Wikidata:Project chat/Archive/2021/01

From Wikidata
Jump to navigation Jump to search
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

Woody Allen

The place of birth (Brooklyn) is incorrect. According to the sources Bronx is correct. I can not change it. Artmax (talk) 17:35, 31 December 2020 (UTC)

For some value of 'fixed'. You have removed, rather than deprecated, the Brooklyn PoB, even though you rely on the same source for the addition of two unmarried partners. If the source is incorrect, the normal policy is to deprecate and add a Wikibase reason for deprecated rank (Q27949697) qualifier of error in referenced source or sources (Q29998666). This guards against the later readdition of the incorrect fact. --Tagishsimon (talk) 20:14, 31 December 2020 (UTC)
the source wasn't incorrect, when it was transcribed to wikidata, that is where the error was made. Funandtrvl (talk) 20:48, 31 December 2020 (UTC)
There was vandalism a while back: [1]. Oddly nothing is being done to prevent this type of edit. --- Jura 23:59, 1 January 2021 (UTC)

A wrong merge and redirect of a given name item affect many other items

Hello. Andreas (Q87263878) This is the item for the Greek name "Ανδρέας". Before some days a user decided to merge with and redirect to item Andrzej (Q14941830) [2]. Which is wrong. Some months ago I have added Andreas (Q87263878) to a lot of items of persons that are Greeks or Cypriots and have the Greek name "Ανδρέας". Now, because of the merge and the redirect, the KrBot changed in all items Andreas (Q87263878) with Andrzej (Q14941830) [3]. That affects more than 400 items! What should I do now? Find items one by one and restore them? Why a change that a user did, make such a mess? Is there an easy way to correct the items? What if someone make wrong merges to 10 or 100 items? Data Gamer play 05:00, 1 January 2021 (UTC)

Ash Crow
Dereckson
Harmonia Amanda
Hsarrazin
Jura
Чаховіч Уладзіслаў
Joxemai
Place Clichy
Branthecan
Azertus
Jon Harald Søby
PKM
Pmt
Sight Contamination
MaksOttoVonStirlitz
BeatrixBelibaste
Moebeus
Dcflyer
Looniverse
Aya Reyad
Infovarius
Tris T7
Klaas 'Z4us' van B. V
Deborahjay
Bruno Biondi
ZI Jony
Laddo
Da Dapper Don
Data Gamer
Luca favorido
The Sir of Data Analytics
Skim
E4024
JhowieNitnek
Envlh
Susanna Giaccai
Epìdosis
Aluxosm
Dnshitobu
Ruky Wunpini
Balû
★Trekker

Notified participants of WikiProject Names

Same happened with Thanos (Q97999192) from another user [4]. Data Gamer play 05:07, 1 January 2021 (UTC)

I have found the same issue to other item. Am sure if I continue checking, I will find more. It's useless Wikidata friends. Weeks of work can be destroyed only by some clicks some users can do. I am not blame the users. The problem is how easily our work can be "destroy" in Wikidata... Data Gamer play 05:28, 1 January 2021 (UTC)

Thanks. Data Gamer play 15:50, 1 January 2021 (UTC)

I will try to figure out how to change the merge tool to prevent merges going forward. --- Jura 23:57, 1 January 2021 (UTC)

Sorting by surnames on Commons (based on Wikidata)

This is a bit of a cross-project question, but I hope someone can help. I've been trying lately to create some elements for surnames, in order to enable automatic sorting of categories on Commons using Wikidata Infobox template there. But I can't get it right. Can someone explain to me how it works? For example, what's wrong with Chwalba (Q104622943)? Powerek38 (talk) 20:07, 1 January 2021 (UTC)

@Powerek38: Probably the only issue is that you’re impatient. There are some caches that take a few days to expire, maybe even more. I did a null edit on the c:Category:Andrzej Chwalba to force clearing the cache (opened the page for edit and published it without any changes), and it looks OK for me now. —Tacsipacsi (talk) 20:31, 1 January 2021 (UTC)
This is probably a cache issue as Tacsipacsi says. But for future-proofing, it's best to also add the Commons sitelink to the Wikidata item for the surname. Thanks. Mike Peel (talk) 21:08, 1 January 2021 (UTC)

Suggestion for new sparql prefix: WDO that selects only statements that are ongoing

Ongoing is defined like this:

  • has start time (P580)
  • does not have end time (P582) or end time (P582)unknown

(this was inspired by @salgo60:) WDYT?--So9q (talk) 23:17, 1 January 2021 (UTC)

doesn't unknown end time imply there is one but we just don't know it? unlike "no value" which suggests there is no end time (yet)? BrokenSegue (talk) 00:46, 2 January 2021 (UTC)

Fixing coordinates for lakes

Hello,

I've recently been comparing what's in Wikidata on lake (Q23397)s with info available from the U.S. Board on Geographic Names. For items which have a GNIS Feature ID (P590) there are ~2000 claims whose reference is imported from Wikimedia project (P143) the coordinate location (P625) in Wikidata don't match the USBGN data, at least 650 of which are more than just rounding errors. I'd like to amend this information. Since coordinate location (P625) is a single valued constraint, I'm assuming it's best to remove the old claim and replace it with a new claim. I'm also assuming claims which reference an independent source are preferable to those referenced as imported from Wikimedia project (P143). If there are no objections, I'll go ahead and make this change.

I'm also looking at coordinate location (P625) claims which reference Geographic Names Information System (Q136736) but don't match what's currently listed. Should I update these as well? Should such changes as well as those above supply a publication date (P577)?

Regards, Gettinwikiwidit (talk) 01:13, 30 December 2020 (UTC)

I just mixed up the DEC and DMS data on the GNIS website. Assuming you're working with the (correct) DMS data, it may be helpful if you could link one or two examples. It may be interesting to see if either dataset had a tendency to better approximate the centroid, for example.
For data sourced as imported from Wikimedia project (P143) I'd opt to just replace it. Where the reference is Geographic Names Information System (Q136736), it depends: if the data changed because either the maps were updated or the actual geographic feature changed, the existing data should probably remain and the current data should be added with preferred rank. publication date (P577) (on the reference) would definitely be useful, especially in that case. I'd add retrieved (P813) as well. --Matthias Winkelmann (talk) 03:14, 30 December 2020 (UTC)
@Matthias Winkelmann: I'm not sure what you're saying about DEC vs. DMS here. Entering wikibase:geoLatitude and wikibase:geoLongitude is done in DEC format though they're displayed in DMS. Here is a sample of what I propose to change. 
SELECT ?ps ?lakeLabel ?latDEC ?lonDEC ?lat ?long ?gnisURL WHERE {
  {
    SELECT ?ps ?lake ?latDEC ?lonDEC ?lat ?long ?gnisid WHERE {
      VALUES (?ps ?lake ?gnisid ?latDEC ?lonDEC) {
        (wds:q7335953-42836CD9-7979-4F71-87D7-91F1519254E6 wd:Q7335953 "1614835" "45.1918512"^^xsd:double "-84.7642201"^^xsd:double)
        (wds:Q6477828-027DF6A9-FE91-41BD-A0F3-0C5FA7373100 wd:Q6477828 "291599" "29.3079458"^^xsd:double "-82.4791358"^^xsd:double)
        (wds:q6475804-48CD1BCD-4573-4385-95CD-8C3BDB15FEA1 wd:Q6475804 "771253" "48.7499294"^^xsd:double "-113.9135721"^^xsd:double)
        (wds:q6477077-08DF184C-258E-4FD8-84BB-F811EC7427A2 wd:Q6477077 "958316" "43.3081268"^^xsd:double "-76.4369827"^^xsd:double)
        (wds:q6475994-69586DA3-56F2-4DEA-8482-EEC5E1802109 wd:Q6475994 "951103" "41.3981179"^^xsd:double "-73.6754025"^^xsd:double)
        (wds:q6377827-B18B7A53-D6AF-4DE6-B7A5-B5F3F831751D wd:Q6377827 "772940" "48.5282363"^^xsd:double "-113.4481318"^^xsd:double)
        (wds:q6475271-A124F3C5-56DE-4F4A-A8FE-56FDC80B5612 wd:Q6475271 "553954" "30.2282451"^^xsd:double "-93.2347173"^^xsd:double)
        (wds:q6477999-0D69C812-5795-4976-B1AB-5CA0CBBBBE60 wd:Q6477999 "967114" "44.1070059"^^xsd:double "-73.9356783"^^xsd:double)
        (wds:q6477136-35814F23-B203-439A-A0B7-6F38CD948C79 wd:Q6477136 "1570331" "45.2566848"^^xsd:double "-87.9080654"^^xsd:double)
        (wds:q6476368-2F724E96-DABC-4346-9857-14260CA74979 wd:Q6476368 "772914" "48.7810128"^^xsd:double "-113.6756959"^^xsd:double)
        (wds:q7235286-832B6767-EEE5-47EA-8109-4F609CF362E6 wd:Q7235286 "1217748" "41.3873242"^^xsd:double "-71.5311693"^^xsd:double)
        (wds:q6477143-763568E7-6158-4A91-BCB3-A7FAB06F660E wd:Q6477143 "306117" "28.939347"^^xsd:double "-81.5410478"^^xsd:double)
        (wds:q6938987-2454D549-D883-4EC7-81C9-3F678F033AB5 wd:Q6938987 "1570045" "45.6965111"^^xsd:double "-91.2081829"^^xsd:double)
        (wds:Q6476282-B81F7238-E49F-46DA-8EED-2375226D6DCB wd:Q6476282 "477318" "38.2472036"^^xsd:double "-97.7191383"^^xsd:double)
        (wds:q6693772-8075AC7F-DA34-4570-8D35-634FE0D19BFA wd:Q6693772 "956107" "44.3028384"^^xsd:double "-74.1931089"^^xsd:double)
        (wds:q6476251-89FAD984-5259-452C-A87A-AEE4B803AE36 wd:Q6476251 "284452" "30.6312498"^^xsd:double "-84.2390238"^^xsd:double)
        (wds:q5665721-81992C6B-A823-430F-9504-091EBFE78FFF wd:Q5665721 "784584" "48.5167701"^^xsd:double "-113.7705404"^^xsd:double)
        (wds:q6477828-6BDDFE36-5F26-4643-A942-E600D1088C8C wd:Q6477828 "291599" "29.3079458"^^xsd:double "-82.4791358"^^xsd:double)
        (wds:q6477231-85C105DC-33C1-4645-828D-26F62B11BE69 wd:Q6477231 "288259" "26.5970392"^^xsd:double "-80.0780448"^^xsd:double)
      }
      ?ps psv:P625 ?node.
      ?node wikibase:geoLatitude ?lat;
        wikibase:geoLongitude ?long.
    }
  }
  BIND(URI(CONCAT("https://geonames.usgs.gov/apex/f?p=GNISPQ:3:::NO::P3_FID:", ?gnisid)) AS ?gnisURL)
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!
I definitely have a preference to use the info as referenced by a third party source rather than try to calculate a centroid on my own. My reasoning is that makes it easier to run reconciliation checks against the source. The sample above are all referenced with imported from Wikimedia project (P143). I'll post back when I move on to the next stage. For clarity, I understand you as not having a problem making this change. Please correct me if I've misunderstood. Regards, Gettinwikiwidit (talk) 04:56, 30 December 2020 (UTC)
Perhaps I should say that geodata *may* be entered in DEC. I'm not exactly sure what you tried. Gettinwikiwidit (talk) 04:45, 30 December 2020 (UTC)
It might also be worth noting that I fixed some claims to satisfy the single value constraint on GNIS Feature ID (P590) by supplying the alternate name with an alternative name (P4970) qualifier. See my recent edits if you have any questions or concerns. Gettinwikiwidit (talk) 04:45, 30 December 2020 (UTC)
I object to your plan to blindly replace one value with another. P625 is, like all properties, capable of having multiple statements on an item, ideally one of these having best rank. There is nothing in principle wrong with coordinate values imported from a language wikipedia, and no guarantee that GNIS coordinates are an improvement on language wikipedia sourced data. Your definate preference for info as referenced by a third party source is noted, but you should appreciate that we hold a wealth of info as referenced by a third party source which on examination turns out to be complete horseshit. By all means add new data, but do not presume to remove data without proper examination and reason, and do not presume to make unexamined and uncompared data best rank. --Tagishsimon (talk) 04:57, 30 December 2020 (UTC)
@Tagishsimon: Please try to tone down the rhetoric. First of all as mentioned there are property constraints on coordinate location (P625). Second, I've referenced my sources whereas data imported from Wikimedia project (P143) comes without any references that can be effectively cross checked. If you have reason to believe this data to be better than the referenced data or that the source I've referenced is unreliable by all means bring that to the discussion. But if this is just a personal attack on my approach I'll move on without regard to it. Regards, Gettinwikiwidit (talk) 05:12, 30 December 2020 (UTC)
Please try to understand what the property constraint is, before going off half-cocked. The single value constraint calls for a single best rank value, not a single value. There are any number of articles with a plurality of coordinates. Problems only arise where there are two or more best ranked values. An example, for your inspection is, Q17843364#P625. If you wish to add new coordinates backed by a source, knock yourself out. But you should not, as I noted above, blindly remove coordinates nor blindly set your latest-imported-coordinate to best rank. From my perspective, this is not a personal attack. If you see being appraised of your lack of understanding of the constraint, rank, and custom & practice on wikipedia as an attack, that is for you to deal with. --Tagishsimon (talk) 09:39, 30 December 2020 (UTC)
@Tagishsimon: Please stop using accusatory language. It's truly inappropriate. You have no actual questions about the work that went into this and you can't even acknowledge the fact that bringing this to this forum is not running off half cocked. So if you can adjust your approach we can have a productive conversation here. But I will not be thrown on the defensive by someone who has offered no evidence that they've put any energy whatsoever into their comments. Regards, Gettinwikiwidit (talk) 10:29, 30 December 2020 (UTC)
@Gettinwikiwidit: Tagishsimon is making a valid point; please do not ignore their comments just because you unhappy about the wording used. Simon Cobb (User:Sic19 ; talk page) 13:15, 30 December 2020 (UTC)
@Sic19: Maybe you could reword it so that we can discuss this properly. What would you say his core point is? Gettinwikiwidit (talk) 14:53, 30 December 2020 (UTC)
@Gettinwikiwidit: I'll try. The issue is the removal or deprecation of statements that are not necessarily incorrect, without a process to manually review or otherwise analyse the existing coordinate location (P625) value(s) beforehand. Just because you have coordinate data from an authoritative source does not make the existing data wrong. To put it another way, the P625 coordinates claim, i.e. a point, for a lake is always somewhat arbitrary because, I hope we can agree, as long as the point is within the area of the lake it is not incorrect. Unless the existing coordinates completely miss the lake, it is questionable whether any other coordinates are an improvement. The addition of coordinates for coordinates of northernmost point (P1332), coordinates of southernmost point (P1333) etc., where these claims do not exist, would, on the other hand, be a significant impovement to our existing lake data, as would geoshape (P3896) (polygon) data.
I should say, this is not a subject I feel strongly about nor do I have any desire to speak on behalf of @Tagishsimon: but I do hope you can now have a constructive discussion about what appear to be very real concerns. Best, Simon Cobb (User:Sic19 ; talk page) 16:16, 30 December 2020 (UTC)
@Sic19: I'm happy to have that discussion but without anyone offering a strong view on how this should be done, there's not much of a discussion to be had. In that environment, I'm inclined to go with my instinct. My argument that a new point is an improvement is that it's clearly sourced and thus more effectively serves as a reference that can be cross-checked. Your point about it being arbitrary can be flipped to ask why preserving old values serves any function here? @Matthias Winkelmann: above makes an interesting point about geological structures perhaps changing over time so that old values may be information worth preserving. But again with data which has no references it's hard to follow up. Regards, Gettinwikiwidit (talk) 22:13, 30 December 2020 (UTC)
@Sic19: For clarity, my problem with the initial wording is that it seems more clearly focused on characterizing the effort I put into this work than actually offering a viewpoint on the work. I dare say your first reply can read like you took that characterization at face value. I think personal attacks not only have no place on this forum but discourage the very volunteer activity that maintains it. Such attacks should be universally pushed back on. Gettinwikiwidit (talk) 22:18, 30 December 2020 (UTC)
Although no one is offering a strong view about how this task should be undertaken, there is an objection to your proposed approach which should be discussed before you do any furthur work on this task.
Please can you clarify what you mean when you say "I dare say your first reply can read like you took that characterization at face value"? What are you implying? I have merely interjected to help two fellow editors find an amicable solution in a situation I perceive as unnecessarily hostile. Simon Cobb (User:Sic19 ; talk page) 23:26, 30 December 2020 (UTC)
@Sic19: I agree it was entirely unnecessarily hostile. And it should be clear what the source of that was. You said that I was dismissing claims because I didn't like the wording, but I was simply rejecting the hostility. The objection is not on any clear grounds. Regards, Gettinwikiwidit (talk) 00:04, 31 December 2020 (UTC)
  • Just saying : for lakes that do not clearly look like a circle, how does someone calculate the centroïd of the lake? Sources are here to help give that centroïd. Adding P625 is a nice help, but do not remove other P625, just have them ranked down if these P625 are clearly pratting around. (which is a human task, not quickstatementable). --Bouzinac 💬●✒️●💛 23:47, 30 December 2020 (UTC)
@Bouzinac: I'm not sure what you mean by pratting around. Is the argument that claims should never be removed? If not, then what is the reasoning behind keeping around claims which have no clear reference to a source especially claims that @Sic19: above points out are necessarily somewhat arbitrary? I'm not sure that the USGS uses the centroid or that the centroid is even necessary for a coordinate location (P625) claim. I am suggesting that the USGS is a recognized authority. And that using such authorities makes it easier to cross-check and cross-reference against. Regards, Gettinwikiwidit (talk) 00:04, 31 December 2020 (UTC)

Being a great user of GNIS Feature ID (P590), actually the third according to Navel Gazer, I can assert that their coordinates are not more accurate than anything else. They are sometimes way off and I thus do not support any form of methodical replacement. Time-consuming manual work has been invested into bettering the coordinates and it would be a great loss should a single move suddenly erase days, hours of careful checks. Thierry Caro (talk) 00:29, 31 December 2020 (UTC)

