Wikidata:Requests for permissions/Bot/ZabesBot 3
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
- Withdrawn, lets go with T299525 instead. --Zabe (talk) 14:41, 19 January 2022 (UTC)[reply]
ZabesBot 3 edit
ZabesBot (talk • contribs • new items • new lexemes • SUL • Block log • User rights log • User rights • xtools)
Operator: Zabe (talk • contribs • logs)
Task/s: Copy 'de' label of humans to 'de-formal', 'de-at' and 'de-ch'
Code: code
Function details: It goes through the items which represent humans (P31=Q5) with an german label and without an 'de-formal' label and then copies the 'de' label to 'de-at', 'de-ch' and 'de-formal' when there is no label in those languages yet. This can be done as names are always the same in 'de', 'de-at', 'de-ch' and 'de-formal'. --Zabe (talk) 00:20, 16 January 2022 (UTC)[reply]
- I recommend to maintain a list of terms that are different from de in de-at/de-ch, in order to avoid copying labels that contain these terms in de. For de-formal, it is probably much easier since there shouldn't be much difference to de anyways. —MisterSynergy (talk) 00:37, 16 January 2022 (UTC)[reply]
- I was mainly talking about items which represent humans for now (and potentially other stuff in the future). I changed my original statement to clarify this a bit, since it was not really mentioned. For humans there should be no differences between de-at, de-ch, de-formal and de since the labels are only the names. --Zabe (talk) 00:50, 16 January 2022 (UTC)[reply]
- Alright, it's fine then :-) —MisterSynergy (talk) 00:55, 16 January 2022 (UTC)[reply]
- I was mainly talking about items which represent humans for now (and potentially other stuff in the future). I changed my original statement to clarify this a bit, since it was not really mentioned. For humans there should be no differences between de-at, de-ch, de-formal and de since the labels are only the names. --Zabe (talk) 00:50, 16 January 2022 (UTC)[reply]
- Strong oppose There is zero need to copy these labels, that’s exactly what language fallbacks are for. (For language variants, we don’t even display an indicator that a fallback took place, see T174318.) If you copy the labels, then the copied labels will just become outdated when someone changes the 'de' label without noticing that the other languages should also be updated. Lucas Werkmeister (talk) 19:24, 17 January 2022 (UTC)[reply]
- Can you please explain in which context "language fallback" works? As much as I am aware, it is a useful feature of the Wikidata web UI; in many other scenarios, users would simply experience missing labels, right? I am not aware how the "language fallback" could be used by third-party data users. —MisterSynergy (talk) 19:31, 17 January 2022 (UTC)[reply]
- I would say that multilingual interfaces to Wikidata’s data that don’t support language fallbacks are incomplete, and if they show missing labels, that’s a missing feature which should be added. The language fallback graph is available as an API; APIs like wbformatentities and wbformatvalue include language fallbacks automatically; even wbgetentities can apply language fallbacks for you. Lucas Werkmeister (talk) 21:44, 17 January 2022 (UTC)[reply]
- Mh. The query service does not have anything like this, right? —MisterSynergy (talk) 22:12, 17 January 2022 (UTC)[reply]
- Kinda same thought. When performing a query which results in Barack Obama (Q76) and then also printing the de-at label, there is no fallback happening. It simply prints 'Q76'. --Zabe (talk) 10:42, 18 January 2022 (UTC)[reply]
- I thought
SERVICE wikibase:label { bd:serviceParam wikibase:language "de-formal". }
would do that, but yeah, it doesn’t look like that’s the case. I’m not sure if users are expected to put the fallback chain into the parameters themselves (i.e.SERVICE wikibase:label { bd:serviceParam wikibase:language "de-formal,de,en". }
), or if this warrants a Phabricator task (feel free to create one). Lucas Werkmeister (talk) 20:35, 18 January 2022 (UTC)[reply]- Sure, language fallbacks would make stuff quite a lot easier. Created phab:T299525. --Zabe (talk) 14:38, 19 January 2022 (UTC)[reply]
- I thought
- I would say that multilingual interfaces to Wikidata’s data that don’t support language fallbacks are incomplete, and if they show missing labels, that’s a missing feature which should be added. The language fallback graph is available as an API; APIs like wbformatentities and wbformatvalue include language fallbacks automatically; even wbgetentities can apply language fallbacks for you. Lucas Werkmeister (talk) 21:44, 17 January 2022 (UTC)[reply]
- Can you please explain in which context "language fallback" works? As much as I am aware, it is a useful feature of the Wikidata web UI; in many other scenarios, users would simply experience missing labels, right? I am not aware how the "language fallback" could be used by third-party data users. —MisterSynergy (talk) 19:31, 17 January 2022 (UTC)[reply]
- "de-formal" shouldn't actually be used in labels AFAIK. For "de-ch" shouldn't you substitute "ss" for "ß"? --- Jura 11:47, 18 January 2022 (UTC)[reply]
- Comment I think this is similar to Wikidata:Requests for permissions/Bot/Pi bot 20. I wrote that code to be easily expandable to other language pairs - so if we want to do this, that bot script could do the job. But Lucas's points about language fallback in these cases is a very good one. Thanks. Mike Peel (talk) 21:07, 18 January 2022 (UTC)[reply]
- @Mike Peel does that mean you will discontinue your bot's task (identical, but for different language codes)? --- Jura 14:31, 19 January 2022 (UTC)[reply]
New section edit
Let's finish the discussion even if Zabesbot wont be doing it. I'd still be interested in the Pibot status for the same task. --- Jura 14:45, 19 January 2022 (UTC)[reply]
- Sorry, I didn't want to interrupt it. --Zabe (talk) 15:10, 19 January 2022 (UTC)[reply]
I'm not planning on making changes to Pi bot, it seemed to work well and is important e.g. for pt-br, I was just offering it as an alternative way to implement this. Thanks. Mike Peel (talk) 15:14, 19 January 2022 (UTC)[reply]
- @Mike Peel You seemed to agree with Lucas to use language fallbacks instead. --- Jura 08:38, 22 January 2022 (UTC)[reply]