories and others by requiring a (possibly invisible) {{navcat}} tag, and leaving propagation/enforcement for the subcategories to the bots? I. e. just tag and appropriate others in as {{navcat}}s, and make a bot with these tasks:

• tag as {{navcat}} every subcategory of a category tagged as {{navcat}}
• flag/zap/whatever any inappropriate content in any category designated as {{navcat}}

How does that sound? Paradoctor (talk) 19:59, 19 December 2009 (UTC)

Previous discussions notwithstanding, I find it very handy to have the template close at hand when I am browsing a category, so I oppose this proposal and any other like it on said grounds. The only exception I would make if that if there is a template category available, then obviously the template should be in it and not in the content category. Debresser (talk) 21:56, 19 December 2009 (UTC)

Could you give an example of what you mean? Paradoctor (talk) 23:14, 19 December 2009 (UTC)

[edit] Exchange of material with Help:Category

The difference between WP:categorization and Help:category determines the content and sizes. I'd like to discuss the intentions of the content of the two, and propose moving some things from here to there to get both articles to about 30kb in size. Categorization is 42kb, and category is only 24kb.

I think the namespaces are like this:
Help is hue-collar: task-oriented, how-to-ish, procedural.
WP is white-collar: managing, guiding, conceptualizing, planning, and engineering.
Categorization should be understood for the most part, and then the Help:category article shows the tools.

Indeed WP:categorization is linked more often than Help:category. I checked out what links here and what links there.

Here a section outline in this article that belongs there, and would shave off 9 kb.

 3 Display of category pages  3.1 Form of entries 3.2 Sort order  3.2.1 Using sort keys 3.2.2 Typical sort keys 3.3 Split display 3.4 Category description 

Both articles discuss sorting, and both articles discuss hiding categories.

Here is the outline for Help:Category, a 24kb sized article:

 1 Summary 2 Putting pages in categories 3 Category pages 4 Hiding categories 5 Sort order  5.1 Default sort key 6 Categories and templates 7 Categorizing redirect pages 8 Moving and redirecting category pages 9 Applying "Related Changes" to a category  9.1 Detection of additions to a category 10 Comparison with "What links here" 11 Category considerations 12 Extended tools  12.1 List of all categories 12.2 Visualizing category trees 12.3 Dynamic page list 12.4 Category intersection, union, etc. 12.5 Category flattening 12.6 Count 12.7 Retrieving category information 

There we should give-up 7, Categorizating redirect pages. There we should move 11, Category considerations to the lead section there telling them to come here first.

Here is our outline:

 1 Quick summary 2 The category system  2.1 Categorizing pages  2.1.1 Eponymous categories 2.2 What categories should be created 2.3 Subcategorization  2.3.1 Duplicate categorization rule 3 Display of category pages  3.1 Form of entries 3.2 Sort order  3.2.1 Using sort keys 3.2.2 Typical sort keys 3.3 Split display 3.4 Category description 4 Project categories  4.1 User pages 4.2 Images 4.3 Hiding categories 5 Categorization using templates 6 Redirected categories 7 Interlanguage links to categories 8 Tips  8.1 Displaying category contents on pages 8.2 Retrieving category information 8.3 Linking to categories 8.4 Searching for articles in categories 9 See also 

We should lose 5, Categorization using templates to Help. Sections 7 and 8 are also their territory. They are short, 2.7 kb.

That would make categorization 30 kb and help 35 kb, but... the redone-dancy will square, and all will be well.

Also, we should interact more with WP:classification. We could discuss WP:classification more here. We barely mention the name. The classification article is 7.8 kb, and has both theory and practice content.

— CpiralCpiral 01:04, 21 December 2009 (UTC)

[edit] Category system - description too narrow

I've had the impression before, but now it's official: The section is way to narrow in its description of applications of categories. Check out this section on on category headers. One thing that should be kept in mind is that categories are basically just sets, and seeing that almost all of mathematics can be expressed in terms of sets, it doesn't make sense to restrict categories to being navigational aids for the readers. Rather, rules for categories should be described in relation to a specific application. Paradoctor (talk) 10:47, 22 December 2009 (UTC)