Property talk:P6712

Documentation

[create Create a translatable help page (preferably in English) for this property to be included here]

Datatype

Hi,

This is a very useful property that I could use (for ar (L35217) for instance), but why is the datatype lexeme and not item (which was the initial request). Why link to a (L20817) and not simply to A (Q9659)?

Both seems good but the second seem a bit easier (no need to create all letters of all languages as these items already exists), I just want to know before using it (which would require to create all letter of the Breton alphabet for which I would have no use except for this property).

@Liamjamesperritt, Jura1, ArthurPSmith:

Cheers, VIGNERON (talk) 15:27, 22 June 2019 (UTC)

PS: if this property is kept as it is now, we should add constraint to make sure it's correctly used.

• Just use it as it's defined, it should be fine. --- Jura 15:35, 22 June 2019 (UTC)
  • @Jura1: maybe but first, could you please answer my question? VIGNERON (talk) 16:40, 22 June 2019 (UTC)
•  Comment That's odd, when I supported the proposal it was using item datatype. I don't think lexeme datatype is right at all, letters are not lexemes because (1) the same letters apply to many languages that use the same alphabet (or at least somewhat overlap), and (2) there are no grammatical or sense meanings for the letters in themselves, so there's no purpose having them as lexemes. I would strongly support replacing this with an item-valued version of the property. ArthurPSmith (talk) 16:11, 22 June 2019 (UTC)
  • @ArthurPSmith, VIGNERON, Jura1: I agree that it makes much more sense to have the datatype as Items rather than Lexemes. However, Jura was insistent that it be Lexeme, and so for the sake of finally being able to use this property, I decided to yield and let the datatype be changed to Lexeme. Creating letters for each language just for this property does seem a little excessive... Liamjamesperritt (talk) 13:37, 29 June 2019 (UTC)
  • I think it's preferable over making a separate page for such letters in a given language and then asking other users to translate this unstructured content into every other language. Wikidata is here to provide structured content, not standard mediawiki pages. --- Jura 13:42, 29 June 2019 (UTC)
    • @Jura1: why? it's may or may not be preferable and I may or may not agree with you but beforehand I need at least a basic explanation. From what I've seen until now, there is nothing depending on the language so therefore, there is no need to use lexeme as value. « Wikidata is here to provide structured content, not standard mediawiki pages. » I don't understand, to have an item entity or an lexeme entity is equality structured (if I look at a (L20817) and A (Q9659), I would even say that the second is more and better structured). Can't you answer and give some reason for using lexemes as value? Cheers, VIGNERON (talk) 15:25, 29 June 2019 (UTC)
      • I don't really understand you either. You are the one who makes such pages and than relies on the translation extension. I'm sure there is some benefit for you in this approach (it might be easier for you to make a normal wiki page, but it places a burden on everyone else.) --- Jura 15:28, 29 June 2019 (UTC)
        • @Jura1: but you don't need to understand me, you just need to give us at least one example where is a clear benefit of the lexeme entity as value over the item entity (entity in both case, no one talk about "normal wiki page" - whatever it means - except you). Cheers, VIGNERON (talk) 15:50, 29 June 2019 (UTC)
        • Apparently you add content to a normal wikipage, because you can't add it in a structured way on items. Sounds like a clear benefit for the approach. --- Jura 16:11, 29 June 2019 (UTC)
          • Wikidata:Lexicographical data/Documentation/Languages/br has nothing to do with it here. As stated in the title is just a documentation page to help people working on lexemes in this language (like the pages for others languages). It's obviously not structured data (that said, it could make use of them, I will think about it) and certainly it has nothing to do with my question here. Please answer the simple question so we can move forward. Cheers, VIGNERON (talk) 18:27, 29 June 2019 (UTC)
            • I think it illustrates that you haven't added the same in a structured way into Wikidata and could easily have avoided building the page if you had created corresponding lexemes. Now users end up trying to translate your text as you haven't entered it into Wikidata directly. Obviously, you might have "no use", but others would have been spared the burden. --- Jura 19:33, 29 June 2019 (UTC)
              • Stop making assumptions (a lot of these data are already stored in a structured way, that's why I said I will think about refactoring the page) and please answer the question. Cheers, VIGNERON (talk) 20:06, 29 June 2019 (UTC)
                • I don't see it. Maybe you can link to it. People already started translating the page into 5 languages. If you had used lexemes as others, it would already be translated. So it's not "a bit easier". Given the amount of people's time you are wasting with having to explain to you what a "normal wiki page" is and why that isn't a good idea .. --- Jura 09:40, 30 June 2019 (UTC)
                  • @Jura1: please read, understand and answer the question (if not, the alternative would be a deletion and recreation which no-one really want). The documentation page has nothing to do with it and I will completely rework it soon, so stop talking about it. My question is simple : should the datatype of this property be lexeme or item, nothing more, nothing less, stop digressing and making diversions. Cheers, VIGNERON (talk) 13:42, 30 June 2019 (UTC)
• @VIGNERON, ArthurPSmith, Jura1: Perhaps the proper datatype for this property should not be Item nor Lexeme, but instead should be Text (Monolingual Text). There are some instances where the word-initial preceded is not a single letter, but a multi-letter grapheme. An example of this is the Latin prefix in- (L35043) which takes the form ig- before words starting with "gn" (as outlined in this Wiktionary page). In this case, a letter represented by an Item or a Lexeme does not suffice. I think the current datatype of Lexeme has deterred people from using this property who would be otherwise inclined to using it. A Text datatype would make it much easier to use. Liamjamesperritt (talk) 03:04, 5 November 2019 (UTC)
  • @Liamjamesperritt: agreed. Should it be a Monolingual text or String? It's already clear that in- (L35043) concern only Latin and if we aim an easy use, then String seems to be an easier datatype. Cdlt, VIGNERON (talk) 08:57, 5 November 2019 (UTC)
  • @VIGNERON: Yes, you're right, my mistake. 'String' would be the correct Datatype. Liamjamesperritt (talk) 22:39, 5 November 2019 (UTC)