@Thierry Caro: This is interesting. Way off as in they don't actually point to the item indicated? Can you give an example? Can you point to an example that has been manually altered to be more accurate? Would such entries be labelled as imported from Wikimedia project (P143). My main concern is that manually adjusted coordinates make it more difficult to reconcile information. That is to say, when adding new items it's much more difficult to be assured that two items are not the same item when you can't check against a consistent reference. I would imagine many hours went into the USGS work as well. Regards, Gettinwikiwidit (talk) 00:37, 31 December 2020 (UTC)
FWIW, each of the claims in the sample above (by no means handpicked) were imported by a bot. Or is your point that the Wikipedia data was manually edited? I'd still be concerned that without a clear reference it's not clear what standards applied to each of these entries. But perhaps this is the reasoning behind keeping around old data?? I.e. to have something to fall back on should the primary source be inaccurate. Gettinwikiwidit (talk) 00:48, 31 December 2020 (UTC)
GNIS is notoriously error-prone and should not be given priority over any other data source. Abductive (talk) 02:14, 31 December 2020 (UTC)
@Abductive: Can you point me at any examples? Regards, Gettinwikiwidit (talk) 02:28, 31 December 2020 (UTC)
A non-lake one I remember is Drawbridge, CA which missed the island the ghost town was on, and certainly didn't point to the remaining road and buildings. Many features that are small are missed by over 1°. How about Brentwood Lakes? It took me about 20 looks at random lakes to find this one. The coordinates are on land, but also don't point in between the two lakes. Anyway, GNIS gives six or seven decimal points of precision (Q962365), which is bad because it looks authoritative but is not finding centroids or anything, and is inaccurate often enough to not be trustworthy. Abductive (talk) 03:08, 31 December 2020 (UTC)
  • If there was a property "GNIS coordinates", that would be fine. But at present if the are two sets of coordinates in an item, the little lightning arrow icon appears saying that there shouldn't be two sets of coordinates. Abductive (talk) 03:11, 31 December 2020 (UTC)
@Abductive: Sorry. I was asking if you had any examples of GNIS data being wrong. Do you have any? Gettinwikiwidit (talk) 03:46, 31 December 2020 (UTC)
Just above. See? Do you need more? Abductive (talk) 03:54, 31 December 2020 (UTC)
@Abductive: Thanks. Obviously a handful of examples among the millions of entries isn't enough to establish GNIS as an unreliable source. Jc3s5h's experience below speaks to their responsiveness when errors are reported to them. The examples do show they're not infallible though. Gettinwikiwidit (talk) 05:12, 31 December 2020 (UTC)
I showed you 5% error rate on lakes, and it is worse on smaller objects. Abductive (talk) 05:16, 31 December 2020 (UTC)
I don't think that checking 20 cases and finding one counts as statistically significant does it? Also you should report those cases to GNIS. It sounds like they're responsive to feedback. Gettinwikiwidit (talk) 06:31, 31 December 2020 (UTC)
Also checking those coordinates for Brentwood Lakes with Google Maps, I wouldn't call that wildly off, nor would I consider finding the precise centroid critical. Moreover, I didn't find a reference to this lake in Wikidata. Gettinwikiwidit (talk) 06:57, 31 December 2020 (UTC)
@Abductive: Hmmm.. Also you pointed to "Brentwood Lakes", plural. There is a separate entry specifically for Brentwood Lake, which does seem to be located within the lake itself. Gettinwikiwidit (talk) 07:18, 31 December 2020 (UTC)
I have nearly 90,000 edits on en.wiki, and quite a lot of them involve correcting GNIS coordinates. Remember kids, GNIS stands for Garbage Notoriously Inaccurate Shit. Abductive (talk) 07:24, 31 December 2020 (UTC)
I'm not generally persuaded by anecdotal evidence. Also, it would seem like you would have easy access to plenty of examples. Gettinwikiwidit (talk) 07:33, 31 December 2020 (UTC)
I'll add another error: Grandpa Knob (Q49032661). I noticed an error two years ago, and deprecated the rank of the value supported by GNIS. I wrote to GNIS; they checked their old maps, and decided a map from 1944 showed the feature more clearly than more recent maps, so I was correct. They fixed it. Jc3s5h (talk) 04:08, 31 December 2020 (UTC)
@Jc3s5h: Thanks. And thanks for doing the right thing and having GNIS fix their entry. Gettinwikiwidit (talk) 05:12, 31 December 2020 (UTC)
Apart from that Wikidata:Verifiability is not Wikidata policy and thus doesn't dictate a mission. ChristianKl ❪✉❫ 11:43, 31 December 2020 (UTC)
@ChristianKl: I personally think not holding to a verifiability policy is a big mistake and plays into the common conception of how much Wikimedia data can be trusted. Also, there was some indication that GNIS data should *not* be the preferred rank, but a preferred rank is required by the property constraint for coordinate location (P625). There was never any indication that the statements added would not have a clear reference and the main intent of the effort was to ensure that there were clear references for the data provided. In any event I'm put off enough by the general tone that I've lost interest. Gettinwikiwidit (talk) 12:09, 31 December 2020 (UTC)
@Sic19: I don't believe that to be true. I believe that is Dead Lake. Is this something you're working on? Gettinwikiwidit (talk) 15:56, 31 December 2020 (UTC)
Or rather the value that I changed was pointing at Dead Lake. It's not clear to me what your claim is at this point. Gettinwikiwidit (talk) 16:10, 31 December 2020 (UTC)
It doesn't really matter what you believe; it is correct per Google Maps and various other mapping services. Coordinates are included on the list of items not requiring a source for a simple reason - we can verify them using a map. Please cease your disruptive editing.
@ChristianKl: as an administrator, could I ask you to intervene as appropriate. Unfortunately, Gettinwikiwidit does not appear to be listening to the concerns raised. Simon Cobb (User:Sic19 ; talk page) 16:13, 31 December 2020 (UTC)
@Sic19: Please explain how this latest concern is related to the discussion above. FWIW, I have changed it now and the only difference is that I hadn't changed the precision from the previous entry. The entry was accurate. If you're concern now is that the precision of any changes must be updated to the highest precision, then feel free to make that argument. Regards, Gettinwikiwidit (talk) 16:18, 31 December 2020 (UTC)
You are removing data when you have been asked not to and replacing it with inaccurate data per the example I provided above. Thank you for correcting the precision; that does not, however, change the fact that the coordinates you added were not accurate. How can we be sure that the other coordinates you are changing do not have similar problems?
Further, the reference you have added is insufficient because it does not provide a link which can be used to verify the data. Simon Cobb (User:Sic19 ; talk page) 16:33, 31 December 2020 (UTC)

I agree that coordinates imported from Wikipedias can't be blindly replaced with data imported from an official dataset. I don't know about GNIS but a lot of official sources have very often inaccurate coordinates. If coordinates are going to be replaced, first they need to be checked one by one with a map to check that they are better than the existing ones - and there are tools to help in that task.--Pere prlpz (talk) 16:55, 31 December 2020 (UTC)

Swedish lakes all have id-numbers. Those id-numbers are always the coordinates (in some system) of the drain of the lake. I doubt they are the best choice of main coordinate for the lake, but it is a perfect source for locating the drain. 62 etc (talk) 18:12, 31 December 2020 (UTC)
As far as manually adding references that don't include a direct link, I don't think we have rules against that, however if a batch is added via QuickStatements it would be good if it uses the GNIS Feature ID (P590) in it's reference. ChristianKl ❪✉❫ 18:20, 31 December 2020 (UTC)
@ChristianKl: thanks. Yes, I agree there are not rules against adding references without a link. However, in this instance, without a link we can only assume the source data is available via GNIS Feature ID (P590) → 1433778. If that is the source of the coordinates for Uintah Lake (Q16256196), the reference is incorrect because the values do not exactly match, which seems to go against the stated aim of importing referenced data from GNIS to make it verifiable.
@Gettinwikiwidit: apologies if you have found my comments in this thread unhelpful; that is the complete opposite of my intention and I am not going to be involved in any further discussion on this topic as it clear will not be constructive. Simon Cobb (User:Sic19 ; talk page) 18:48, 31 December 2020 (UTC)
@Sic19: FWIW, I chose not to put the elink in the qualifiers as it seemed redundant because it existed in a separate claim for the entity. Gettinwikiwidit (talk) 21:57, 31 December 2020 (UTC)
Also, no worries. I opened this thread because I was looking for guidance. I did feel I was treated harshly in spots, but I pushed back mainly because I'd like this to be a kinder space where people can work together not because I was looking for a quarrel. Regards, Gettinwikiwidit (talk) 22:09, 31 December 2020 (UTC)
  • FWIW, I've eyeballed quite a few entries at this point and the GNIS data seems to overall be very good. I don't detect any interest in finding the centroid which many times will not be a point in the body of water. I still don't see the point in keeping the unreferenced data when the referenced data is just as good, but I've let that stand. FWIW, here are a few entries which seem to be misentered in GNIS (or at least oddly chosen reference points): 
    #defaultView:Map
SELECT DISTINCT ?place ?placeLabel ?location ?layer ?distance WITH {
  SELECT ?place ?gnisPt ?distance WHERE {
    VALUES (?ps ?place ?gnisPt ?gnisid ?distance) {
(wds:Q34920071-BF7F5206-0A46-4D09-BFCB-CE19ABB75EC3 wd:Q34920071 "Point(-89.8838357,32.5608125)"^^geo:wktLiteral "676892" "9.635"^^xsd:double )
(wds:Q34920071-BF7F5206-0A46-4D09-BFCB-CE19ABB75EC3 wd:Q34920071 "Point(-89.8838357,32.5608125)"^^geo:wktLiteral "676892" "9.635"^^xsd:double )
(wds:Q34920071-BF7F5206-0A46-4D09-BFCB-CE19ABB75EC3 wd:Q34920071 "Point(-89.8838357,32.5608125)"^^geo:wktLiteral "676892" "9.635"^^xsd:double  )
(wds:q4491409-D1B5021F-C153-418C-BC64-09CCA34A1DCB wd:Q4491409 "Point(-78.3957169,35.9040142)"^^geo:wktLiteral "999761" "10.139"^^xsd:double )
(wds:q6476491-BED67AE3-7F33-4E16-9D24-09ADD5392B9C wd:Q6476491 "Point(-115.2357799,48.5436406)"^^geo:wktLiteral "806845" "11.819"^^xsd:double )
(wds:q4311340-A539AE60-F711-4273-AA21-6BB82E7615A3 wd:Q4311340 "Point(-107.051753,36.6717657)"^^geo:wktLiteral "892389" "12.911"^^xsd:double )
(wds:q4311340-A539AE60-F711-4273-AA21-6BB82E7615A3 wd:Q4311340 "Point(-107.051753,36.6717657)"^^geo:wktLiteral "892389" "12.911"^^xsd:double )
(wds:q7331396-87B5F12B-6C39-436B-BFEB-3881CE34AECD wd:Q7331396 "Point(-96.9508621,35.2056252)"^^geo:wktLiteral "1946039" "18.016"^^xsd:double )
(wds:Q49701832-B00165FB-F5DA-45DE-9A41-B9526E69517C wd:Q49701832 "Point(-79.4523432,40.2355775)"^^geo:wktLiteral "1180051" "18.736"^^xsd:double )
(wds:Q49701832-B00165FB-F5DA-45DE-9A41-B9526E69517C wd:Q49701832 "Point(-79.4523432,40.2355775)"^^geo:wktLiteral "1180051" "18.736"^^xsd:double )
(wds:Q49701832-B00165FB-F5DA-45DE-9A41-B9526E69517C wd:Q49701832 "Point(-79.4523432,40.2355775)"^^geo:wktLiteral "1180051" "18.736"^^xsd:double )
(wds:Q1616130-ef10340f-4255-557b-8b04-e059d77b0128 wd:Q1616130 "Point(-121.3699417,37.6260417)"^^geo:wktLiteral "243393" "28.251"^^xsd:double )
    }
    FILTER EXISTS { ?place wdt:P590 [] }
  }
} AS %vals
WHERE {
  {
    INCLUDE %vals
    ?place wdt:P625 ?location.
    BIND( 'wiki' as ?layer )
  } UNION {
    INCLUDE %vals
    BIND( ?gnisPt AS ?location )
    BIND( 'gnis' as ?layer )
  }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
} ORDER BY ?label
    Try it!
    Regards, Gettinwikiwidit (talk) 09:08, 2 January 2021 (UTC)

How to create stable data tempate for wiktionary?

How to do that? My language has only auto generated template. -- 12:44, 31 December 2020 Bicolino34

Item:Q304675 - MF Doom

What should we do in cases like this where multiple databases (VIAF, ISNI, etc) assign multiple identifiers to the same unique individual, and aren't consistent about what natural language labels they give the individual? This person – MF Doom – was known by numerous pseudonyms and aliases, and the external databases have taken that up. Consequently, this item shows multiple errors. I created Item:Q104604985 to consolidate identifiers associated only with one of those pseudonyms – Viktor Vaughn – but I wasn't sure if that was correct. AleatoryPonderings (talk) 22:51, 1 January 2021 (UTC)

I don't think that creating items for pseudonms is a good idea. You can add the multiple entries on a single item, and use a subject named as (P1810) qualifier. Ghouston (talk) 23:43, 1 January 2021 (UTC)
  • The fact that the item shows an error is good. It suggests to VIAF that they have a problem of two IDs pointing to the same person. On the Wikidata site we can give one of the VIAF IDs a best rank to handle the issue on our side. ChristianKl ❪✉❫ 01:09, 2 January 2021 (UTC)
  • At one time we had a representative from VIAF that responded to emails and was performing merges and detangling conflations, but they stopped responding over a year ago. Does anyone have a new contact at VIAF? We should arrange to have a contact for each of our Identifiers. --RAN (talk) 03:03, 2 January 2021 (UTC)

Wikidata:Database reports/Constraint violations/P2671

The bot cannot save the constraint violations list at Wikidata:Database reports/Constraint violations/P2671 because g.co seems to be in a spam blacklist (although it's the main url formatter of Google Knowledge Graph ID (P2671)). This is a weird situation. How can this be resolved? Steak (talk) 12:34, 2 January 2021 (UTC)

botany (Q441) Arabic aliases are disruptive

A bot User:ASammourBot added the majority of Arabic aliases for this entity. These include hundreds of specific plant names, including very common plants with their own entities, such as "apple" and "lemon". I would fix it myself, but the entity is locked.  – The preceding unsigned comment was added by High surv (talk • contribs) at December 14, 2020 (UTC).

Merge Silver lake entities?

I believe that Silver Lake (Q65118798) and Silver Lake (Q49702647) are the same entity. Is this the place to make a merge request? Thanks, Gettinwikiwidit (talk) 06:37, 2 January 2021 (UTC)

Probably easier if you enable the preferences/gadgets/merge tool and do these merges yourself. There are, in particular, lots of duplicate articles arising from ceb wikipedia. Where they are the same thing they can be quietly merged. --Tagishsimon (talk) 06:42, 2 January 2021 (UTC)
✓ Merged. Tagishsimon: As said above, you can enable the merge tool in your preferences. See Help:Merge. Esteban16 (talk) 17:04, 2 January 2021 (UTC)
  • Ugh! Ceb Wikipedia strike again. I have merged hundreds, and there are still thousands that are poorly defined geonames entries, some synonyms, some direct duplicates, some so poorly defined, they will never be properly disambiguated. --RAN (talk) 21:39, 2 January 2021 (UTC)

QueryGraph : Tool to create a query from a graph

Hello,

I come to present QueryGraph which is a web tool for creating queries in SPARQL for Wikidata, from graphs.

The goal is to make query creation easier.

The concept is to draw a graph composed of nodes representing a data and edges representing the links between these data, from this the software generates a SPARQL query.

Link to test the tool.

neat. would be cool if I could see the generated SPARQL code from the graph. is this open source? BrokenSegue (talk) 14:33, 2 January 2021 (UTC)
@BrokenSegue Yes it's Open-Source, The code is available on github https://github.com/shevekk/QueryGraph. In the next version the SPARQL generated will be visible in the software. --Shevekk (talk) 20:06, 2 January 2021 (UTC)
A detail. Mountains situated in the Alps would be stored through mountain range (P4552), not part of (P361). Thierry Caro (talk) 17:00, 2 January 2021 (UTC)
@Thierry Caro Thank you for your feedback. I will correct it in the next few days. --Shevekk (talk) 20:06, 2 January 2021 (UTC)

Joseph Nash the younger

Joseph Nash the younger (Q104631253) and Joseph Nash (Q19842174) are the same person. (Q104631253) requires deleting. The creator needs fixed too, and yes. I did take the trouble to look for the original entry first before making the later. Thanks. Broichmore (talk) 16:47, 2 January 2021 (UTC)

@Broichmore: Done. Remaining item is Joseph Nash (Q19842174). The merge gadget is your friend for cases like this -- go to merge under Wikidata:Tools/Edit items and hit activate to add it to your user interface; or go to Preferences from the links across the top, then the Gadgets tab, and select the 'merge' check-button. Jheald (talk) 17:30, 2 January 2021 (UTC)
Oh, and your edit summary is entirely true. The search function is rubbish. I find that Reasonator (activate from Preferences -> Gadgets -> Reasonator) gives slightly more useful search results, as it returns a summary based on the statements in the item, rather than just what is in the 'description' box. But neither search works well with slight variances in spelling. Jheald (talk) 17:36, 2 January 2021 (UTC)

Compositions versus recordings

Newbie question here: do we typically segregate recordings of a work into a distinct item from the composition? Asking because Music for 18 Musicians (Q2714697) seems to mix identifiers referring to specific recordings with identifiers referring to the composition as an abstract entity distinct from its various recordings. Same question would also apply to books (say, Emma (Q223880)) versus specific editions (say, this one). AleatoryPonderings (talk) 20:31, 2 January 2021 (UTC)

  • Things usually start of with a single generic entry, and if someone wants to do the work, they can break it apart into the various recordings by other artists, and re-releases. When a book is a particular edition, we use "instance_of=version, edition, or translation". If it is an entry for the generic version of the book we use "instance_of=written_work". I generally use "instance_of=version" if I am adding in a "Google Book ID" since that is tied to a specific version, edition, or translation. I assume a recording should be broken down the same way, if you want to do the work. --RAN (talk) 21:34, 2 January 2021 (UTC)

Problem in descriptions of Korean given name objects

Hello. I am currently frequently editing Korean given name objects. Most Korean given names are gender neutral. However, in wikidata, also a lot of those gender neutral names were declared as "female given name" or "male given name" by the Propery instance of (P31). Then, a bot comes and writes translations of that in the descriptions for most languages. Now, I changed it to "gender neutral name". Is there a way to trigger the bot to change all the descriptions again? Recent examples are Eun-seon (Q69507702), Nae-yeong (Q69512018) and Jeong-seong (Q69511040). --Christian140 (talk) 10:32, 1 January 2021 (UTC)

