Wikidata

Wikidata talk:WikiProject Mineralogy/Properties/Archive/2018/06

< Wikidata talk:WikiProject Mineralogy‎ | Properties‎ | Archive
Latest comment: 5 years ago by Chris.urs-o in topic Use of instance of P31
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

Use of instance of P31

Latest comment: 5 years ago13 comments3 people in discussion

Hi! I note that currently there are over 4000 items that are said be instances of mineral (Q7946). These seem to be entirely or almost entirely mineral species. Mineral species, like diamond (Q5283), however are not instances of mineral. Instead, mineral samples, like Centenary Diamond (Q2757039), are instances of (subclass of) mineral. Mineral species is a class of individual mineral samples and hence should be marked as a subclass of mineral. See Help:Basic membership properties about the use of these properites. If it's needed to distinguish items that are about mineral species, and there already isn't some other sufficent way to do it, then possibly you could introduce a metaclass item entitled "mineral species" (cf. species (Q7432)), and mark items like diamond (Q5283) as instance of it. --90.191.81.65 14:32, 14 June 2018 (UTC)

This is new to me. You can be wrong. I can argue that mineral samples (items) are instances of mineral samples.
The tree is tectonic plates, terrains, etc. > strata > rocks > facies > inorganic non-mineral compounds > valid mineral species > mineral varities and species polytypes. --Chris.urs-o (talk) 14:58, 14 June 2018 (UTC)
Chris.urs-o this is probably the right place to discuss this - as I mentioned on Project Chat, the decision is basically on whether mineral (Q7946) should be considered a metaclass or a regular class whose instances are items. However, you've several times mentioned a "tree" with particular members - I think what you are referring to would be a subclass tree, but can you clarify if you are talking about existing wikidata items, and whether they currently fit this relation, or something else? ArthurPSmith (talk) 20:51, 14 June 2018 (UTC)
Earth Sciences:
Nickel-Strunz Mineral Classification:
IMA-CNMNC:
@Tobias1984: Should we create a new item (Valid mineral species)? Is 'Valid mineral species' an instance of taxonomic rank? Regards --Chris.urs-o (talk) 02:37, 15 June 2018 (UTC)
It didn't occur to me that statements in mineral (Q7946) might be the problem, because, after all, mineral is a kind of compound, a kind of solid, a kind of substance etc. This is what Wikipedia articles are about. So saying that it is non of that and that it's instead a metaclass would seem odd. Being a kind of mineral(s) is a common characteric of mineral groups like Q178977. If Q7946 is to be moved from subclass tree of substance (Q27166344) to subclass tree of class (Q16889133), then you would simply skip "mineral" between Q11173 and Q178977 in subclass tree?
If a metaclass item for individual mineral species is needed then I believe that introducing a new item for this purpose is a way to go. If Chris.urs-o or someone else can confirm that all current uses of "instance of mineral (Q7946)" stand for instances of (valid) mineral species (no mineral groups, no mineral samples etc. have this use), then I suppose that letting someone handy at using automatic tools replace these uses with something like "instance of [(valid) mineral species]" would be sufficent solution. This class item would be a subclass of class (Q16889133) or some other appropriate item in that tree (but probably not taxonomic rank (Q427626) as this is a concept in biology).
I suppose by 'valid mineral species' you mean specific status of a mineral species in IMA database? I don't have a strict opinion about whether this would be better as a class-item, instead of more generic 'mineral species'. If a mineral species is not approved by IMA, then would you indicate this as mineral species here? Are there such examples here currently? If there are, then perhaps the use of some qualifier could be considered instead to show specific status? This might compare to taxonomic species (Q7432) that seem to be used when species is valid.
If the purpose of bringing up Nickel-Strunz class tree is to show that particular mineral species can be subclass of some mineral group too, particular mineral species doesn't have to be a direct subclass of Q7946, then sure, I'm not arguing that. 90.191.81.65 08:40, 15 June 2018 (UTC)

subclass tree

@Chris.urs-o: I'm trying to clarify what your meaning is here regarding the trees and Nickel-Strunz (by the way do you have a direct reference that illustrates this tree as part of that classification? The enwiki page doesn't list all those divisions for instance). For example, mineral class (Q3679577) in wikidata is stated to be a metaclass, with mineral subclass (Q3965272) as a subclass, currently. However, that doesn't exactly make sense by the meaning of "subclass" - an instance of the smaller class should be a valid instance of the larger one, but I assume a "mineral subclass" is a different thing from a "mineral class"? On the other hand, mineral supergroup (Q3977918) is not declared as a subclass of anything. mineral series (Q24241569) and mineral group (Q1936581) do have the subclass relations you indicate but then if they are metaclasses they have this same issue; unfortunately they also subclass "matter" which is (through its tree) considered a class of concrete object (Q4406616) and so should not be metaclasses. But that may be wrong. So it seems to be a bit of a mess generally as it stands. The logical subclass relations I think should be not at this meta-level, but at the specific level of the mineral groupings themselves, for example zincowoodwardite (Q204305) subclasses woodwardite mineral group (Q3777866) < hydrotalcite supergroup (Q3977911) < mineral (Q7946) which seems fine, though maybe there should be a mineral subclass and class in there also? In any case, changing the instance of (P31) mineral (Q7946) to have a "mineral variety" value instead would certainly be fine. ArthurPSmith (talk) 15:10, 15 June 2018 (UTC)
changing the instance of (P31) mineral (Q7946) to have a "mineral variety" value instead would certainly be fine
This of course would be the case for uses of "instances of [mineral]" for items that are about mineral varieties (not mineral species). Are there such uses, though?
There's now mineral species (Q55076514), but this "valid" part confuses me a little. Per IMA status, is it the grandfathered (pre-IMA) species in particular, or is it also species that were later approved? Earlier I didn't note that there's this property: IMA status and/or rank (P579). Isn't "valid" in label for mineral species (Q55076514) kind of redundant to value of this property? This property also seems to be more clear about the status. Perhaps omit "valid" in label? 90.191.81.65 17:06, 19 June 2018 (UTC)
@Chris.urs-o: I think what the anon. user proposes makes sense - shorten mineral species (Q55076514) to "mineral species", to cover valid or obsolete or otherwise invalid cases, and indicate the status with a different property. Then we should convert the P31 mineral (Q7946) to be instance of mineral species (Q55076514). ArthurPSmith (talk) 17:27, 19 June 2018 (UTC)
There is always a redundancy. Mineral species: items w chemical formula need to be instance of chemical compound. Grandfathered minerals: most grandfathered minerals were approved in the List of Minerals of 2007. --Chris.urs-o (talk) 17:50, 19 June 2018 (UTC)
A discredited mineral is a 'inorganic non-mineral compounds' or a 'mineral variety' of a synonym but not a 'mineral species'. --Chris.urs-o (talk) 02:14, 20 June 2018 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata_talk:WikiProject_Mineralogy/Properties/Archive/2018/06&oldid=1554505902"
Return to the project page "WikiProject Mineralogy/Properties/Archive/2018/06".