Wikidata:Property proposal/beer style

‎beer style

Return to Wikidata:Property proposal/Generic

Under discussion

Description	classification of a beer based on its style
Represents	beer style (Q1998962)
Data type	Item
Domain	beer (Q44)
Allowed values	subclasses or instances of beer style (Q1998962)
Example 1	Hertog Jan Grand Prestige (Q2176488) → barley wine (Q808329)
Example 2	Coors Light (Q5168060) → light beer (Q1814622)
Example 3	Grolsch 0.0% (Q61043304) → low-alcohol beer (Q524679)
Example 4	Brand IPA (Q19884261) → India Pale Ale (Q1543214)
See also	beer bitterness (P6088), beer color (P6089)

Motivation

Beer style is a key classification for differentiating beers by their characteristics, complementing existing properties like beer color and bitterness. – The preceding unsigned comment was added by Mitsjol (talk • contribs) at 10:43, September 30, 2024‎ (UTC).

Discussion

Comment This seems related to the proposal here: Wikidata:Requests for comment/object vs design class vs functional class for manufactured objects - it sounds like your intent is to replace overspecifity of subclass of (P279)/instance of (P31) in this domain? ArthurPSmith (talk) 20:02, 30 September 2024 (UTC)[reply]
Yes, that's correct, the intent is to reduce the overspecificity of subclass of (P279) and instance of (P31) by introducing a clear "beer style" property, similar to cuisine (P2012) for food or genre (P136) for media. Mitsjol (talk) 07:52, 1 October 2024 (UTC)[reply]
Oppose on those grounds, instance of (P31) and subclass of (P279) are sufficient, we don't need special subclass items for types of object. Vicarage (talk) 17:56, 7 October 2024 (UTC)[reply]
Comment Why do we need this property for beer specifically? Why not "alcohol style" or "beverage style"? The other beer properties in "See Also" have quantity values for a specific standard, so it would make sense for them to exist. -wd-Ryan (Talk/Edits) 21:39, 30 September 2024 (UTC)[reply]
You raise good points about beverage classification. My proposal for a "beer style" property stems from the wide variety of distinct beer styles. However, I recognize the value in exploring broader categories like "alcohol style" or "beverage style." That said, using "alcohol style" for instances of vodka, for example, may not be beneficial due to its limited recognized styles. Mitsjol (talk) 09:18, 1 October 2024 (UTC)[reply]
Grolsch 0.0% is indeed a low-alcohol beer; but that is not its style, which is Pilsner. Guinness 0.0% is a stout. The two beers are not of the same style. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:42, 7 October 2024 (UTC)[reply]
I think a shared property with food would be more useful, since we have quite a few of type of food or dish (Q19861951): type of pizza (Q116392487), type of cheese (Q3546121), type of cooked meat (Q129175686). There's also type of drink (Q116861721) for drinks, but no specific classes have been created for this yet, and the metaclass itself doesn't have much use as well yet. Although I can find lots of items with description "type of cocktail" and the like, so there is a great potential to extend both branches. And I think we can distinguish basic types for all of them - sort of genres or broad groups. Well, we can use "genre" for this purpose, but I'm not sure if it's really appropriate for food. Solidest (talk) 07:16, 8 October 2024 (UTC)[reply]
subclass of (P279) could be used with all of those. No need for a property tree Vicarage (talk) 09:09, 8 October 2024 (UTC)[reply]
Sorry, I didn't word what I meant precisely. That the new shared property could be used to specify food types for branded food\beverages and such. Something that is not a specific type, but rather an implementation of a specific sort. For example, Mukuzani (Q4306822) -> sort/type -> red dry wine; Bavaria Pilsener (Q13436879) -> sort/type -> pilsner (Q152281); César Régalis (Q23211927) -> sort/type -> sheep milk cheese (Q1411808); Coca-Cola Zero (Q937783) -> sort/type -> diet cola (Q50425277). At the moment the link between brands and meal/beverage types is specified via p31 or p279 and I'm not sure if this is correct for branded reproduction of a particular type. For example, here you can see how often p31/p279 of "beer style" are used in p31: https://w.wiki/BThF . These are mostly sorts of beer used by the brands. It would be much more orderly to mark all of them with the class "beverage brand" or something like that. Using multiple types in p31 is not really database friendly, and continuing the type hierarchy through p279 with brands is definitely clogging it up. You may have hundreds of varieties of one product, for example cheese. And you may have 2-3 varieties each placed in p31 for a particular cheese product under its own name, whereas it's just a branded reproduction of a particular type of cheese. If we have hundreds of types filled randomly in p31 or p279 - it's definitely less usable than if it was just p31 = "branded food" or "branded dairy product" or "branded cheese". And at this point, the hierarchy of types and sorts is already heavily cluttered with commercial branded reproductions. Compared to other realms, it's like if you use music genres in p31/p279 for individual songs - something the Music project dealt with by creating a single metaclass. And, as I said, we can use "genre" to refer to types of branded food, but that seems like a stretch for the realm of food. Solidest (talk) 15:07, 8 October 2024 (UTC)[reply]
See the RFC mentioned at the top which explains how product_model is currently used. It works for brands without the need for extra properties. Your approach goes against that. Vicarage (talk) 20:13, 8 October 2024 (UTC)[reply]