Bots tend not to change existing descriptions. What I'd do is blank the existing descriptions using MediaWiki:Gadget-dataDrainer.js, and hopefully it will get repopulated some time. Ghouston (talk) 21:34, 1 January 2021 (UTC)
To use that script, you add it to a personal common.js, like in User:Ghouston/common.js. Ghouston (talk) 21:35, 1 January 2021 (UTC)
It might be easier to determine a suitable format, set it on one prototype item, and then use wikibase-cli (Q87194660) to apply to all similar items. --- Jura 23:55, 1 January 2021 (UTC)
If you are interested, I can set up a page for Korean names like the one discussed at Wikidata_talk:WikiProject_Names#Reports_for_Italian_given_names_(or_another_language). --- Jura 10:19, 2 January 2021 (UTC)
Thanks. I will look into it. --Christian140 (talk) 10:27, 3 January 2021 (UTC)

Brexit in Wikidata

What is the correct modelling of United Kingdom (Q145) and European Union (Q458) see WD Q145#P463 - Salgo60 (talk) 13:23, 1 January 2021 (UTC)

That looks fine. Start date 1 January 1973, end date 31 January 2020. (Yesterday's change wasn't withdrawal from the EU - it was the end of the post-withdrawal transition period.) The only quibble I would have is whether the start date is correct - the UK joined the EEC, there was no EU in 1973, so you might want to model that as two separate claims. However, I guess that's something you'd want to check is handled consistently across all member state items rather than just changing it for the UK. Andrew Gray (talk) 14:51, 1 January 2021 (UTC)
It looks correct to me too - the United Kingdom (Q145) stopped being a member of the European Union (Q458) on 31 January 2020 because of Brexit (Q7888194). Simon Cobb (User:Sic19 ; talk page) 14:57, 1 January 2021 (UTC)
The currently favoured inception date on European Union (Q458) is 7 february 1992, presumably memberships should match that. Ghouston (talk) 21:30, 1 January 2021 (UTC)
This should be 1 November, 1993, the day that Maastricht Treaty (Q11146) took effect. Ghouston (talk) 21:38, 3 January 2021 (UTC)

Thanks... we dont lower rank? --> I need to check for end time (P582)... my test query that I need to change 

  BIND (exists{?cid wdt:P463 wd:Q458} AS ?memberEU)

- Salgo60 (talk) 16:00, 1 January 2021 (UTC)

@Salgo60: normal rank is appropriate since the information in the statement is correct. The following change to your query should work:
BIND (exists{?cid wdt:P463 wd:Q458} && not exists{?cid p:P463 [ ps:P463 wd:Q458 ; pq:P582 [] ]} AS ?memberEU)
Simon Cobb (User:Sic19 ; talk page) 18:17, 1 January 2021 (UTC)
@Salgo60: Yes, agree with Simon - the original query will get "everyone who's ever been a member of the EU". Similarly, it will get "anyone who's ever been a member of the IMF", but there are a couple of former countries - Cuba and Taiwan both left in the past, though their items on Wikidata don't seem to reflect that yet. (In fact, only eight countries currently have "member of IMF" set!) So you might need to apply the end-date filter to a couple of the other elements as well.
The wdt: format will also break down if it's not the "best" statement available. So for example with the USA, while it is assigned as a member of the IMF, for some reason a handful of P463 statements are "preferred" and the rest "normal", so the "normal" ones aren't found by the wdt: query. Not quite sure what's going on there but a good reason to prefer the p/ps format. Andrew Gray (talk) 22:11, 1 January 2021 (UTC)
Many thanks .... my reflexion is that how we now model things makes the SPARQL very complex --> less people will be able to use the data in Wikidata.... maybe we need a new prefix that fetch relations with no end time (P582) or end time (P582) is a future date
OT: I have used this query to do some statistics on the progress of the Corona Vaccination GITHUB / Jupyter Notebook - Salgo60 (talk) 22:49, 1 January 2021 (UTC)
Q155#P463 for Brazil makes a different use of the ranks on these. With this, wdt:P463 should only get you current members. --- Jura 23:42, 1 January 2021 (UTC)

Mayflower passengers

Found that page: Google Maps with 86 of the 103 passengers of the Mayflower. All of them are already in wikidata but for many birth and death dates/places are missing. So i crawled that into a csv though the googlemaps-API.

Now I have 2 Questions: Is it allowed to copy such "bigger" amounts of data to wikidata and how to cite it. The googlemaps-API needs API tokens, which are short lived... But everyone can verify the data though the website. – At least for the moment. I assume such an interactive website won't be archived through archive.org. I tried to make a source-object: Q104588852 But it didn't go very well. Can someone more experienced help? --Fabiwanne (talk) 12:06, 3 January 2021 (UTC)

@Fabiwanne: it's fine (but not optimal) to have references that aren't trivially verified. Better than no reference. You could upload your scraped data to archive.org and link to it if you want. Also this isn't really a "large" amount of data to import so I wouldn't worry. BrokenSegue (talk) 15:27, 3 January 2021 (UTC)

SMB-digital ID

as long as this statement is not shown in tl artwork of the file one should not delete statements in P973--Oursana (talk) 13:20, 3 January 2021 (UTC) @Thierry Caro:--Oursana (talk) 23:25, 3 January 2021 (UTC)

Write into wikidata, multiple (very complete) family trees I own

Hello! Is it allowed to import (hand written) family tree to wikidata? Wikidata could be the place to store and traverse the largest unified family tree of the world!  – The preceding unsigned comment was added by 2A01:CB05:8AE3:E500:E8C5:7E39:8E76:CCAD (talk • contribs).

all items here must meet our notability standards BrokenSegue (talk) 19:36, 3 January 2021 (UTC)
Wikidata has a completely different goal, you may want to visit geni.com (Q2621214). 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 12:57, 4 January 2021 (UTC)
@1234qwer1234qwer4: Wikidata has many different goals and genealogy is one of them. However we care more about sources. In this case uploading the core family tree documents to WikiSource is a way to have sources for the Wikidata items. ChristianKl ❪✉❫ 14:56, 4 January 2021 (UTC)
@ChristianKl: I was talking specifically about "Wikidata could be the place to store and traverse the largest unified family tree of the world!" 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 14:57, 4 January 2021 (UTC)
Wikidata's sourcing standards are a bit higher then those of some other websites and thus there are reasons for Wikidata's tree to be smaller then that of other websites but storing high quality geneology data is within Wikidata's scope. ChristianKl ❪✉❫ 15:10, 4 January 2021 (UTC)

Wikidata weekly summary #449

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Other Noteworthy Stuff
  • Development
    • Due to the winter holidays, no deployment has happened for Wikidata in the last two weeks.

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

Too old

I look for people over 110 years old since they are most likely errors, unless marked as a supercentenarian. Most result from typos, some from vandalism. We have to figure out how to mark people that are exaggerations, and not true supercentenarians. Please help at Wikidata:Database_reports/unmarked_supercentenarians --RAN (talk) 02:29, 3 January 2021 (UTC)

@Richard Arthur Norton (1958- ):some of them are actually alleged to have lived that long but just probably didn't. for example Sreeman Tapaswiji Maharaj (Q94506405) could be sourced to [6] but it's probably nonsense. Not clear what to do here. Maybe add a "dubious" qualifier or something. BrokenSegue (talk) 02:55, 3 January 2021 (UTC)
Yet we still need a way to get him and the hundred others off the list, so we can find and fix errors, and figure out how to distinguish real from exaggerated people. That list grows by a dozen people or more each month, most new entries are typos. Ultimately we want a list of true certified supercentenarians. Anyone can make a claim, but we will need a way to distinguish true certified supercentenarians, so we can auto-generate a list just from our data. We still do not have that yet, but many suggestions. We wouldn't want a person who claimed to be President of the United States or claimed to be The Pope in an interview in a reliable source, to show up in our master list of those positions. --RAN (talk) 02:03, 5 January 2021 (UTC)

Coordinates now in line, but not matching the label

Hello,

The coordinate location (P625) and GeoNames ID (P1566) for Ca Ira (Q48982124) had lined up but mismatched the GNIS Feature ID (P590). I've changed this to match the other coordinates, but the Wikipedia entry points to the dam, which is not the reservoir and has a different GNIS Feature ID (P590) (the previous one). There is a different from (P1889) to a somewhat cryptic disambiguation page. I'm somewhat confused about how this should be cleaned up. Should there be a separate entry for the reservoir? Should we repurpose this for the reservoir? The entry is only on ceb-wiki and not I don't speak Cebuano, so I'm not sure how best to follow up. Any advice?

Regards, Gettinwikiwidit (talk) 09:10, 3 January 2021 (UTC)

From what I can see, we have a Ceb article for a dam attached to a wikidata item for a reservoir. In the item, as you note, one ID thinks it is pointing at a dam - https://www.geonames.org/4750316/ca-ira-dam.html - and the other at a reservoir which is called a pond - https://geonames.usgs.gov/apex/f?p=GNISPQ:3:::NO::P3_FID:1464214 ... meanwhile the coordinate on the item points (in google maps) to a wooded hillside above the course of what may well be a dammed stream/river, in the vicinity of a structure which might be a dam. I think the desired structure for wikidata would be a dam item linked to a reservoir item using reservoir created (P4661) and, or couse, ideally having coordinates which actually point at the items on a map remembering that the coordinates in the sources are often not accurate, as in this case. --Tagishsimon (talk) 10:12, 3 January 2021 (UTC)
While sometimes both reseveoir and dam are mixed it's better when both have their own item and linked via dam (P4792). As far as geographic items in ceb-wiki go, see https://www.wikidata.org/wiki/Wikidata:WikiProject_Territorial_Entities/Geonames_and_CebWiki . Any information in the ceb wiki articles comes from Geonames. ChristianKl ❪✉❫ 22:03, 4 January 2021 (UTC)

Location of Lower Hope Point (Thames Estuary)

The location of Lower Hope Point in the Wikidata entry (Q24639320) appears to be incorrect. It is shown as 51°30'0"N, 0°29'0"E, which puts it on the north (Essex) side of the Thames. Consistent with this, the historical county is shown as Essex. But according to both an Admiralty chart and the Ordnance Survey website it is at 51°28'59"N 0°28'8"E (to nearest second) putting it over a mile to the S and clearly on the south (Kent) side of the River. The reference for the location is GEOnet Names Server feature ID -2602119. I cannot find a way to access this to check. What is an acceptable reference for the location? Can I use the OS website? If so what detail do I give? I have not done much editing in Wikidata so help would be appreciated. Kognos (talk) 12:02, 5 January 2021 (UTC)

Referencing it to an OS map would be fine. It's clear on this National Library of Scotland rendering of a 1960 OS map - https://maps.nls.uk/geo/explore/#zoom=14&lat=51.47981&lon=0.46657&layers=10&b=1&marker=51.482246,0.468369 at 51.482246,0.468369 so you might use that as a URL? Location / located in statements should be changed to point at the appropriate places; they'll likely have been set from the erroneous coodinate. --Tagishsimon (talk) 13:06, 5 January 2021 (UTC)
Tend to agree with using OS as the reference. However, I would suggest using the TOID (P3120) (4000000074746674) in the reference to link to OS open data site. Simon Cobb (User:Sic19 ; talk page) 19:30, 5 January 2021 (UTC)
@Sic19:Thanks for this,very helpful. I've edited the entry using TOID. Would you have a moment to check that what I've done is appropriate? Kognos (talk) 21:51, 5 January 2021 (UTC)
@Kognos: It's mostly ok - I've copied the reference from coordinate location (P625) to located in the administrative territorial entity (P131) so it can be verified. The only thing you might want to reconsider was the removal the GNS Unique Feature ID (P2326) claim. If it is incorrect, retaining the claim and setting the rank to deprecated with an appropriate reason for deprecated rank (P2241) qualifier would be preferable to prevent it from being added again in the future. For information about rankings see Help:Ranking. Hope that helps. Best, Simon Cobb (User:Sic19 ; talk page) 22:09, 5 January 2021 (UTC)
@Sic19:I'll check out the help as suggested. Thanks again. Kognos (talk) 22:18, 5 January 2021 (UTC)

Template for new people?

These representatives are new members of United States Congress (Q11268) and appear not to have any Wikidata item. I don't mind creating them myself, but I was wondering if there's a good template for people/representatives. I've not done this before. I can wait for some bot to pick them up as well. Just looking to some guidance. Regards, Gettinwikiwidit (talk) 13:49, 5 January 2021 (UTC)

All had items, apart from S000168 which is faulty (and probably refers to C000127). I added Bioguide-Information, but there’s a lot of work still to be done. --Emu (talk) 14:10, 5 January 2021 (UTC)
Hm.. They didn't have US Congress Bio ID (P1157) when I checked and reconciliation didn't match them. I believe that bioid is not faulty. See https://www.congress.gov/member/maria-salazar/S000168. Regards, Gettinwikiwidit (talk) 14:22, 5 January 2021 (UTC)
Yeah, reconciliation sometimes doesn’t really work reliably. You are right about María Elvira Salazar (Q6003715), I corrected the ID, but there is still no Bioguide info on her. --Emu (talk) 14:29, 5 January 2021 (UTC)
I'm still a little confused. The following still doesn't return anything, though I see that Jerry Carl (Q102277702) does have a US Congress Bio ID (P1157). What am I doing wrong? 
SELECT ?state ?stateLabel ?district ?districtLabel ?repName ?party ?bioURL WHERE {
  {
    SELECT * WHERE {
      VALUES (?state ?district ?rep ?bioid ?party) {
        (wd:Q173 wd:Q4705154 "Carl, Jerry L." "C001054" "R")
        (wd:Q173 wd:Q4705155 "Moore, Barry" "M001212" "R")
        (wd:Q99 wd:Q864997 "Obernolte, Jay" "O000019" "R")
        (wd:Q812 wd:Q5461103 "Franklin, C. Scott" "F000472" "R")
        (wd:Q812 wd:Q5461122 "Salazar, Maria Elvira" "S000168" "R")
        (wd:Q1428 wd:Q5547273 "Clyde, Andrew S." "C001116" "R")
        (wd:Q782 wd:Q1446201 "Kahele, Kaialiʻi" "K000396" "D")
        (wd:Q1558 wd:Q6364518 "Mann, Tracey" "M000871" "R")
        (wd:Q1166 wd:Q6837402 "McClain, Lisa C." "M001136" "R")
      }
      ?rep wdt:P1157 ?bioid.
    }
  }
  BIND(URI(CONCAT("https://bioguide.congress.gov/search/bio/", ?bioid)) AS ?bioURL)
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!
The answer is I'm an idiot. Need to use a different variable to bind besides ?rep. Gettinwikiwidit (talk) 14:34, 5 January 2021 (UTC)

Extending domain of 2 properties related to usage examples on lexemes

Please join the discussions Property_talk:P6191#Extend_domain_to_qualifiers_on_usage_examples Property_talk:P3865#Extend_domain_to_qualifiers_on_usage_examples--So9q (talk) 23:55, 5 January 2021 (UTC)

de Wiki '2020 United States presidential election' article misdirecting to nonexistent en Wiki article suddenly

Hi,

Not sure if this is the right place to ask because I don't know where the problem is.

The de Wiki article Präsidentschaftswahl in den Vereinigten Staaten 2020 was linked correctly to 2020 United States presidential election until a few days ago. All of a sudden it now links to a nonexistent article Felony disenfranchisement in Florida Artikel aus der englischsprachigen Wikipedia.

First thing I did was to check the site links at https://www.wikidata.org/wiki/Q22923830#sitelinks-wikipedia, but everything there looks OK, so I'm stumped.

What should I do next / whom to ask for help?

Best, == Peter NYC (talk) 22:28, 4 January 2021 (UTC)

The page has a rogue interwiki link, [[en:Felony disenfranchisement in Florida Artikel aus der englischsprachigen Wikipedia]], in the Swing States section. Ghouston (talk) 22:37, 4 January 2021 (UTC)
@Peter NYC: The Wikidata sitelink was overridden by a local interwiki after this edit. Now fixed. 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 22:38, 4 January 2021 (UTC)
@1234qwer1234qwer4: Thank you very much for the quick response! Just so I understand correctly, the badly formed interwiki link en:Felony_disenfranchisement_in_Florida within the article text caused the whole page to link incorrectly? Is this a general problem or exclusive to de Wiki, for whatever reasons? Best, == Peter NYC (talk) 23:01, 4 January 2021 (UTC)
@Peter NYC: This is the general formatting of interwiki links to the wiki with the language code written before the colon. It was used before the start of Wikidata, and is still used sometimes, when a sections is wanted to be the interwiki link, for example (although, if possible, an appropriate redirect should be linked to Wikidata instead, as the interwiki syntax would have to be inserted manually on each wiki). 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 23:08, 4 January 2021 (UTC)

@Peter NYC: If you did actually want to link to enwiki, you can use the format [[:en:Felony disenfranchisement in Florida article from the English language Wikipedia]] (notice the colon in front of en) MSGJ (talk) 11:42, 6 January 2021 (UTC)

Alternative account

Hello. I have an alternative account that I use in eswiki: User:Títere malvinero. ¿Can I use it here too?

Ok, I have already edit WD with that account, but just now I realised that maybe in WD the alternatives accounts it are prohibited. Regards, Malvinero10 (talk) 04:15, 6 January 2021 (UTC)

I disclose it. I use it in eswiki for minor edits and category edits. I see that this isn't a valid reason, so I won't use it in WD no more. Thank you for your attention. With regards, Malvinero10 (talk) 18:23, 6 January 2021 (UTC)

Template combines topics

Hello. Do we have property "Template combines topics"? Something similar to category combines topics (P971) for templates? template has topic (P1423) is not enough in some cases (as category's main topic (P301) is not enough for some categories). Of course is better to add P1423 if there is such item.

For example,

or

or

Data Gamer play 08:06, 6 January 2021 (UTC)

I am not aware of such a property - but agree it should be added to better describe the templates. Especially for the Wikimedia navigational template (Q11753321) there also should be the equivalent to category contains (P4224) where the second item is added as a qualifier. For these templates the equivalents to category related to list (P1754) and list related to category (P1753) could also be useful. Ahoerstemeier (talk) 18:27, 6 January 2021 (UTC)
Thanks. Wikidata:Property proposal/Template combines topics Data Gamer play 21:19, 6 January 2021 (UTC)

Inception of a building or bridge

We are mapping inception (P571) to the date that a bridge opened. What would be the best way to record the date that construction started? If with a qualifier, which qualifier? Thank you MSGJ (talk) 11:32, 6 January 2021 (UTC)

First, date of official opening (P1619) is arguably the best place to store, err, an opening date. Presuming inception (P571) is being used, it needs to be qualified with object has role (P3831) stating the nature of the date. There is also significant event (P793) which is good for a chronology of events with dates as point in time (P585) qualifiers. I don't think I'd store an opening date in inception (P571) for a bridge, but ymmv. --Tagishsimon (talk) 15:45, 6 January 2021 (UTC)
We're already mapping date of official opening (P1619) to "Inaugurated". I think it is possible that an official opening ceremony would not occur on the same day as the actual opening to the public. Or a bridge might just open without having an inauguration. MSGJ (talk) 11:22, 7 January 2021 (UTC)

I am interested in contributing protein-small molecule binding data from our BindingDB database

Dear Wikidata community

My lab runs the BindingDB database of protein-small molecule binding affinity data, bindingdb.org, and we are interested in contributing our data to WikiData.