Deletion ?

@Liamjamesperritt, Hameryko, Mahir256, Luca.favorido, عُثمان, ArthurPSmith:

Is this property still useful? 4 years later, it's still widely unused.

Plus, now we have Wikifunctions that can solve the same issue (totally or maybe partially? here I'd like more info). For instance, I never used P6712 (P6712) on ar (L35217) (as it would requires to create 25 lexemes just to be used only on one item) and I created Z11700 (which is much more easy and useful).

Cheers, VIGNERON (talk) 16:24, 19 December 2023 (UTC)

Ooups, I missed that it is already propose to be deleted: Wikidata:Properties for deletion/P6712. Cheers, VIGNERON (talk) 16:27, 19 December 2023 (UTC)

I agree, I translated the label, but I never used it, and I cannot imagine a meaningful use for it... Luca.favorido (talk) 20:01, 19 December 2023 (UTC)

Ooups, wrong link, I corrected the link above to the deletion request. @Luca.favorido: according to NavelGazer you used it twice. I also see that someone removed almost half of the uses, we are down from 57 to 26 uses now. Cdlt, VIGNERON (talk) 08:17, 20 December 2023 (UTC)

In the case of Spanish language, P6712 (P6712) is being used, for example, in the prefix in- (L708363), to indicate that it has the form "im-" when the following lexeme begins with "b" or "p", or the form "i-" if it begins for "l" or "r". Could appears before phonological feature (P11950) be used in the same way? --Hameryko (talk) 09:11, 20 December 2023 (UTC)

@VIGNERON, you're right, I forgot I used that property and I even don't remember in which lexemes. :-) I probably used it to model some Italian grammatical rules, which behaves similarly to Spanish in some cases. Anyway, I don't think it is a good idea to model all the grammatical rules in Wikidata Lexemes: some of them are really difficult or with too many exceptions. Wikifunctions seems interesting for the cause. Luca.favorido (talk) 04:47, 21 December 2023 (UTC)