|This is the talk page for discussing improvements to Commons:Categories.|
|Archives: 1, 2, 3|
Branching and crossing of categorization threads
I've removed this section from COM:OVERCAT (I've copied it below). If there was discussion with a consensus to change OVERCAT to add this exception, then my apologies and please let me know. Otherwise, it is problematic. First, it's very difficult to understand. I'm a lawyer, I interpret dense legalese every day, yet I do not find this particularly clear. There must be a better way to describe this than referring to "a multifactorial net with multihierarchic traits". Second, I am not sure that I necessarily agree. Just to use the same example of the regional administrative office. It isn't a given that the file should be categorized directly in the subcat for the street as well as in the main parent category for the region. I would consider that COM:OVERCAT. The better solution would be to create a subcategory for regional government buildings or for the regional government. Anyway, hopefully we can sort this out. --Skeezix1000 (talk) 18:38, 7 July 2014 (UTC)
"Note that Commons category structure (as well as Wikipedia categorization) is not a simple hierarchy (like biological taxonomy) but rather a multifactorial net with multihierarchic traits. That's why one subject can be categorized into any category by one factor and into its parent category by another factor and such categorization should be not considered as overcategorization.
- Example: a category of the Regional Office is categorized to the category of the street where the office bulding is. As regards location, the building should be not categorized directly into the category of the city or the region. However, by attribution it should be categorized into the category of the region because the office is an administrative body of the region and the building is owned by the self-governed region. Such categorization is not overcategorization even though the second category is a parent category of the first category.
- Similarly, a category of a village which is part of a municipality can be categorized under the neighbouring village by cadastral division (because it fall under its cadastral area) and simultaneously directly to the category of the municipality (because both villages are co-equal municipal parts administratively).
Exceptionally, different categorization threads can even meet each other in the opposite direction (category A is subcategory of B by one factor and simultaneously category B is subcategory of A by another factor) and can create a quasi-cycle. However, such a solution is not preferred and should be avoided if possible."
@Skeezix1000 The fact itself that the categorization on wiki projects is not a simple hierarchy but is multifactorial (use more types of relations, not only hyponymy, not only meronymy) is established and consensual enough.
However, some users don't comprehend consequences of this basal and consensual fact. We should expect that not all uploaders (users, editors) are capable to understand all consequences of this system, though they can be capable to consume their benefits. The fact that "one subject can be categorized into any category by one factor and into its parent category by another factor and such categorization should be not considered as overcategorization" follows from the consensual categorization principles - removing of this warning from the policy can cause that people which are not capable to understand the categorization system will disrupt and destroy it under their faulty sense that they rectify "over-categorization".
Surely, as you propose, such a seeming "over-categorization" can and should be effectively reduced or avoided by such a "explaining middlecategory" which indicates - in our example - what is the reason why the building is categorized under the category of the region (because the categorization tag itself does not enable to mark and distinguish a type of categorization relation). However, we should emphasize that the fact that the file is categorized into a subcategory by one categorization criterion doesn't imply that it should be removed automatically and thoughtlessly from the parent category - it can belong to it from a different categorization reason/criterion (if the parelel subcategorization structure by such a criterion is not created yet). Wee should and need to underline that true over-categorization can be judged and stated only within one categorization criterion and type of relation.
I agree with you, such a "seeming over-categorization" should be also avoided. However, it should be treated in a different way than the "true over-categorization" and distinguished from it. The true over-categorization should be treated by simply removing from the parent category in most cases (or can be deputized by a paralel flat category - by name, by ID etc.). The "seeming over-categorization" can be treated by explaining auxiliary mezzocategories, never by removing from the parent category. However, in some cases, too many categorization levels are worse problem than the "seeming over-categorization" of one or two files. --ŠJů (talk) 19:44, 3 September 2014 (UTC)
Some types of categories are meant as so-called flat categories. Such categories are not intended to be organized and to group their content into their subcategories. Most of them are hidden "technical" or maintenance categories which group files tagged with an appropriate tagging template. However, some item categories, especially categories "by name" or "by ID number" etc. (as People by name) have also less or more such a flat character. Such flat categories can be used by maintenance bots etc. Try to discuss before you create subcategories in such cases.