We have about 2 million measurements of binding between about 8,000 different proteins and 900K small molecules. The data come from scientific articles and US Patents. Many of the data come with a CC BY attribution license. Another big subset comes with a CC BY-SA license. We can treat these two sets differently, if needed.

The proteins are outfitted with UniProt IDs and/or protein sequences. The small molecules have SMILES strings and InChI strings, and we can generate other identifiers for them if needed. The scientific articles are identifiable with DOIs, and the US Patent are identifiable with US Patent numbers. So it should be fairly straightforward to link our data in with other WikiData using these identifiers.

Would these data be a good fit to Wikidata?

Thanks, Mkg001 (talk) 13:57, 6 January 2021 (UTC)

WikiProject Molecular biology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

  • As far as licenses go, there's a general question to what extend this is simply factual data that isn't protected or whether it is protected. Wikidata itself is cc-zero, and I think there's a good case to be made that the factual nature of the data allows the import and thus it would be great to have the data in Wikidata. ChristianKl ❪✉❫ 17:48, 6 January 2021 (UTC)
  • Most important procedure-wise would be to avoid creating duplicate substances at all costs. Any protein in WD has a unique UniProt ID so any non-existing ID will need creation. Small molecules in WD have unique InChi keys (modulo exceptions) so any substance with InChi key not in WD is good to create. The central addition will be a statement A physically interacts with (P129) B, and as reference the statement stated in (P248) Reference where Reference is the WD item of the paper. So you face similar task as the recently started Wikidata:WikiProject_Chemistry/Natural_products, except your data is already digital, why not learn from them? Best wishes, have fun with Wikidata. --SCIdude (talk) 10:20, 7 January 2021 (UTC)
    • PS You can see the >9,700 already existing statements with this query, so there is another type of duplicate to avoid:
SELECT DISTINCT ?item1 ?item1Label ?item2 ?item2Label
{  
  ?item1 wdt:P129 ?item2
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!

Something wrong with Property:P3056

I think there's something amiss with formatter URL (P1630) on TCM Movie Database person ID (P3056), but I can't figure out how to fix it. When you click on the auto-generated URL on an item that uses this property (e.g. Tanya Roberts (Q232356)), the TCM website returns a 404. They use a URL format that I've never seen before, and I can't figure out how to fix the error. (Apologies if this post should go elsewhere—I saw that not many people were watching TCM Movie Database person ID (P3056), so I thought I'd post somewhere more central.) AleatoryPonderings (talk) 17:15, 5 January 2021 (UTC)

haven't investigated fully but on first glance seems they changed their url format. At least some of the URLs used to work. We need to figure out if there is a new format or if we should re-point the property at the archive. BrokenSegue (talk) 17:58, 5 January 2021 (UTC)
I adjusted the value to one that works. Ghouston (talk) 21:59, 5 January 2021 (UTC)
It seems they changed their format some time to use two numbers separated with a bar. Other items such as Stanley Kubrick (Q2001) also have values that don't link. Ghouston (talk) 22:07, 5 January 2021 (UTC)
IIRC, they've been using the two number format for years. I think the first number is the actual ID used by TCM itself, the second one is American Film Institute person ID (P5340) (probably for entries with data they got from the AFI database). If no AFI ID exists - or they simply didn't get their data for that entry from AFI - the second number is a zero. The links worked fine with just the first ID for many years. But the website has been acting strange at times. Two or three years back they changed the URLs and the new URL simply didn't work for various people (including me) and just redirected everything to https://www.tcm.com/unavailable. The old URL still worked at totally random times for a while but then also stopped working alltogether. And now they seem to be switching to a new design, who knows what they're doing with the URLs now. But I'd be hesitant to switch to using both numbers as the ID right now, even if it makes the link work at the moment. --Kam Solusar (talk) 18:09, 7 January 2021 (UTC)

"ORCID public data file 2020" in references statement

Hey, I improving OrcBot which connects authors and their publications. Besides other stuff the new version will introduce the reference statement for the registration as author (P50) author which is "ORCID public data file 2020" .

What would be the most appropriate property for "ORCID public data file 2020" as references stated in (P248)? We have: 

1) ORCID, Inc. (Q19861084): non-profit organisation that issues ORCID Identifiers  
2) ORCID EU (Q27050209): European organisation promoting ORCID
3) ORCID iD (Q51044): nonproprietary alphanumeric code to uniquely identify scientific and other academic authors

None appears appropriate to me. what do you think? Thank you!  – The preceding unsigned comment was added by EvaSeidlmayer (talk • contribs) at 14:32, January 6, 2021‎ (UTC).

@EvaSeidlmayer: Agree with ArthurPSmith and I see you have created ORCID Public Data File 2020 (Q104707600).
Do you intend to include any other claims in the reference statement, such as ORCID iD (P496)? Although it might be less of an issue with works than affiliations from ORCID, I think including the put-code for the work will be beneficial, especially for bibliographic data that was input manually by the author, because data in ORCID records can change, be deleted or become inaccessible due to changes to visibility settings. I have been verifying imported employer (P108) and educated at (P69) claims to catch reconciliation errors and without the put-code it is, in some instances, very difficult to know whether the reconciliation was incorrect or the record has been updated as I cannot be certain I am referring to the same data as used in the ingest. Simon Cobb (User:Sic19 ; talk page) 00:03, 8 January 2021 (UTC)

Kill a runaway QuickStatements batch

I recently submitted a QS batch via [7], which I've used successfully & extensively in the past, but it appears to be hanging now with no feedback. I checked my contributions, and it should have only created 31 new items & populated them with some information, but instead it just keeps creating hundreds of blank new items... —Tom.Reding (talk) 18:14, 7 January 2021 (UTC)

@Tom.Reding: I blocked your account for 24 hours for editing in the main space so the tool won't be able to create more empty items for now. That gives us some time to figure out how to stop it. Multichill (talk) 18:23, 7 January 2021 (UTC)
I'm going to delete the empty items. --Martin Urbanec (talk) 18:41, 7 January 2021 (UTC)
IIRC your own computer makes that edits, so closing _should_ be enough Tom.Reding. Feel free to ping me at IRC (Urbanecm) if you need any real-time assistance. Best, --Martin Urbanec (talk) 18:53, 7 January 2021 (UTC)

@Tom.Reding: I just unblocked your account, hoping to see the script already finished. No items seem to be created now, so I guess it's fixed now? --Martin Urbanec (talk) 20:35, 7 January 2021 (UTC)

@Multichill, Martin Urbanec: thank you both for your help. I'm still not sure why this happened - I used the same local files to create my QS commands as I did last time in November 2019, and the QS documentation doesn't seem to have changed materially since then either (the message to use a newer version of the tool existed then, too). I'll use the newer version, and possible retry the older one later, but only on batches of 1 at first. —Tom.Reding (talk) 23:32, 7 January 2021 (UTC)

Preferred country of London (Q84)

The page for London is locked so I can't edit it. Can someone put the preferred country as the most recent one? This also seems like a good opportunity for a bot to go through and set these. Lordgilman (talk) 21:59, 2 January 2021 (UTC)

The most recent value (per start date) is already the preferred one. --- Jura 12:13, 8 January 2021 (UTC)

Labels for languages that don't have an existing Wikipedia project

Are there any plans to allow language labels to be added to Wikidata items, where the language has its own valid ISO 639-3 code, however does not (yet) have its own Wikipedia project? For example, the ISO 639-3 code for Okinawan Ryukyuan is "ryu", however currently users are unable to add ryu labels to items. Also of note is that English Wiktionary has hundreds of Okinawan entries. Benlisquare (talk) 16:08, 5 January 2021 (UTC)

List

How to make a list on Wikidata item? I guess something other than is a list of (P360) schould be useed. Eurohunter (talk) 20:20, 5 January 2021 (UTC)

@Eurohunter: not sure what you mean. do you mean as in list of presidents of the United States (Q35073)? BrokenSegue (talk) 23:13, 5 January 2021 (UTC)
@BrokenSegue: Yes. So how we could list presidents at WD item? Eurohunter (talk) 23:20, 5 January 2021 (UTC)
@Eurohunter: The list of US presidents is here. BrokenSegue (talk) 00:43, 6 January 2021 (UTC)
@BrokenSegue: I'm talking about Wikidata. Btw. why your bot do such edits? What about Facebook, YouTube and other social media platforms? Eurohunter (talk) 08:43, 6 January 2021 (UTC)
@Eurohunter: because that is what the bot was approved to do. you are free to make your own bot for those platforms. BrokenSegue (talk)
  • Is this what you mean? See for example the one I just created here: Talk:Q50579815. Create a position entry in Wikidata. Go to the position that you want to make into a list and click on the discussion page and add {{PositionHolderHistory|id=}} fill in the id= the Qid for that position, then hit Refresh and it will do the calculations and formatting. It expects a start date, and end date, an ordinal and preceded by, and succeeded by. It will sort on start_date when they are filled in. Here is a fully formatted one I did last month: Talk:Q47459093. --RAN (talk) 08:52, 6 January 2021 (UTC)
Thing of beauty, RAN. So to sum up: we do not solve the "List of Presients of the USA" by trying to list each of them within a single item. Rather we have discrete items for each person, and we indicate using a position held (P39) statement that they held the position of President. And then we use SPARQL, or tools such as PositionHolderHistory, to provide the compiled list on demand. (And in addition we have item(s) that point to list articles held on language wikipedias; again, such items do not list the Presidents within the item, but rather point to a list somewhere else.)--Tagishsimon (talk) 09:58, 6 January 2021 (UTC)
@Richard Arthur Norton (1958- ): @Tagishsimon: What is the point to have list on talk page? There should be property to list anything like you list awards or organisations etc. Eurohunter (talk) 12:30, 6 January 2021 (UTC)
It's not an efficient way to do things in wikidata. It is possible to list all office-holders witin a single item; so for President of the United States (Q11696) we could add 45 officeholder (P1308) statements. We don't. Instead, just the current one - Q11696#P1308. Wikidata is a database. It is about splitting information into small units, and using reporting to bring data together to produce large lists. It is not for storing those large lists as lists in a single item. As to why: mainly, it's very unwieldy. Consider Lord Mayor of London (Q73341) where there are about 600+ position holders. The item - the record we need to edit - becomes impossibly big and ergonomically hugely difficult to maintain, quite apart from causing technical problem because of its size. And it's not required because, as demonstrated, when we want a list of presidents we can get one, easily, with a report or via a tool. Language wikipedia text pages are good places for single-document lists. Wikidata has another purpose entirely, which is to hold that data that allows such lists ton be produced, but not to hold the list as a list. --Tagishsimon (talk) 12:46, 6 January 2021 (UTC)
  • You asked: "What is the point to have list on talk page?" You need a place for it to display so that you can work on it to correct flagrant errors. You are ignorant of all the errors in our data until you see it displayed, with error messages built in. When you look at Talk:Q73341 aren't you troubled by all the error messages? Were you even aware that there are inconsistencies in the dating and numbering and successions before you looked at the generated table? It took me a week to build the Rutgers table, and it took me a month to build one for the mayors of a large New Jersey city, it will take me another month to finish fixing all the errors in the pope data. The beauty of the Wikidata table, that the Wikipedia table, does not address is that once completed the table can be translated into other alphabets like Greek and Chinese and Cyrillic and Arabic, so long as they have been filled in at Wikidata. You would have to build the table at the Greek Wikipedia from scratch. --RAN (talk) 18:29, 6 January 2021 (UTC)
  • The list is useful, but is the talk page the best place to put it? Perhaps it should be somewhere on the public web, where it shows up in search engines, so that others can find it. Ghouston (talk) 21:54, 6 January 2021 (UTC)
@Ghouston: The PositionHolderHistory lists on talk pages are a convenience for users who are trying to ensure that all position holders are listed in wikidata They're a tool. The most useful place for such a tool designed to assist in ensuring that all position holders are listed in wikidata, is on the talk page of the position item in question. I hope that clears this up. --Tagishsimon (talk) 15:06, 8 January 2021 (UTC)
We still have somewhat redundant officeholder (P1308) for this. --Matěj Suchánek (talk) 09:57, 7 January 2021 (UTC)

Data donation - worldwide information on banks, hospitals and universities

Dear WikiData Community,

I am writing this post to inquire whether the community would be interested in data spanning majority of world's hospitals, banks and universities. We have been involved with a start-up that for over 2 years has gathered basic and detailed information on those institutions, however due to financing issues with COVID-19 we had to cease our operation. We have over 120 000 institutions gathered.

For example: Please use the link below to access example data: https://drive.google.com/drive/folders/11Y-ohbYvPyMTPebj5hFM-wxIuINqUNbz?usp=sharing For hospitals -> look at this Canada File attached below in row 179 in Excel and you can see this hospital with general information alongside more specific info such as their departments: which goes along with https://www.wikidata.org/wiki/Q7914187 and https://en.wikipedia.org/wiki/Vancouver_General_Hospital and https://arz.wikipedia.org/wiki/مستشفى_فانكوفر_العام

I have also attached two files representing our research on banks and universities and what data we have gathered. Universities in Spain and Banks in South Korea.

All of the data in those sheets has been collected directly by our researchers, who were scholars at London's universities all across the globe through either internet research or calling up those institutions. Previously, we had all of the data displayed on our website, however as we have shut down we had to close this too. I could supply all of the data through Dropbox sharing option so anyone can have access to it, however, our main questions is how all of this could be updated on those respective wikipedia pages?

Thank you in advance, Kacpermic (talk) 13:06, 8 January 2021 (UTC)

@Kacpermic: Definitely this data looks useful. 120k institutions is a lot. The most difficult thing would be matching your identifiers to ours (and potentially creating new entities). I'm guessing your data doesn't have a unique identifier that trivially joins to us. I assume the "Hospital_ID" is your internal identifier. Maybe our best bet would be to use mix n match to match the data (we can match on name/url/city/etc). I would suggest uploading the raw data to achive.org so we can point our references for the data there (And you can write an explanation of how it was generated, give it a name, etc). Once we have the data joined we can use a bot to import missing pieces of data. BrokenSegue (talk) 15:21, 8 January 2021 (UTC)
@BrokenSegue: Thank you for the response. Yes, our identifiers are based on ISO lists that I could share here, if it helps? I will sort out through our data and upload it to the archive.org in the coming days and reply to this thread. Thank you.
@Kacpermic: unsure what the ISO lists are (this isn't my area of expertise) but any additional data would be good. If there's a standard identifier we don't currently support for these institutions then we should probably add support. BrokenSegue (talk) 15:56, 8 January 2021 (UTC)
  • If it's easily split, maybe the three fields could be separated. Depending on the volume, maybe also per country (or separate for some countries with many entries). I think our coverage is fairly good for universities, it varies a lot for hospitals from one country to the other, and I'm not really sure about the third (except for India). --- Jura 15:32, 8 January 2021 (UTC)

@BrokenSegue: and :@Jura: thank you for the comments. I think I will proceed with creating three different folders - one Health, Universities and Finance and within them we can have a split by country for each one of them and I can provide further explanations on our own internal identifier. If our data is in the template, is it ok if I keep the template rows? For example in Hospital file you can see the one in first few rows: Name, Link etc... or should I clear those and explain it somewhere? 16:04, 8 January 2021 (UTC)

my guess is that minor things like that won't matter as some manipulation will have to be done manually no matter what. the public upload is mainly good as a canonical location for the data. Also you should probably note that you are releasing it into the public domain or cc-0 since we require that. BrokenSegue (talk) 16:33, 8 January 2021 (UTC)

Tabular data in Wikipedia for updating Wikidata

Hello,

Apologies is this is the wrong place for this question. Some light background - I've been attempting to look at federal judges appointed by US presidents. I was able to scrape a couple different sources and get all the information I need. However, while I was doing that I noticed that there's some (understandable) inconsistency in the data that Wikidata holds on the judges. For instance for most modern judges such as Sri Srinivasan (Q7586290), their dates of service and appointing president is associated with the position they held position held (P39). That is not true of older judges, e.g. Learned Hand (Q493155) - as far as I can tell dates of service and appointing president are not present anywhere on the wikidata item.

That information is already present in tables on Wikipedia for both judges (here and here respectively) in the same format but on different pages. I think it would be cool to have better coverage of historical judges as the data is already there. All that for my potentially dumb question:

Q - Is tabular data from Wikipedia like that present on the above pages pulled into anything that is easy to access directly for general purposes, and more specifically for then updating Wikidata? Or do the wiki pages themselves have to get scraped to extract that information and subsequently update wikidata?

I took a look through what I could find in tools for Wikidata and on Wikipedia, but I struggled a little bit to navigate through so I easily have missed something. Let me know if I'm off base completely or if there are any other questions. Thanks.

