Wikipedia talk:Categorization and subcategories

I think this is a bad idea because it encourages overcategorization. It encourages people to add half a dozen categories to an article, instead of the most appropriate one. This is important - the more categories an article is in, the less meaningful they become.

This is a debate that should be held some time anyway - in my opinion, certain classes of categorization are undesirable and should be avoided. In particular, any categorization by gender, skin color, or sexual or political preference (unless, of course, this is directly relevant to the person we're talking about). I do believe this obviates the problem that the "all or nothing" rule is intended to solve.

It seems topic articles are a good exception, as you state on the page. R adiant _>|< 09:48, August 23, 2005 (UTC)

Articles should be in categories that help people find the articles. Some articles should be in many categories. Some shouldn't. If you want to convince me that this is a bad idea, you have to explain what overcategorization is, why it is bad, and why this will encourage it. This proposal is trying to fill holes due to UNDERcategorization without opening up the floodgates where everything ends up in every possible related category.

While I disagree with your desire to remove any categorization by gender, skin color, or sexual or political preference, I don't believe that argument is relevant to this discussion. I just happened to use African American actors as an examples, but I could have used Academy award winning actors to illustrate the same problem. Just because Marlon Brando is in Category:Best Actor Oscar that doesn't mean that he should be removed from the parent Category:Film actors. This also is an all or nothing example.

I also want to point out that most of what I am trying to codify is already common practice. -- Samuel Wantman 07:43, 24 August 2005 (UTC)[reply]

I think the main difference is whether a category's subcategories, if put together, would form the entire category. For instance, "people" is subcatted by "profession" (and also "nationality"), and possibly subcatted further. Since every person has a profession, there should be no articles on individuals in the category:people. They should all be in the subcategories.
A category on actors could arguably be subcatted in oscar winners and non-oscar winners. However, the latter is rather silly so we don't have it. Thus, Marlon can be both in cat:actors and cat:oscar winners.
On the other hand, if we're going to subcat by ethnicity, it is obvious that every person has an ethnicity. So if cat:Americans is subcatted with "African Americans", "European Americans" and "Native Americans" (etc) then it follows that all articles on individuals should be put in the latter, not the former. R adiant _>|< 14:17, August 24, 2005 (UTC)

I think we are basically agreeing. There are some minor differences in our reasons for things, but I don't think they are worth fighting over here. I would say that categorizing by ethnicity may occasionally be useful, such as with "African Americans", that doesn't mean it makes sense to categorize everyone by ethnicity, I don't think it adds to the usefulness of the categorization scheme. And, while everyone may have an ethnicity, it often is far from obvious or clear what someone's ethnicity may be. To me that would fall into your "rather silly" category. (I find some of the subcategorization by nationality to fall into the same silliness. For instance, I don't think it makes it useful to sub-categorize all professions by nationality. Some professions are international, like Category:Film directors. Where should Roman Polanski be found? I don't think his nationality is an important distinction. But this is all besides the point.) I'm going to change the example from Halle Berry to Marlon Brando, because it will make this proposal less controversial. -- Samuel Wantman 19:40, 24 August 2005 (UTC)[reply]

more on all or nothing rule

I'm not sure the "all or nothing rule" as stated reflects widespread current practice. I think in some cases it's reasonable, like the example of Category:Best Actor Oscar (and we should perhaps discuss in another forum whether Category:Best Actor Oscar should even be a category in the first place). I think there are other cases where it would lead to massive redundancy, like the ethnicity or national origin subcats of Category:American people. I think the difference might be the relative size of the incomplete subsets to the higher level supercat. Back to the "utility" yardstick, or perhaps more related to the Principle of least astonishment, if the supercat contains a collection of "relatively small" subsets (and the subsets are incomplete) the preponderance of the articles will be directly in the supercat. In this case, I think duplicate categorization is warranted. If the subcats contain a large fraction of the articles based on a relatively obvious division I think duplicate categorization is not warranted. This may make for an "ugly" rule, not pleasantly in line with any kind of formal database view, but IMO viewing categorization as a formal database is not realistic. -- Rick Block (talk) 17:45, September 11, 2005 (UTC)

I agree with the utility yardstick. Can you think of a good category to illustrate the principle of least astonishment? How would you rephrase things? BTW, most of the members in Category:African-American actors are also in Category:American actors, which is in my opinion, the way things should be. -- Samuel Wantman 04:43, 12 September 2005 (UTC)[reply]