RBAnderson (talk) 18:44, 8 January 2021 (UTC)

@RBAnderson: currently there isn't a great way to store tabular data on wikidata. Instead we store the data on each item and then write a SPARQL query to convert that into a table. So for example you would write a query that says "show me everyone with the position 'US federal judge' who was appointed by 'obama'". Someone can help you write such a query if you need help. It is possible as you say wikidata lacks the information to retrieve everyone properly but you or someone else can go in and annotate all such judges with this information. BrokenSegue (talk) 18:57, 8 January 2021 (UTC)
@BrokenSegue: Makes sense, I've got the query and the list of judges that are missing the information. Good to know I didn't miss any obvious data repositories. I'm on my way, thanks for the quick answer! RBAnderson (talk) 19:56, 8 January 2021 (UTC)
  • So far no general tool is available for migrating data from Wikipedia tables to Wikidata tables, but people have devised bots for specific tasks, I see that the pope table must have used a bot because of the duplication. Where I have been building lists of governors I have had to migrate the data by hand. The nice thing about Wikidata is when you build the table for a position holder, you have tools that flag errors in the succession of a position, and the dates. In many cases there are errors at both Wikidata and at the various language Wikipedias where the date was imported. See for instance the one I am working on for popes Talk:Q19546. Sometimes there is duplication and quadruplication, because each language Wikipedia uses a different start date. I discussed earlier in the week on a script that creates the tables:
"Go to the position entry in Wikidata and click on the discussion page and add {{PositionHolderHistory|id=}} fill in the id= the Qid for that position, then hit Refresh and it will do the calculations and formatting. It expects a start date, and end date, an ordinal, and preceded_by, and succeeded_by. It will sort on start_date when they are filled in. Here is a fully formatted one I did last month: Talk:Q47459093." --RAN (talk) 01:18, 9 January 2021 (UTC)

Q18609226

Quick question: do we have a property corresponding to a work (not author) ID for works in the Trove (Q18609226) library catalogue? I see Trove newspaper ID (P5603) and NLA Trove people ID (P1315), but not one for works. I figured it must exist, but didn't see it in my search. Would be useful for The Doctor and Student (Q104636335) (Trove identifier is 7077765), among many others. If not, I'll set up a property creation request. AleatoryPonderings (talk) 18:55, 8 January 2021 (UTC)

I think you are right, the property is missing and worth creating. There's also a separate namespace for full text newspaper articles, such as https://trove.nla.gov.au/newspaper/article/122267775. Ghouston (talk) 22:04, 8 January 2021 (UTC)

Site links to MediaWiki, Meta-Wiki, and Commons, etc

Why don't the site links to MediaWiki, Meta-Wiki, and Commons, etc (the wikis that use Extension:Translate) automatically use Special:MyLanguage? That way items like Huvudsida (Q21169674), Main page/de (Q21167895) and most of the others on Wikimedia main page (Q5296) in the said to be the same as (P460) section would be rendered obsolete for instance. --Sabelöga (talk) 21:47, 8 January 2021 (UTC)

Conflation

Q104603619 is a conflation of at least three people with the same name, I started teasing them apart, but I need help. I assume we will keep 104603619 for the person born in 1962, and tease out the other two artists that already have entries. I don't think any of the Identifiers belong to the 1962 person. I am not sure if the landscape architect is even a notable person. --RAN (talk) 03:28, 2 January 2021 (UTC)

  • Maybe @Tm: should be made aware of it. See Help:Conflation of two persons for a how-to. --- Jura 10:16, 2 January 2021 (UTC)
    • What's current recommended best practice when multiple external sources (eg library catalogues, VIAF) conflate two people, or at least confuse some aspects about them, such as dates or works?
Eg an example I came across recently John Talbot Dillon (Q6260196) (1739-1805) vs John Talbot Dillon (Q19361094) (1734-1806), who were routinely conflated until the 1960s, and still quite often confused (eg: as to who wrote what, and which dates apply to who) in many library catalogues (and, indeed, by en-wiki until last week). Which item should confused library catalogues be linked to?
Should we try to distinguish, and hope that VIAF will notice our assignments are now different to old ones it has copied? Should a new 'conflation' item be created, and the links dumped there? Or, given that library catalogues should probably mostly represent the one who wrote the books (John Talbot Dillon (Q19361094)), should we link them to him regardless of any biographical data they present? Will bots merely re-write VIAF's assignments here, whatever we manually change? Interested to know people's thoughts. Jheald (talk) 12:49, 2 January 2021 (UTC)
A bit of previous discussion here: /Archive/2020/11#How_best_to_handle_conflations_in_external_identifiers? Jheald (talk) 13:16, 2 January 2021 (UTC)

Carlinhos

May someone merge Q1041746 and Q21285777? I tried on my side, but this is my first time and am not quite getting it.--Ortizesp (talk) 01:44, 9 January 2021 (UTC)

Carlinhos (Q1041746) and Carlinhos (Q21285777): it's against the Wikidata policy, although this is an example where I think it would do no harm to put all the sitelinks on Carlinhos (Q21285777), and get rid of the disambiguation item. Ghouston (talk) 02:12, 9 January 2021 (UTC)
I think there is some error. Did you check the sitelinks? Most are disambiguation pages so they should go on the item for disambiugation pages. As you said, we don't mix the two. --- Jura 07
  • 55, 9 January 2021 (UTC)

The metadata of many book reviews is incorrect at Crossref, Google Scholar...

At the moment I am adding all citations using the "cites work" statement. When a paper has no links I google for it.. Often there is a link to Crossref and/or Google Scholar. This is one example.. Q93615130 is the item ... When you analyse the statements, they all refer to Crossref, they refer to the reviewed book and are NOT about the review. Consequently, MANY books have a DOI as well but they are of the review NOT of the book. Given the prevalence and the small set of data, it is likely a huge issue and not so much an incident.

I have reached out to Crossref, I have corresponded with Roderic Page who added the item from the data at Crossref. Thanks, GerardM (talk) 10:13, 3 January 2021 (UTC)

GerardM: I'm a little confused. When I run the doi 10.2307/4109011 on Crossref, it returns https://search.crossref.org/?from_ui=&q=10.2307%2F4109011+, and that's tagged by Crossref as a book review (or, rather, a journal article), not a book. AleatoryPonderings (talk) 16:51, 3 January 2021 (UTC)
The review has only one author. Many of such reviews are linked as citations in stead of the work they review. Many books DO have a DOI as well as a ISBN however, quite often the review is used in stead of the book. Thanks, GerardM (talk) 18:01, 3 January 2021 (UTC)
The review was published in Kew Bulletin (Q2675794) and not the book itself with was published by Blackburn Press. The statements in the item seem correctly about the review and not the book. ChristianKl ❪✉❫ 11:53, 4 January 2021 (UTC)
Did you check the history ... I needed that item. thanks, GerardM (talk) 19:48, 10 January 2021 (UTC)
  • I think I once noticed a case where the review metadata (from cross-reference) was added to the item about the work here. The reviewer ended up being seen as a co-author. --- Jura 12:19, 8 January 2021 (UTC)

"Compress" same statements with different ref URL

Hello, is there bots to merge two "exacts" same statements into one? Same value, same point in time (P585), but different references . Eg : https://www.wikidata.org/w/index.php?title=Q630524#P3872 : go see the year 2019 ; it should be reading only one statement, not two. Thanks --Bouzinac 💬●✒️●💛 13:08, 7 January 2021 (UTC)

My bot has a subroutine for this. However, when I have it run over that item it does nothing. There is probably some internal "invisible" difference that makes Pywikibot treat them as different. --Matěj Suchánek (talk) 08:23, 8 January 2021 (UTC)
Dates are different: one has "+2019-00-00T00:00:00Z", the other "+2019-01-01T00:00:00Z"; both with "Precision 1 year" [8]. --- Jura 11:34, 8 January 2021 (UTC)
I understand what you said, but to my functional point of view, they look the same.
Could your bot consider these precisions as the same technical date ? Other example of pb with month precision : Q854130#P3872, see eg months of 2019 Bouzinac 💬●✒️●💛 15:41, 8 January 2021 (UTC)
You made me do it. --Matěj Suchánek (talk) 11:22, 9 January 2021 (UTC)
Good! --Bouzinac 💬●✒️●💛 16:19, 10 January 2021 (UTC)

Abbey of Monte Cassino (Q765737, Q334051)

I think "Territorial Abbey of Mont-Cassin (Q765737)" and "Abbey of Monte Cassino (Q334051)" should be merged. There is no difference!--Utilo (talk) 10:14, 10 January 2021 (UTC)

No. Per the italian wikipedia articles (which in any event preclude the merge) one is mainly about the C6 building, the other about a regional territory of the Catholic church. Obvs there is a great deal of overlap, but equally it is possible to tease out differences. Please take a little more time to investigate properly, especially, things linked to items that you suppose should be merged. Whilst the need to merge is frequent enough that your report really did not need its final exclamation mark, sadly it is also a commonplace for users to suggest or enact merges becuse they have not troubled to look properly at the information before them. --Tagishsimon (talk) 10:43, 10 January 2021 (UTC)

How to develop and deploy documentation for Q items ?

Here is a request for comments : Wikidata:Requests for comment/How should we develop and deploy documentation for items ? PAC2 (talk) 17:55, 10 January 2021 (UTC)

BorkedBot non-valid additions

User:BrokenSegue's bot BorkedBot has been non-validly adding (&limit=2500) social media followers (P8687) as sole statement to person's items filling data with number of followers – right for Twitter accounts. It is biased (maybe very biased) addition and should be reverted (by bot, there is a lot of pages, if possible earlier because following edits can prevent undo). --5.43.83.177 23:40, 7 January 2021 (UTC)

Addition: Instead of this, diffs should look like this (for the item I gave as example, that number was duplicated by adding it in a wrong place, as a statement; only change that can be useful there is to give number of followers/subscribers with social media followers (P8687) instead of number of subscribers (P3744) but that can be other bot task as broader-meaning P3744 is used widely too accompanied with point in time (P585) for social media services beside other website services such as databases as I know). --5.43.83.177 23:54, 7 January 2021 (UTC)

With all due respect, you are mistaken. that way of adding this data is no longer best practice. BrokenSegue (talk) 00:01, 8 January 2021 (UTC) (sorry for being so harsh but you could've privately messaged me and I would've explained your confusion but you choose this venue instead BrokenSegue (talk))
@BrokenSegue: You got "approval" (not from me, so that is not kind of approval at all but 'partial consensus' if that exists) for "Task/s: Populate meta-data/qualifiers for entities with Twitter username (P2002) and YouTube channel ID (P2397). Qualifiers I will populate include: number of subscribers (P3744), Twitter user numeric ID (P6552), start time (P580), number of viewers/listeners (P5436) and named as (P1810).". You populated items with social media followers as Statements statement (adding Twitter numbers, which is completely arbitrary [why not add Facebook, then YouTube, then Twitter, then other social media followers numbers then in that statement qualifiers], plus controversial because Twitter was used for allegedly, effectively or not, bot-campaigning US politicians [one reason why it is biased, aside from obvious inconsistency, and you weirdly say "that way of adding this data is no longer best practice.") instead of adding it as qualifier to Identifiers statement (what's usual use of that property and what was intended task, as you said: "... Qualifiers I will populate include..."). Regarding User:Trade's comment on the page Rfp you cited, "Regarding Twitter: named as > Twitter user numeric ID > number of subscribers > start time > has quality > point in time. [...]", I think it was better to arrange them Twitter user numeric ID > start time > named as > number of subscribers [or social media followers, which one to choose is question also for YouTube which obviously has or should have followers but website itself cites 'subscribers' for channels even if it's free] > point in time > has quality. Point in time should follow qualifier it can be applied to (number stats). --5.43.83.177 00:21, 8 January 2021 (UTC) [e]
please read the full conversation in the request for permission and remember that the policy states that "Bots must stay within reasonable bounds of their approved tasks. The general guideline is to use common sense, and if in doubt, file another request for approval" which this falls under regardless. It is not my fault you did not participate in the request for permission discussion. We cannot store multiple number of subscriber counts using the method you describe which is the point of the new property. Also I have no idea what "plus controversial because Twitter was used for allegedly, effectively or not, bot-campaigning US politicians" has to do with anything involving this work. I intend to do the same thing for the other social networks later but we have the best coverage for twitter at the moment. It seems like you have an axe to grind that has nothing to do with me. BrokenSegue (talk) 00:30, 8 January 2021 (UTC)
No need to read the full conversation if you are doing both wrong and outside of your task. I do not want to bite you but say that these need to be reverted; bounds of tasks (e.g. possible coverage, number of qualifiers operated) is not same as wrong-way adding properties (in this case, properties as statements instead of as qualifiers of Identifier statements). You might have used sense but it is still wrong, and not falling under that request (your bot's contribution seemed good until 2021 started, changing youtube qualifier number statistics can be compared to what needs to be done with Twitter qualifiers, it's basically same). It's false argument; it is not your fault, true, but consensus is still invalid unless everyone participated (if we want consensus and not foolish show-off terms). It is old property and should works same as for YouTube channel (as done in these edits); everything can be stored. It could be related to recent contributions of BorkedBot because you did not stop to see discussion outcome but are still adding it wrong way, in the midst of problems in the US. Do not do that for other social networks because these changes will be reverted; that falls under Identifiers statements' qualifiers, why would you duplicate it (instead, number of subscribers should be changed to social media followers as Twitter username's (P2002) qualifier). Who 'we'? Wikidata is not made for coverage but for having structured data in one place. I simply don't see birth date or birth place statement comparable to social media followers as statement. --5.43.83.177 01:24, 8 January 2021 (UTC)
I cannot follow what you are saying (appears we have a language barrier). Nothing about "problems in the US" has anything to do with anything. You don't seem to get that the old way of doing this had limitations and so a new way was introduced which is what I'm following. You are simply mistaken about "number of subscribers should be changed to social media followers". I'm unsure why you are so confident you are right when you appear to have limited experience here. Please wait for more people (with experience in this area) to weigh in before saying things you might regret. BrokenSegue (talk) 01:43, 8 January 2021 (UTC)
@ChristianKl: I don't want to comment this that way. Why is registered account or comment with signature more valued? --5.43.83.177 23:55, 9 January 2021 (UTC)
@5.43.83.177: registered accounts have stable identity and as such can be trusted based on previous behavior. Nonregistered accounts that raise issues against registered account might be nonregistered because of a desire to avoid accountability. ChristianKl ❪✉❫ 12:12, 10 January 2021 (UTC)
Please stop the Bot at once. The Twitter Followers are already present in the objects, so the bot is adding qualifiers that the object already has. It would be tedious to revert all edits by this bot regarding Twitte by hand. Sigh. --Gereon K. (talk) 13:56, 8 January 2021 (UTC)
@Gereon K.: if consensus were reached to revert I would revert it all with my bot. So no worry about tedium. I do not expect that to happen as it is doing the right thing and people are just confused. BrokenSegue (talk) 15:05, 8 January 2021 (UTC)
You are not even able to explain why would that be right. It is expected from your side to stop doing that action (if not to revert what's already done, just because you want more proofs even if it is obvious) until its described positively or negatively in a conclusion... --5.43.83.177 23:55, 9 January 2021 (UTC)
I'm sorry. I can't fully make out what you are trying to say here. I am not going to stop a bot because one anonymous user has a problem with it. If consensus is reached that this is a mistaken I will undo it all. There is no rush to act. BrokenSegue (talk) 00:43, 10 January 2021 (UTC)
  • The bot's approach is actually consistent with the data model we chose. The previously favored approach was to add the same Twitter ID multiple times, but that didn't really work out. --- Jura 15:01, 8 January 2021 (UTC)
Could you explain how it's consistent and what's that 'data model' (I don't see that first of two sections, Statements in Statements and Identifiers, is free to define in it just any description one finds – especially when it is unpopular or controversial, and in this case it's actually inconsistent because there is not only one social media to define its followers [property is named 'social media followers'; cannot be compared to instance of, sex or gender, date of birth, official website etc. that require one value nor occupation, educated at, genre etc. that require one or more values as it other way implicates more values])? Your next sentence is unclear and might be contrary to previous one, so needs explanation. --5.43.83.177 23:55, 9 January 2021 (UTC)
"especially when it is unpopular or controversial". It isn't, and it isn't. hth. --Tagishsimon (talk) 23:58, 9 January 2021 (UTC)
It's clear that Twitter is rubbish, unused and nonsense website, so it is unpopular. It is controversial because bases itself on hashtagging instead of content (in difference to Facebook, Instagram, YouTube etc. which have applicable and normal flow use). I don't know what's current purpose of which but the latter mentioned sites are at least with supposed applicable usage. --5.43.83.177 00:14, 10 January 2021 (UTC)
You are making a ton of unrelated claims to support your position all of which are wrong. Twitter is popular in many countries and is widely used by important figures. Hashtags are not unique to twitter. There is no intent to favor twitter here as I intend to do this for e.g. youtube later. Even if twitter were totally terrible we already store follower counts but in a schema that is problematic. All I am doing is moving the data to a new schema. I do not see how this hurts wikidata at all. I'm guessing english is not your native language so you may not realize this but you are coming across as slightly unhinged. I will not reply to you further unless someone else steps forward to oppose the bot. BrokenSegue (talk) 00:40, 10 January 2021 (UTC)
I should have read this thread bottom-up, because that "Twitter is rubbish, unused and nonsense" comment makes it really easy to dismiss this complaint.--Matthias Winkelmann (talk) 01:12, 10 January 2021 (UTC)
And that would be both ordinary mistake and logical error. --Ageuser (talk) 21:02, 10 January 2021 (UTC)

I'm tired of reverting or formatting all pages but you now have task to format that statement like this or it is not valid or consistent and thus needs to be reverted sooner or later to get desired format. Those websites are ordered: first useful and normal Facebook (which does not have publicly available [since recent times] followers or subscribers so can be omitted in the discussion), second (as unneeded also-as-Twitter:-rubish Facebook's limpet) Instagram (which disables access to its website but numbers can be get via archiving URL in https://archive.is/), third YouTube (which also some way hides numbers by not giving full number but thousands precision), [...], and Twitter (which hides day in the date of account creation) as last. --Ageuser (talk) 21:02, 10 January 2021 (UTC) [e]

Ageuser, it would be helpful if you knew what you were doing. You do not. Reordering statements on an item - which is what I understand you to be doing here - is just nonsense. For a set of statements on social media audience, it does not matter, at all, what order the statements are presented in. I really cannot even put into words just how unnecessary and futile your action is. You have been here fpr 5 minutes. Please take it from those who have been here every day for years that you are not helping. In very short: items are a collection of statements. They are mainly designed to be discovered via reports of one sort or another. The presentation order of statements through the manual edit user interface is a matter of little or no moment; especially where the preferred order you specify has no relation to anything beyond your own preference. --Tagishsimon (talk) 21:13, 10 January 2021 (UTC)
I will arrange them for myself however I want; currently I like to follow present everything-cleaned-up style. --Ageuser (talk) 01:08, 11 January 2021 (UTC)
@Ageuser: Reverting is a waste of time. My bot can edit much faster than you can revert and will redo whatever you revert. There is no need to "reorder" things as wikidata has no concept of "order". Also you are clearly evading a ban so I will notify WD:AN promptly. It didn't have to end this way. BrokenSegue (talk) 21:14, 10 January 2021 (UTC)
You obviously want then your edit as last and don't want to cooperate.
I disagree with that about ordering.
We are alive so be happy!
--Ageuser (talk) 01:08, 11 January 2021 (UTC)

I am not sure if there is a consensus to record in items social media followers numbers of multiple social networks on a regular basis, as it requires a huge number of edits, which is evident from the number of edits of BorkedBot. For example in the case of weather data, there is consensus to use Commons tabular data (see for weather history (P4150)). --Jklamo (talk) 10:48, 11 January 2021 (UTC)

@Jklamo: this is a legitimate concern. I personally think there is value in having the data searchable via SPARQL. Plus we were already storing this data just in a worse format. If we record the data annually then we're talking about 260k updates per year for twitter (the rest are lower volume). The works out to about 0.1% of all edits in a year. However, I would also be willing to compromise and only add social media data for accounts with lots of followers. Currently I only add data for accounts with >1k followers but that could be 10k or 50k. That would massively reduce the edit volume and preserve much of the value. BrokenSegue (talk) 14:49, 11 January 2021 (UTC) The 50k limit reduces the number of edits by 75%, 10k halves it and in my opinion both preserve most of the value. BrokenSegue (talk) 14:59, 11 January 2021 (UTC)

Associated Wikipedia page removed. Should entry be removed too?

Crystal Lake (Q5191245) points to a water feature in a mobile home park. The associated Wikipedia page was removed. Should we remove this entity as well? Gettinwikiwidit (talk) 12:38, 10 January 2021 (UTC)

No. Why on earth would we? It satisfies Wikidata:Notability criterion 2 per the GNIS ID. --Tagishsimon (talk) 12:49, 10 January 2021 (UTC)
@Tagishsimon: I just added the GNIS ID, which actually points to the Mobile Home populated area rather than the water feature itself. Should we remove that and then delete the entity? Gettinwikiwidit (talk) 12:57, 10 January 2021 (UTC)
My inclination is still to keep the lake item; I imagine it will still satisfy criterion 2, although perhaps not via GNIS nor Geonames. It satisfies C3, the structural criterion, in that it gives its name to the settlement built around it (named after (P138)); on which subject, the GNIS ID should not be on the lake item, but on an item for the village. There's also a geonames ref - https://www.geonames.org/7193367/crystal-lake-mobile-home-park.html I think. --Tagishsimon (talk) 15:29, 10 January 2021 (UTC)
  • I changed it to the mobile home park, which is what the link is to. There is no water on Google maps or in any available satellite imagery to date, if they are going to build a lake we can always make a new entry. --RAN (talk) 05:07, 11 January 2021 (UTC)
Well. There is a lake on the ESRI image. RAN, changing an item from one thing (a lake) to another thing (a village) IS NOT HOW WIKIDATA WORKS. It is bad and damaging and wrong. Users are entitled to expect that the nature of items is consistent over time. Although it is unlikely anyone has a dependency on this specific item, as a general point, you do not know what external uses of the item have been made and are not in a position to know whether changing the item's nature will break anything. I have reverted your edit. Please do not do this ever again. --Tagishsimon (talk) 12:06, 11 January 2021 (UTC)
@Tagishsimon: This was an odd entry, which is why I brought it up here. It was originally linked to a en-wiki page which described the mobile home park but was removed (possibly because it's a commercial venture). That page was removed and left the intent of the entry unclear. Can you please explain what type of breakage you're referring to? Regards, Gettinwikiwidit (talk) 12:37, 11 January 2021 (UTC)
I take it you're saying it's better to create a new entry and possibly delete the old in this case? I can understand this, but I want to make sure I understand your point. Regards, Gettinwikiwidit (talk) 12:40, 11 January 2021 (UTC)

Wikiversity article representation in Wikidata and P31 values?

I've created a Wikidata object for a Wikiversity learning resource neurodiversity movement (Q104773840) "The Neurodiversity Movement". I considered adding an instance of Wikimedia article page (Q15138389) to that Wikidata item but seeing P31 has a none constraint for Q15138389 I'm puzzled at how to go forward with this. Even if I added P31 = "Wikiversity Learning Resource"(if I made a new Wikidata item for that) would that also be a none contraint for P31? How should I tag learning resource for P31 on Wikiversity articles so they get their Wikidata objects? Can I skip adding an instance of (P31) or subclass of (P279) altogether for Wikiversity articles and just tag them as good as I can? LotsofTheories (talk) 23:14, 10 January 2021 (UTC)

Property proposals: should they close?

Hello. In Wikidata:Property proposal/Generic there are more than 75 proposals. The discussion of some of the them started at 2019 and the last comment was 6 or 7 or 12 months ago. Why these proposals are still "under discussion"? Data Gamer play 06:52, 10 January 2021 (UTC)

  • The question is if we have a better solution to address the modeling question or not. Delay as such isn't an issue if the users are still interested. Proposals sometimes just lack sufficient interest to be implemented.
Obviously, there are a few proposals that are seen as incompatible with Wikidata's modeling or are insufficiently specified to be implemented. --- Jura 16:21, 11 January 2021 (UTC)

Wikidata weekly summary #450

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
    • QueryGraph is a web tool for creating queries in SPARQL from graphs.
  • Development
    • Continued working on the Query Builder. We focused on making it possible to query for Item values and being able to limit the number of Items in the result. We also looked into how to query for quantity values.
    • Changed P920’s data type from string to external identifier (phab:T269205)
    • Finished working on the problem of Items that have the page prop wikibase:statements set to 0 but actually do have statements (phab:T145712)
    • Continued work on the “Publishing” step of the Wikibase release pipeline by building and testing three components of the software suite: Wikibase (phab: T267893), Query Service frontend (phab:T268025), and Query Service backend (phab:T268022)

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.


Could someone do this?

I don't know if it's the right place to ask for this, but I'd like to ask for someone to add the link to an article in Silesian Wikipedia (code: szl) titled "Zwjyrzynta" to a semi-protected page Q729. I can't do this myself. --Psiŏczek (talk) 17:46, 11 January 2021 (UTC)

publisher

What is the structured data problem with the P123(property:publisher) Q88359373(Nels)? This is simply a postcard published by 'Nels'.Smiley.toerist (talk) 23:58, 11 January 2021 (UTC)

I think it's a constraint added by Uzume recently, restricting usage to Wikibase item (Q29934200). Ghouston (talk) 00:35, 12 January 2021 (UTC)
@Ghouston, Smiley.toerist: Is this Special:Diff/1339340150/prev what you are looking for? That should fix the SDC issue with c:Special:EntityPage/M20856711. Previously there was no allowed-entity-types constraint (Q52004125), so I added one but I did not realize it was being used on Wikibase MediaInfo (Q59712033) at SDC. It was an easy enough fix. —Uzume (talk) 04:16, 12 January 2021 (UTC)
I have been adding a publisher to the SDC in postcard images in the Commons for several months. This is certainly more than one case. Everytime I encounter a postcard Commons file I add the Q192425 property. In the future bots will probably automaticaly add Q192425 and other properties (a recent one: Wikidata:Property proposal/date posted), to file s in postcard Commons categories.Smiley.toerist (talk) 11:22, 12 January 2021 (UTC)
@Smiley.toerist: I do not really understand what you are saying. postcard (Q192425) is a Wikibase item (Q29934200) not a Wikibase property (Q29934218) (both are Wikibase entities). Perhaps you mean you add statements to Wikibase MediaInfo (Q59712033) entities (SDC: structured data at Commons) claiming instance of (P31) postcard (Q192425)? You are welcome to add such statements along with others claiming publisher (P123) as well. The reason instance of (P31) did not complain when you added it to a Wikibase MediaInfo (Q59712033) is because it already had a constraint allowing it to be added to Wikibase MediaInfo (Q59712033) (look at the qualifiers on the property constraint (P2302) statements specifying allowed-entity-types constraint (Q52004125) at Property:P31#P2302). —Uzume (talk) 18:17, 12 January 2021 (UTC)
Question: File:Kasteel de “Mick”.jpg. The castle is located in 'Brasschaat'. I get an error message if I try to use the location property. Should I always set the location in the 'depicts' property? The scanned postcard itself can be located anywhere, but this is not really relevant.Smiley.toerist (talk) 12:08, 12 January 2021 (UTC)
@Smiley.toerist: Properties like location (P276) have constraints. One of the possible constraints is allowed-entity-types constraint (Q52004125). For example one can look for this at Property:P276#P2302 (you might have to scroll to find it as there are often many constraints). You can find help on that constraint at Special:MyLanguage/Help:Property constraints portal/Allowed entity types. From the help and your issue you can surmise that you want to add an item of property constraint (P2305) qualifier to the allowed-entity-types constraint (Q52004125) constraint statement with the value of Wikibase MediaInfo (Q59712033). That is what I did for publisher (P123) above and the same can be done for location (P276) removing the error when added to SDC MediaInfo items like c:Special:EntityPage/M30301411 (I refer to it by its SDC Wikibase entity name and not its Commons File name c:File:Kasteel de “Mick”.jpg since we are talking about the associated Wikibase JSON structured data and not the File wikitext description or the File contents, in this case a JPEG image). —Uzume (talk) 18:17, 12 January 2021 (UTC)
@Smiley.toerist: I added Special:Diff/1339687796/prev the constraints qualifier allowing location (P276) to be used on SDC Wikibase MediaInfo (Q59712033) rectifying your issue with c:Special:EntityPage/M30301411. —Uzume (talk) 18:25, 12 January 2021 (UTC)

Verses of the Bible

Currently, thanks to the work of LesserJerome, we have elements for all Bible verses from the gospels, like Luke 3:14 (Q64688743). Also, due to Ben Skála, parts of the verses of Genesis, Job, and Lamentations exist. However, the vast majority of Bible verses don't have an item yet. Anybody who could create them? 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 13:37, 4 January 2021 (UTC)

It's my understanding that Catholic and Protestant bibles divide the text into verses a little differently, and consequently, number the verses differently. How do you account for this? Jc3s5h (talk) 14:50, 4 January 2021 (UTC)
A possibility would be to use aliases and include the text of the verse in a property for identification. 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 09:43, 11 January 2021 (UTC)
It seems to me like it would be good if the items of the verses would point to the corresponding text on Wikisource. ChristianKl ❪✉❫ 14:54, 4 January 2021 (UTC)
It is not as simple, as that one verse in one bible corresponds exactly with another verse in another bible. One translation can make a split mid sentence, while another let a whole sentence stay inside only one verse. Its the editors privilege to choose that. To make any sense of this, we should have one verse-item for every edition of the bible. 62 etc (talk) 07:11, 13 January 2021 (UTC)
Yes, that would be helpful and I thought about this, too. Is there any property for this? 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 14:56, 4 January 2021 (UTC)
Sitelinks aren't modeled via properties within Wikidata but on the right side in the sitelink menu that goes for Wikisource pages the same way it goes for Wikipedia pages. ChristianKl ❪✉❫ 15:06, 4 January 2021 (UTC)
@ChristianKl: Er, but we are talking about sections, right? Plus, there are multiple translations of the Bible on Wikisource, so there is no one-to-one correspondence anyway. 𝟙𝟤𝟯𝟺𝐪𝑤𝒆𝓇𝟷𝟮𝟥𝟜𝓺𝔴𝕖𝖗𝟰 (𝗍𝗮𝘭𝙠) 19:26, 4 January 2021 (UTC)
Unfortunately, I don't know enough about WikiSource to know what would be best here. ChristianKl ❪✉❫ 15:59, 5 January 2021 (UTC)
@1234qwer1234qwer4, ChristianKl: On en.wikisource, the versicles have ids, so you can do Genesis 1:1 (Q4101743)full work available at URL (P953)https://en.wikisource.org/wiki/Bible_(King_James)/Genesis#1:1part of (P361)King James Version (Q623398) --Tinker Bell ★ ♥ 21:00, 11 January 2021 (UTC)

References displayed in one vertical line

Hi, the displaying of the references and of some qualifiers is sometimes really boring, more details there, although I'm not sure this is the best place there. See also this screenshot. Christian Ferrer (talk) 11:52, 13 January 2021 (UTC)

Should probably go to WD:DEV. --Matěj Suchánek (talk) 14:13, 13 January 2021 (UTC)
Ok it's done, thanks you. Christian Ferrer (talk) 14:41, 13 January 2021 (UTC)

Indicating favorites

Is there any way to indicate a person's favorite color, lucky number, or similar? {{u|Sdkb}} talk 11:53, 11 January 2021 (UTC)

I hope not. --Tagishsimon (talk) 14:28, 13 January 2021 (UTC)
@Sdkb: You can state someone's favourite sports team using supported sports team (P6758) Piecesofuk (talk) 18:39, 13 January 2021 (UTC)
If someone actually use their lucky/favorite/important thing in daily life, there is Moka Akashiya (Q3860183)uses (P2283)Rosario Cross (Q63184654) and Thor (Q717588)owner of (P1830)Mjolnir (Q1401384). --Lockal (talk) 19:58, 13 January 2021 (UTC)

subclass of (P279) for city districts of Tallinn (Q2864118)

  1. Territory of Estonia Administrative Division Act, Chapter 1, § 2. Administrative division of territory of Estonia is directly stated that Rural municipality districts or city districts (linnaosi) may be formed in a rural municipality or city, respectively, pursuant to the procedure provided by law;
  2. Local governments in Estonia work about administrative reform in Estonia (Q16410284) by Ministry of Finance (Q12373345) says that Local governments may form rural municipality and city districts in their territories for the decentralisation of power;
  3. Tallinn (Q1770) is divided into 8 areas with certain boundaries (city districts of Tallinn (Q2864118));
  4. These areas have their own statutes and local governments which form a hierarchy with Tallinn goverment. For example, we have clear information from the statute of Kesklinn (Q1230929) that it has own local two-branch government (Linnaosa haldusorganid: 1. Tallinna Kesklinna linnaosakogu; 2. Tallinna Kesklinna Valitsus). Also § 2 (3) says that Linnaosa täidab oma haldusterritooriumil lisaks käesolevas põhimääruses sätestatule teisi riigi ja Tallinna õigusaktides talle pandud ülesandeid (In addition to the provisions of these statutes, a district shall perform other functions assigned to it by state and Tallinn legislation in its administrative territory).

With the above in mind, what should be specified in subclass of (P279) of city districts of Tallinn (Q2864118) and can we use items like Kesklinn (Q1230929) with located in the administrative territorial entity (P131)? Сидик из ПТУ (talk) 17:40, 11 January 2021 (UTC)

city district (Q4286337) is the right subclass for either city districts of Tallinn (Q2864118) or an intermediate "city districts of Estonia" and, yes, I see no reason why located in the administrative territorial entity (P131) would not work in the way you describe. There is perhaps some confusion here because "city district" sometimes describes one of multiple suburbs within one city, and at other times describes a structure that has the same territory as a larger city, but takes at the administrative role of a rural district (or "county"), i. e. one level ''up'' from the city government.Wikidata:Administrative_territorial_entity might also be helpful --Matthias Winkelmann (talk) 01:36, 12 January 2021 (UTC)

I agree with you. The problem is that such edits are rolled back by anonymous author for several years. Сидик из ПТУ (talk) 07:40, 12 January 2021 (UTC)
  • It might be easier to use a new item for "district" in Estonia (per specific organizational meaning for (the quoted) "Rural municipality districts or city districts (linnaosi)")
    and add this where applicable. Possibly in addition to the more vague "district of Tallinn".
    As the English label of Q2864118 is "city districts of .." (note the plural), it might not even be a class, but a list or some other item about all districts. --- Jura 13:18, 12 January 2021 (UTC)
city district (Q34985575)? Now Kesklinn (Q1230929) has city district (Q34985575) at instance of (P31). Сидик из ПТУ (talk) 17:33, 12 January 2021 (UTC)
Not sure: there seems to be some back-and-forth around Q34985575. Currently it's labeled "city district", which a "rural municipality district" (from the explanation above) would probably not be, nor is it (per item's current P31 value description) a "human-geographic territorial entity". So P31 and (English) label would need some changes. --- Jura 11:30, 13 January 2021 (UTC)
So, if we want to state that Q34985575 is a "city district" from "Rural municipality districts or city districts" then it should be city district (Q4286337) in instance of (P31). Сидик из ПТУ (talk) 16:52, 13 January 2021 (UTC)

Link hover text feature request

I can't find anywhere else where this has been suggested so I'll ask it here.

Does anyone else think that the hover text of any Wikidata Property, Item, Qualifier, link etc. should show the item description instead of its identifier?

As a new contributor, I often find myself having to click on property, qualifier, or item links for elements that I don't know, or even ones that I know, like instance of (P31) so that I can differentiate it between other extremely-similar properties like subclass of (P279).

This would of course make editing faster for anyone, since they wouldn't have to navigate between the page they want to edit, and the elements in statements that they do not know of.

Here is the current hover text of instance of (P31). I would like it to show that class of which this subject is a particular example and member instead of Property:P31.

Lectrician1 (talk) 14:57, 13 January 2021 (UTC)

Preferences > Gadgets > Descriptions. (MediaWiki:Gadget-Descriptions.js) Ayack (talk) 17:09, 13 January 2021 (UTC)

Fix typo request

I noticed that on Dagbani (Q32238), in the column language, it is wrote “dagbanli” instead of “Dagbanli”. It is possible to change? Thanks in advance!!! --2001:B07:6442:8903:1DA4:7C70:DD0B:2897 16:12, 13 January 2021 (UTC)

The French label was dagbanli so I capitalised the D for you MSGJ (talk) 16:32, 13 January 2021 (UTC)
Why did you that? I suggest to undo the change. Languages are not capitalized in French. --Dipsacus fullonum (talk) 17:23, 13 January 2021 (UTC)
Reverted. --Tagishsimon (talk) 17:25, 13 January 2021 (UTC)

Fix typo request

As reported in https://ibb.co/XSsJdG0 I noticed that on Dagbani (Q32238), in the column language, it is wrote “dagbanli” instead of “Dagbanli”. It is possible to change? Thanks in advance!!! --2001:B07:6442:8903:1DA4:7C70:DD0B:2897 16:12, 13 January 2021 (UTC)

Atoka Reservoir (Q19864939) the same as Atoka Reservoir (Q6474858) but have different e-wiki entries

It seems having distinct wiki entries produces a conflict on merge. Is the idea to first merge the two entries in en-wiki? Does anyone have any experience with that? Regards, Gettinwikiwidit (talk) 12:19, 12 January 2021 (UTC)

They should be merged at enwiki. Lymantria (talk) 12:27, 12 January 2021 (UTC)
@Gettinwikiwidit: Specifically, pick one of them to keep, copy any missing content from the other one, and replace the content of the other one with a #REDIRECT link - see en:Wikipedia:Redirect for details. ArthurPSmith (talk) 18:14, 12 January 2021 (UTC)
@ArthurPSmith: Thanks for the pointers. This is now done. Gettinwikiwidit (talk) 23:28, 13 January 2021 (UTC)

Property for redirects after deprecated rank of official website (P856)

On this item Google Photos (Q20007086) the official website (P856)(URL) links to photos.google.com but when I visit that website it redirects to www.google.com/photos/about. I'd like to add a deprecated rank to photos.google.com and add the reason that it redirects to another URL. Is there a property for redirects that websites do? LotsofTheories (talk) 03:00, 9 January 2021 (UTC)

@LotsofTheories: deprecating it is actually the wrong thing to do. assuming the URL was correct at some point instead add an end time (P582). see Help:Deprecation for more details. BrokenSegue (talk) 05:15, 9 January 2021 (UTC)
The URL stays the same when signed in, otherwise it redirects. Also redirects are not always deprecated, the purpose of a redirect can be to provide a more stable or easier to remember URL. Peter James (talk) 15:47, 9 January 2021 (UTC)

Comment – Leave it as is. That longer URL is variable based on cookies and other (unknown) factors. photos.google.com is the permanent URL. Senator2029 13:40, 14 January 2021 (UTC)

aircraft hijacking (Q898712) and its operator (P137)

Hello, for a aircraft hijacking (Q898712), what should be operator (P137)? The airline that operate(s|d) the flight as I would have first thought or the hijackers that have "performed" the hijacking (they performed the "service" of hijacking...) ? Eg LOT Polish Airlines Flight 165 hijacking (Q1434061) / Air France Flight 8969 (Q1574009) Bouzinac 💬●✒️●💛 12:45, 13 January 2021 (UTC)

I would use perpetrator (P8031) to describe the hijackers, and operator (P137) to describe the airline. --Cavernia (talk) 08:20, 14 January 2021 (UTC)

New dashboard for lexicographical data statistics

Hello all,

Since the middle of December, we have a new dashboard for statistics about lexicographical data on Wikidata: Wikidata Datamodel Lexemes. Similar to other dashboards (e.g. Wikidata Datamodel Statements), it collects various statistics (in this case, using a collection of SPARQL queries). The data is similar to some of the results gathered at Wikidata:Lexicographical data/Statistics, and shows the evolution of data over time.

This dashboard highlights the efforts of the editors to improve the content by adding new Lexemes, Senses and Forms (for example the recent import of Lexemes in Estonian), but also areas where improvements could be made (83% of Lexemes without Senses).

We hope you will find this useful. If you have any feedback or notice an issue, feel free to contact us on this page. Thanks, Lea Lacroix (WMDE) (talk) 12:42, 12 January 2021 (UTC)

In the telegram group we have discussed the need to reference all forms present in the lexeme namespace. Could you include a measure to capture that (forms with "stated in" or similar property)? Also having a sense with no P5137 is not useful at all, so could we have a measure for that too (senses with P5137)?--So9q (talk) 10:51, 14 January 2021 (UTC)
see also Wikidata talk:Lexicographical data#New dashboard for lexicographical data statistic. --- Jura 16:30, 14 January 2021 (UTC)

Help merging

Will anybody merge Q7421267 and Q32430219? I cannot solve "Error: Conflicting descriptions for language fa.". Thanks. --95.127.172.126 18:33, 14 January 2021 (UTC)

✓ Done. iXavier (talk) 18:35, 14 January 2021 (UTC)

Quickstatements not accepting P123

Hi, I have the following quickstament:

CREATE

LAST Len "5 Steps For Improving Old Film Footage - Comparison (Speed Correction, Upscaling, Colorization)"

LAST Den "youtube video from The Great War channel"

LAST P31 Q63412991

LAST P1476 en:"5 Steps For Improving Old Film Footage - Comparison (Speed Correction, Upscaling, Colorization)"

LAST P953 "https://www.youtube.com/watch?v=It_YO7kJL78"

LAST P407 Q1860

LAST P123 Q30598122

LAST P571 +2020-01-02T00:00:00Z/11

All good and right. However, when I go to import the command, P123 is missing. Is there any reason why is that? Tetizeraz (talk) 09:00, 18 January 2021 (UTC)

Works for me (I can see it after import). Check that you have tabs where they should be and don't have them where they shouldn't be. You can use "|" character instead of tabs. --Matěj Suchánek (talk) 10:51, 18 January 2021 (UTC)
That line has multiple spaces before and after "P123", there should be tabs the same as the other lines. Peter James (talk) 12:20, 18 January 2021 (UTC)
Matěj Suchánek and Peter James: I added \t instead of multiple spaces, since I thought QuickStatements allowed four spaces instead of a tab. No big problem. Thanks! Tetizeraz (talk) 12:34, 18 January 2021 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Matěj Suchánek (talk) 10:26, 20 January 2021 (UTC)

I now have a global user page, can my user page be deleted?

I now have a global user page, can my user page be deleted? LotsofTheories (talk) 08:57, 20 January 2021 (UTC)

Done. Hasley+ 18:21, 20 January 2021 (UTC)
This section was archived on a request by: Hasley+ 18:24, 20 January 2021 (UTC)

Songs in an album

What is the best way to express which songs are included in an album?
Example: The album His Hand in Mine (Q1755467) contains the songs In My Father's House Are Many Mansions (Q16996744), Milky White Way (Q6858155), Joshua Fit the Battle of Jericho (Q749116), Swing Down Sweet Chariot (Q7658557) and Mansion Over the Hilltop (Q6751756). I could then go to the album item and add the songs as has part(s) (P527), or I could go to the song items and use part of (P361). Or maybe both? Is there a better property to use? A possibility is published in (P1433) which is used to identify songs published in a songbook. Is a music album similar to a songbook? --Cavernia (talk) 08:12, 14 January 2021 (UTC)

I believe it is tracklist (P658). Jean-Fred (talk) 16:36, 14 January 2021 (UTC)
"tracklist" is correct. I updated the album with the twelve tracks in question. Moebeus (talk) 15:27, 15 January 2021 (UTC)

Sorry for my ignorance

What is "LTA spam", I see it in deletion rationales, but can't find a definition. --RAN (talk) 03:58, 15 January 2021 (UTC)

"Long term abuse", I assume? AleatoryPonderings (talk) 04:06, 15 January 2021 (UTC)

Can we keep at least one example of an exploited entry

For example we have Q104755571 which has over 22 valid Identifiers, but they almost all have no data at the host site. I think we need to keep an example to show which Identifiers can be exploited for SEO. It can be used at a Wikidata project page to discuss which websites we point to can be exploited. For instance you can get an IMDB entry without having an credits or even being an actor. I think more editors need to be aware of what websites that can be exploited, or in this case are already exploited with empty entries. --RAN (talk) 22:31, 15 January 2021 (UTC)

I have archived it at Special:Permalink/1341920262.--GZWDer (talk) 23:20, 15 January 2021 (UTC)
Another example is Prem Raj Pushpakaran (Q61656939), though this item is notable as it is used in other items.--GZWDer (talk) 23:28, 15 January 2021 (UTC)
We generally use Wikidata property for an identifier that does not imply notability (Q62589320) on properties for identifiers where this is the case. ChristianKl ❪✉❫ 00:13, 16 January 2021 (UTC)

Merging given names

Should Tali (Q85439304), Tali (Q96652117), Tali (Q96653951) and Tali (Q96675834) be merged? They seem to be the same name, just from different languages. Or if not merged, could someone go fix the descriptions and labels, as they now are conflicting with each other? The same probably applies for Tali (Q96651126) and Tali (Q96655111), too. Kissa21782 (talk) 11:23, 9 January 2021 (UTC)

Is it possible to add the calorific value of the coal burnt in a power station?

The source of energy (P618) of Kangal power station (Q85773487) is lignite (Q156267). Is it possible to add that the average combustion enthalpy (P2117) or Energy value of coal (Q5377228) of that particular lignite (Q156267) is 1100 kCal/kg please? If so how?

Reference: The owners website

When I try to add an item of property constraint (P2305) of "kilocalorie per kilogram" to combustion enthalpy (P2117) I get an error: "No match was found".

Chidgk1 (talk) 17:29, 13 January 2021 (UTC)

Perhaps energy storage capacity (P4140) - if the constraints are changed to allow kCal/kg? But on an item for the lignite burnt in that power station rather than on the power station item, I think. --Tagishsimon (talk) 17:53, 13 January 2021 (UTC)
Thanks - I am not likely to work on any coal mine items soon but if I do so in future that may be useful.Chidgk1 (talk) 06:09, 14 January 2021 (UTC)
@Tagishsimon, Chidgk1: About allowed units in general: Per property we should allow only units of compatible dimensions. For instance, combustion enthalpy (P2117) currently allows both kJ/kg and kJ/mol. This is bad because it does not allow direct comparison of values: While they are related by, say, the molar mass, it makes consumption of the property difficult in general. Moving forward, we should pay attention to this issue when proposing new properties and before allowing new units in constraints. Ideally we would try to fix existing properties. Toni 001 (talk) 11:08, 16 January 2021 (UTC)

How to delete someone's twitter account

I have to say I do not understand WikiData, despite being (very sporadic) Wikipedia editor for >10 years :(

This person

Steven Frank (Q7614717)

has a wrong Twitter account there. The twitter account is deactivated. I removed {{Twitter}} from his wikipedia entry, but it is still here in wikidata.

I have no clue how to remove it. Clicking on the remove icon causes failure. Uhh, what to do exactly?

--Running (talk) 07:03, 15 January 2021 (UTC)

I guess I should "deprecate" it? How? What to do? --Running (talk) 07:05, 15 January 2021 (UTC)
ok.... I did something I guess. --Running (talk) 07:24, 15 January 2021 (UTC)
@Running: there is no need to deprecate valid statements about deactivated accounts. Setting qualifier end time (P582) is enough in this case (there is also has characteristic (P1552)deactivated account (Q56631052), but this is rare). --Lockal (talk) 15:00, 15 January 2021 (UTC)
@Lockal: Well, I don’t know when did he deactivate it. So I can’t set end time. --Running (talk) 04:44, 16 January 2021 (UTC)
you can set the time to "unknown". BrokenSegue (talk) 14:41, 16 January 2021 (UTC)

Find items of a wikipedia category

Is there a way to find the wikidata items (Q numbers) of articles, subcategories, templates that are included in a category in enwiki? Data Gamer play 13:21, 16 January 2021 (UTC)

yes, you might be interested in petscan which does this and more. BrokenSegue (talk) 14:14, 16 January 2021 (UTC)

Prohibiting batches for P5034

Hello. Is it possible to prohibit the batches or scripts that are running for National Library of Korea ID (P5034)? It will take years or decades till all the wrong IDs that were added get corrected by a human. Apparently, after the National Library of Korea ID (P5034) were then added, a bot runs and then also adds the wrong VIAF-ID based on the wrong Korean Library ID. --Christian140 (talk) 08:42, 8 January 2021 (UTC)

  • Did you try talking to the bot owner? ChristianKl ❪✉❫ 11:40, 8 January 2021 (UTC)
  • If there is a bot that adds incorrect data, please ask its operator to stop it and repair. If it's still running and you don't receive any response, contact a bureaucrat or an administrator to have it blocked.
If some or all data should be deleted, you might want to ask at Wikidata:Bot requests. --- Jura 11:41, 8 January 2021 (UTC)
I don't even know where those batches come from. I only always see them in the version history. --Christian140 (talk) 05:26, 9 January 2021 (UTC)
Just check the history for the account who added them. From statistics, it appears that some 26k values were added around November 27, 2020. --- Jura 07:59, 9 January 2021 (UTC)
Wow. Immediately 26k. Might have been better to test the batches before with a small number instead of going all in immediately. Wouldn't be surprised if more than 25k of these are wrong. If three people would check ten objects each day, it would take 2.4 years to have verified all. So, how can we prevent another disaster like this? --Christian140 (talk) 12:42, 9 January 2021 (UTC)
@Epìdosis: might know more about it [9]. --- Jura 12:53, 9 January 2021 (UTC)

If there are no objections, I would move this discussion to Property talk:P5034. The sequence of the facts is:

  1. November 2019: addition of about 570k VIAF ID (P214) through QuickStatements by @Bargioni: (see Property talk:P214#Recent synchronisation), in all cases in which VIAF linked to Wikidata
  2. November 2020: upload by me in collaboration with Bargioni of four Mix'n'match catalogs for NLK (3992, 3993, 3994, 3995)
  3. November/December 2020: Mix'n'match makes automatic matches on the basis of date of birth (P569)/date of death (P570) and on the basis of VIAF ID (P214) and ISNI (P213); different users, mainly me, have then synchronised Mix'n'match with Wikidata, importing these (about 40-50k) automatic matches from Mix'n'match to Wikidata through QuickStatements

It is plainly sure that in the above process a significant number of errors has been imported, in two phases, from VIAF (which is the main responsible) to Wikidata. I'm particularly suspicious about items in which

  1. the only external IDs are VIAF ID (P214)+National Library of Korea ID (P5034): these 1081 as of now
  2. the only external IDs are VIAF ID (P214)+National Library of Korea ID (P5034)+ISNI (P213): 93 as of now
  3. the only external IDs are VIAF ID (P214)+National Library of Korea ID (P5034)+another external-id different from ISNI (P213): 2446 as of now

I tend to think that the percentage of error outside these items should be very low. Regarding the above 3.5k items, I'm not sure a bot-removal of VIAF ID (P214) and National Library of Korea ID (P5034) would be the best solution, since it may affect the identifiability of the subject - manual check is probably a better solution. When you remove NLK, please be sure to remove the match also on Mix'n'match (enable User:Magnus Manske/mixnmatch gadget.js in your common.js), otherwise some user may reinsert the wrong values from Mix'n'match in good faith. I'm ready to discuss further ways to clean the errors. --Epìdosis 13:35, 9 January 2021 (UTC)

  • If some subsets of statements are found to be incorrect, these should be deleted or deprecated by bot. One shouldn't leave it for other contributors to fix manually. --- Jura 09:11, 11 January 2021 (UTC)
    @Jura1: I agree, but my statement about the incorrectness of the items resulting from the above queries is purely hypothetical, I'm not able to verify them and I would tend not to remove possibly correct data as a precaution; I could also think about VIAF doing a good job, until some systematic verification is done. If a Korean-speaking user verifies i.e. 20 items resulting from query 1 and affirms that VIAF ID (P214) and National Library of Korea ID (P5034) are both wrong in 15 cases or more, I've no problem in removing the properties from all the 1081 items; same for query 2 and query 3. --Epìdosis 14:25, 11 January 2021 (UTC) P.S. Removing all the VIAF ID (P214) and National Library of Korea ID (P5034) has the problem that, leaving the incorrect match between NLK entry on Mix'n'match and the Wikidata item from which the wrong NLK was removed, the NLK will probably be reinserted sooner or later from MnM and than VIAF will be added on its basis. Mix'n'match should have a system of batch-removing incorrect matches, which is at the moment absent. --Epìdosis 14:29, 11 January 2021 (UTC)
I don't get the argument why objects should be looked into where the only external IDs are VIAF, ISNI and Korean library number. I usually add those three to objects, so, also the first that comes when doing the SPARQL query was added by me. I would rather look into everything that was added through batches. I went to maybe the first 15 of these 1081 and Q12172968, Kang Ghil-boo (Q12582750), Q12582880, Q12583170. --Christian140 (talk) 16:48, 16 January 2021 (UTC)

Alexander Altmann

Hi, There is a big confusion being two painters: Q37846275 and Q546341. One is Ukrainian (1878-1932), the other one is Russian (1885-1950). Links to Wikipedia and Commons are jumbled, and even the Identifiers are wrong and mixed up. Help needed. Thanks, Yann (talk) 23:12, 12 January 2021 (UTC)

I fixed the links to Wikipedia and Commons. Regards, Yann (talk) 23:45, 12 January 2021 (UTC)
Moved from Administrators' noticeboard. ‐‐1997kB (talk) 05:58, 15 January 2021 (UTC)
@Yann:, they actually are the same, just different sources provide different dates. Also, Kiev Governorate was a part of Russian Empire until 1919, and Altmann left Russian Empire in 1905 (by all sources) and never returned back (which was very common for people of that time). --Lockal (talk) 16:05, 15 January 2021 (UTC)
@Lockal: The same? I could understand that there are uncertainities about the birth date, but how do you explain that one died in 1932 and the other in 1950? And that one is born in Odessa, and the other one in Kiev? [10] seems to support your claim, but we have a precise date and place of death in France: September 14th, 1932, Nemours (77). It should be easy to get the death certificate. Regards, Yann (talk) 21:11, 15 January 2021 (UTC)
@Yann: Yes, just dates are wrong. There are no contradictions about birth place (and no statements anywhere that he was born in Odessa). He was born in Sobolevka village, Kiev province (frwiki states that as "nee a Kiev", but that's just an approximation). There is a bigger article about him, which mention both 1932 and 1950 as death years. But the main evidence that that they are the same are signatures on paintings, on every source, on every auction, they are just the same no matter which period of life is mentioned. As for the question about primary source of "died in 1950" claim - I have no answers. 1932 has much stronger sources, 1950 just looks like a widely distibuted alternative value out of thin air. --Lockal (talk) 11:06, 16 January 2021 (UTC)
@Lockal: If you are sure about this, the 2 WD entries should be merged, as well as links to WP and Commons. Regards, Yann (talk) 17:40, 16 January 2021 (UTC)

Can I add the property "country of origin" in Structured commons data on files?

Like in this file, for example. Tetizeraz (talk) 21:58, 16 January 2021 (UTC)

Yes, that would be very welcome. --Tagishsimon (talk) 22:08, 16 January 2021 (UTC)
Tagishsimon I added country of origin to the linked file and there's a warning. Should I ignore it? Tetizeraz (talk) 22:19, 16 January 2021 (UTC)
IMO yes, ignore. There's no logic in constraining the use of that property to wikidata when commons items can usefully be described by it. Constraints are often wrong or not useful. --Tagishsimon (talk) 22:23, 16 January 2021 (UTC)
I've changed the constraint - diff. Warning gone. --Tagishsimon (talk) 22:25, 16 January 2021 (UTC)

Python PEPs in WD

https://www.python.org/dev/peps/ seems to be missing in WD. WDYT about importing them. How best to do it? Scrape with a bot?--So9q (talk) 07:32, 17 January 2021 (UTC)

Before importing a set of entries, remember there may be people don't support adding them. For example, we have the full set of Request for Comments (Q212971) due to User:RfcBot; However, somebody does not agree to import a set of encyclopedia articles. Another example is since 2018 I have an idea to import 14000+ entries from Astronomer's Telegram (Q18205341) although I have not yet started to discuss the idea seriously.--GZWDer (talk) 12:51, 17 January 2021 (UTC)

Why are we importing all science articles we can possibly find?

So linked to the discussion about scope of wikidata in my comment in the discussion above, what is your arguments? I have two: 1) we could need them sooner or later in Wikipedia as reference and then it's easier if they are already in Wikidata (but this accepts a truckload of bloat items that are perhaps never linked from any other WMF project which is possibly a strain on the infrastructure and/or volunteers) 2) science is the basis of knowledge in our culture, so to describe the sum of all knowledge (goal of Wikipedia, no?) needs science and references to it. Importing an article in Wikidata is only the beginning. WDYT?--So9q (talk) 20:43, 15 January 2021 (UTC)

I noticed the scope discussion. Very very odd. Scope is already established in Wikidata:Notability. Ideas such as "Our primary goal is to serve as a data store for Wikipedia content" is for the birds and shows a very profound lack of understanding/appreciation of the non-wikipedia uses of wikidata. Given WD:N, we do not need to make an argument for the inclusion of scientific papers. --Tagishsimon (talk) 20:50, 15 January 2021 (UTC)
This is the proposal of WikiCite - it powers a lot of things including [11]. This also provides information of many researchers. Adding new scientific articles will have building a database of open citations and linked bibliographic data (e.g. eventually you may easily find a PDF copy from a given PubMed ID, if it is in free license). However whether to include non-scientific articles, even those with DOI, is disputed; welcome to participate in Wikidata_talk:WikiProject_Source_MetaData#Encyclopedia_articles_and_notability.--GZWDer (talk) 23:25, 15 January 2021 (UTC)
See also Special:Diff/1342067604.--GZWDer (talk) 05:02, 16 January 2021 (UTC)
Thanks for your comment and for taking my question seriously. We have not met here in years. I recognize the tone of your comment from our previous interactions and I don't recognize it as friendly, correct? I have not heard https://en.wiktionary.org/wiki/for_the_birds before, but then again I'm not so well versed in English culture insults. We are unfortunately missing a lexeme describing the expression it seems, would you be willing to create it now that you seem to have a firm grasp of the concept? You are very welcome to import other insults and English idioms to the lexeme namespace e.g. from Wiktionary to help out (if you know Javascript you can adapt this botcode).
Anyway, you did not clarify why all the millions of scientific papers are in scope, I'm guessing you mean they fill a structural need WD:N 3)? I'm guessing 3) because 1) mentions sitelinks and most science papers are missing that and 2) says "It refers to an instance of a clearly identifiable conceptual or material entity. The entity must be notable, in the sense that it can be described using serious and publicly available references." which is pretty broad and vague (which of course could very much be the reason why I and others have a little trouble understanding what we agreed should be included and what should be excluded).
I don't think every single science paper and book ever written is notable, do you? If the community agree with you I suggest we clarify that in WD:N and clean up the science objects immediately.
I would prefer if we could link to other places where all the science papers we need are, e.g. I don't see why Elsevier and other big journal corporations could not provide a SPARQL endpoint we can federate with, have you or anyone else asked them?
We could even have a selected import of only science papers that are currently used in WP (unfortunately no Wikibase or similar source of truth linked data hub with the worlds science papers exist yet and WD is far from complete and full of errors because of low quality imports by bots if I understood correctly).
This is stated in WikiCite: "The idea is curate the collection of citations in Wikidata, or perhaps in local instances of the Wikibase platform at particular institutions. To the extent that Wikidata has capacity and data is compatible, then information is in Wikidata."
So it's not clear at all that Wikidata is the place for all the articles and books in the world. I suggest we contemplate how to best move forward. What we have now is an incomplete mess (for example when I search Swedish science articles in Wikidata:Scholia I find nothing, no authors, no articles and we cannot import from the Swedish universities because they don't publish metadata as CC0 (I asked them to release it but they answered that they mix in proprietary Elsevier metadata into their stream without marking it up so they can't release it, also the Swedish Royal Library are collecting (in SwePub) but not publishing this data under a suitable license for the same reason apparently). So maybe we should create a Wikibase to hold all the metadata that has not yet been released properly? That metadata database could be populated by many sources and not have the same license related problems as Wikidata.
According to https://www.wikidata.org/wiki/Special:MostLinkedPages we now have 23 mio. science articles out of a total of around 95 mio. on SciHub (that's only 24%). That is only the digitized ones, there are many older articles missing in SciHub.--So9q (talk) 11:33, 16 January 2021 (UTC)
Deletionism is bad for Wikipedia and part of the reason why Wikipedia lost editors about a long time. The idea that we should get more deletionist because someone doesn't like incomplete data seems to be problematic. While it's good when people speak in a friendly tone, you shouldn't surprised when an very unfriendly proposal to delete data in which a lot of users spent a lot of effort get strong pushback.
If there's a mix with proprietary Elsevier metadata, then you also can't publish simply publish the data in an Elsevier database. To have the data in the open you actually need to have a state of affairs where it becomes valuable for researchers to release data about their papers in CC0. Tools like Scholia help provide value. There will be a time where we have enough papers to calculate our own metric that works like the impact factor.
As far as our notability rules are concerned scientific articles are conceptual entities that are described in multiple reliable sources and thus notable. ChristianKl ❪✉❫ 13:16, 16 January 2021 (UTC)
In Wikidata_talk:WikiProject_Source_MetaData#Encyclopedia_articles_and_notability, it is concerned whether encyclopedia articles are notable. An encyclopedia article may be cited by other sources. Previously, there was a similar discussion about news articles in Wikidata:Project_chat/Archive/2020/03#Notability_of_news_articles. GZWDer (talk) 17:04, 16 January 2021 (UTC)
Just because every Wikipedia article can be cited, doesn't mean that we create Wikidata items for every article (items about the article and not the concept the article refers to). We only have those items when we actually model citations. ChristianKl ❪✉❫ 13:29, 17 January 2021 (UTC)

I think the above reaction of User:So9q to "profound lack of understanding/appreciation of the non-wikipedia uses of wikidata" shows there is a need to provide better information about these uses. Compsci students may have no problems immediately associating exciting AI research and its vast broadness of uses but this is apparently not everyone's bread and butter. --SCIdude (talk) 08:30, 17 January 2021 (UTC)

  • GLAM people don't seem to have any problem seeing uses for Wikidata outside of Wikipedia without being compsci students. WikiProject sum of all paintings is one of the biggest on Wikidata. Every weekly announcement contains a list of SPARQL queries and many of them are non-wikipedia uses. ChristianKl ❪✉❫ 12:57, 17 January 2021 (UTC)
    Librarians were the first information scientists, anyway. --SCIdude (talk) 17:49, 17 January 2021 (UTC)

Wikiversity Quiz

I've seen a lot of even fairly experienced people here get confused by the subtleties of wikidata's model. We have help pages but I thought it would be good if we had a quiz that people could take to ensure they've understood everything. I threw together a sample quiz on wikiversity about statement rank. I imagine us writing a collection of quizzes and linking them into our help documents.

I'm interested in feedback on:

  • if this is a good direction
  • if anyone found it helpful
  • if I made any errors
  • if anyone is interested in helping

Feel free to directly edit the quiz too. BrokenSegue (talk) 18:00, 16 January 2021 (UTC)

This does not show qualifiers, which in many cases are important.--GZWDer (talk) 23:01, 16 January 2021 (UTC)
@GZWDer: yeah the intention would be to make a separate quiz about qualifiers but maybe a unified one would make sense. BrokenSegue (talk) 17:54, 17 January 2021 (UTC)

Is it possible to capture bulk metadata from URLs using Zotero?

I found a discussion in the Zotero Forums and came across this code:

var path = '/home/username/Desktop/urls.txt';

var urls = Zotero.File.getContents(path).split('\n').map(url => url);

await Zotero.HTTP.processDocuments(

urls,

async function (doc) {

var translate = new Zotero.Translate.Web();

translate.setDocument(doc);

var translators = await translate.getTranslators();

if (translators.length) {

translate.setTranslator(translators[0]);

try {

await translate.translate();

return;

}

catch (e) {}

}

await ZoteroPane.addItemFromDocument(doc);

}

)

The idea is, you save the URLs you gathered in the urls.txt and Zotero parses it, adding to the library of citations you have. However, I can't choose the folder I'm saving, and when I export the saved citation to QuickStatements, I get generic metadata, like the link is a webpage, not a video recording (i'm trying to pull metadata from youtube videos, so it should be at least video recording). I'm wondering if anyone has a solution to this problem! Thanks for any help! Tetizeraz (talk) 02:28, 17 January 2021 (UTC)

Yeah, ChristianKl, I'm trying to use differnt alternatives, like using Jupyter notebook and OpenRefine together. Still researching though. Tetizeraz (talk) 15:08, 17 January 2021 (UTC)

Inflated bogus edit statistics in WD

Before we fall over ourselves in gratulation about WD contributing with 40% of all WMF edits last year I would like you to take a look at Special:MostRevisions and explain to me why these science article objects have so many edits?

Could it be that pywikibot was used to import it and it was done in a way that did not upload it once after populating the item with authors (similar to what I do in Wikidata:LexUse using wikibaseintegrator) but instead each author added resulted in one edit? If yes, what do we do about that? How does that affect our interpretation of the statistics? WD stats seem bogus to me. WDYT?--So9q (talk) 07:47, 17 January 2021 (UTC)

I wouldn't get excited about edit counts, although BayGenomics: a resource of insertional mutations in mouse embryonic stem cells (Q39790431) having 8326 versions is surprising. It doesn't seem to be authors in that case, but cell lines, added one by one with QuickStatements. Even viewing the item is difficult, but the history is accessable. Most of the others are probably articles with hundreds of authors, and the edits may just be conversions of author name string (P2093) values to author (P50). Ghouston (talk) 08:37, 17 January 2021 (UTC)
WD edit statistics are indeed nothing to be proud of high edits are often more of a sign of inefficient edits then positive signs. ChristianKl ❪✉❫ 12:51, 17 January 2021 (UTC)
WD edits are what they are. Some edits add lots of triples, some only a small handful. The premise of So9q's question is erroneous (we're not falling over ourselves) uninformed (user's scope their edits according to what they or their bots are able to do at the time; there's not a correct way to do things) and tedious. Exactly the same dull argument could be made about language wiki counts: "Could it be that it took 7 edits to get that little bit right, and what about all the semi-automated tiny edits'. Don't get me wrong: knock yourself out with this sort of ill-informed and ill-considered analysis all you want. Only, don't expect to avoid the distain arising. --Tagishsimon (talk) 13:57, 17 January 2021 (UTC)
@Tagishsimon: I don't think he's completely off, in the past we had times where people saw mile stones of edits like having 1 billion edits as something positive. The claim that all users scope their edits by what's possible and that there are no users who make inefficient edits seems wrong to me. The amount of edits that Wikidata infrastructure can do is limited and there were times in the last years where we had problems with too many edits being made. As Wikidata grows there's a good chance that we will again run into situations where we get problems because too many edits are made at a given time. ChristianKl ❪✉❫ 18:47, 17 January 2021 (UTC)

value type constaint in The Outbreak of WWI - How Europe Spiraled Into the GREAT WAR - Week 1 (Q104854062)

In this QID, P123 has the value Q30598122, but for some reason I get the following value type constraint:

value type constraintH

Values of publisher statements should be instances or subclasses of one of the following classes (or of one of their subclasses), but The Great War currently isn't:

organization

human

publisher

software developer

newspaper

periodical

website

samizdat

publisher

brand

Is it a problem in P123 or Youtube channel isn't a valid value for P123? Tetizeraz (talk) 15:07, 17 January 2021 (UTC)

Listeriabot: request to implement

I'm posting here to ask if someone here is able to show me how to set up Listeriabot to run on a new wiki (in this case, Wikiversity; see notice). The original bot creator is typically very busy, and I don't have any bot-running expertise. Is anyone able to show me how to set it up an activate it. I'm not certain whether it's just a case of adding an extra line of code somewhere in the the github repo, or if there's more to it than that. Thanks in advance for any assistance - I'm keen to get it up and running on that site asap. T.Shafee(evo&evo) (talk) 06:36, 14 January 2021 (UTC)

Cleanup after User:GZWDers bots

Now that all bots by this user has now been blocked we have a cleanup effort to do. How do we best organize it? Create a new WikiProject Cleanup? See https://www.wikidata.org/wiki/User_talk:GZWDer#Items_with_mismatched_ORCID_and_name I'm generally in favor of deletion and reimport rather than trying to correct millions(?) of bad items/statements by hand or semi-automatic edits. WDYT?--So9q (talk) 06:45, 14 January 2021 (UTC)

  • @So9q, GZWDer: I don't think this is the same bot account that was blocked; the problem in any case is that it is too trusting of EuropePMC for ORCID id's; they occasionally mess up those assignments. The vast majority however are correct so deleting and reimporting seems unwarranted to me. If we had some way to detect the problem cases that would certainly be useful. ArthurPSmith (talk) 16:45, 14 January 2021 (UTC)
    • As a very general note, many data with varying quality are imported by different users, usually without discussion. We need to discuss good practice of data import in general.--GZWDer (talk) 17:33, 14 January 2021 (UTC)
@GZWDer: Radical idea: staging area?? It could be as simple as a "staging" property that people could FILTER out of their queries knowing that the quality is not assured. Gettinwikiwidit (talk) 02:04, 15 January 2021 (UTC)
This is similar to what is proposed in User:ChristianKl/Draft:New_Ranks#Uncertain. However every imports (by others) may have some issue (while the error rate is low, the total number may still be considerable), and it may not be easy to find out. However, a new rank does not address the issue of possible conflated entities, nor issues in labels, nor duplicated entities (this is the major issue to fix).--GZWDer (talk) 03:00, 15 January 2021 (UTC)
Also, while clean up is ongoing, we still need a guide about what are acceptable practices on new imports (i.e. even you can flag some data as need checking, how should we assure the issues are fixed eventually instead of making backlog growing without end). @Multichill: for notice.--GZWDer (talk) 03:03, 15 January 2021 (UTC)
@GZWDer: The data would presumably be "unstaged" at some point when it's been verified. Only things still tagged would be in the backlog. Maybe it would help to start the conversation with policy goals and technical goals. Should we require some sort of review on imports? Should we have some sort of easy rollback for imports? Should we have some automated reports tracking imports? Are we trying to avoid people using "suspicious" data? What balance are we looking between encouraging contributions and discouraging bad contributions? It seems the current policy is to let people do whatever until they get found out and slapped down. Gettinwikiwidit (talk) 05:06, 15 January 2021 (UTC)
The review is done by the respective WikiProject the import is touching. The problem is that importers can ignore the respective project. I'd suggest allowing imports only when importers can show they posted on a project's talk page about the import. --SCIdude (talk) 07:21, 15 January 2021 (UTC)
Yeah, I will post the dataset at such WikiProject talk page (if exists) before import, unless there are a dedicated bot approval. This may allow users to point out existing issues.--GZWDer (talk) 09:04, 15 January 2021 (UTC)
@Gettinwikiwidit: Such a topic should be discussed in a RfC; there are too many undiscussed imports by different users already.--GZWDer (talk) 09:06, 15 January 2021 (UTC)
@GZWDer: Fair enough. It's not my axe to grind. I was only trying to help. Gettinwikiwidit (talk) 09:09, 15 January 2021 (UTC)
FWIW, I think something as simple as stating "Our primary goal is to serve as a data store for Wikipedia content" would go a long way towards putting people on the same page. I'm not saying this specific goal should be the focus, but just that having a stated focus helps to orient discussions. Gettinwikiwidit (talk) 11:16, 15 January 2021 (UTC)
I totally agree, I try all the time to avoid thinking about Wikidata as THE knowledge graph of the world. It's not, its simply one of multiple spiders in the web of linked data and this particular spider has a big fat "made to interconnect WMF projects"-stamp on the back. WMF projects appear to touch most of the worlds "notable" knowledge concepts and cultural symbols so it's easy to forget. Anyone can create a Wikibase and start linking and storing their data themselves with or without a link to Wikidata. My latest idea was to have all the worlds shelters in Wikidata, it's in the scope of wikitravel, but collecting it on Wikidata first as I began doing is a really no-no I now understand. We link to other stuff. We are not the source of truth for anything.--So9q (talk) 19:45, 15 January 2021 (UTC)
Although not well defined in WD:N, the practice is we may include every knowledge with a reliable source. Maintaining a common Wikidata will make information easily searchable, quaeyable and usable by Wikipedia. Also, it may be a governance and maintanence burden if there are many Wikibase instance in Wikimedia (consider how to deal with conduct disputes - Wikidata already have inefficient procedures on this). In Wikidata:WikiCite/Roadmap, it is proposed that citation data may live in a separate namespace or a new wiki, but it is not a practice accepted by community.--GZWDer (talk) 05:01, 16 January 2021 (UTC)
I suspect this "choice" is what leads much of the effort here to resemble the Wild West. Without a common purpose, people will follow their own whims. Gettinwikiwidit (talk) 10:17, 16 January 2021 (UTC)
@Gettinwikiwidit: That's like saying that every city is like the Wild West because different people who live in the city have different goals. Part of what makes Wikidata great is that allows people with different goals to collaborate and benefit from each others efforts. ChristianKl ❪✉❫ 13:13, 16 January 2021 (UTC)
@ChristianKl: This is probably stretching the metaphor farther than is useful, but in the context of this thread it's worth noting that cities have traffic lights and parking spaces and mass transport that requires a fare. The controls are essential for the smooth interaction of people working closely together. In any event, I'm just sharing a personal opinion. Make of it what you will. Regards, Gettinwikiwidit (talk) 09:29, 17 January 2021 (UTC)
@Gettinwikiwidit: Having rules and policies is indeed important but it's different then having the same goal. Good rules allow people with different goals to happily cooperate. ChristianKl ❪✉❫ 19:21, 17 January 2021 (UTC)
@ChristianKl: As I see it there is a difference between a goal and a motivation. People will obviously bring all sorts of motivations to their efforts and hope to get all sorts of different things out of the project. You can still have an overall (abstract) goal of the project to use as a yardstick to make sure people aren't pulling in opposite directions Gettinwikiwidit (talk) 22:56, 17 January 2021 (UTC)

Can I use media legend (P2096) for the description of a youtube channel?

I have collected urls and the description of each video of a youtube channel that I'm gonna import to Wikidata, but I'm wondering if P2096 is an adequate property for this, or if there's another property more suitable for this. Of if I should do this at all. Tetizeraz (talk) 06:30, 18 January 2021 (UTC)

I think title (P1476) is better. --Matěj Suchánek (talk) 10:52, 18 January 2021 (UTC)

South African Languages

I want to ask why you didn't put all South African languages. Because You put English, Afrikaans, sotho, zulu, Xhosa but why not Tsonga, Siswati & Venda?

phab:T263946.--GZWDer (talk) 10:46, 18 January 2021 (UTC)

New URLs for the Wikidata Analytics dashboard

Hello all,

The Wikidata Analytics dashboards are now gathered on a new portal: https://wikidata-analytics.wmcloud.org/ . You can for example find the dashboard about Wikidata usage in the WIkimedia projects on https://wikidata-analytics.wmcloud.org/app/WD_percentUsageDashboard.

The new dashboards include more interactive visualizations, a new analytics systems scale allowing a better share of resources and therefore an easier access to the boards, and various improvements to the code that will make it easier to add new dashboards.

The boards now have new URLs, and the old ones (for example https://wmdeanalytics.wmflabs.org/WDCM_UsageDashboard/ ) will be maintained until February 28th. We apologize for this inconvenience.

Feel free to update your bookmarks and links. Here you can find the list of all boards with the old and new URLs. We will take care of updating them on the documentation pages that we maintain.

Thanks for your understanding, feel free to reach out to me if you have any questions or issues. You can also add a comment under this Phabricator task. Lea Lacroix (WMDE) (talk) 10:03, 18 January 2021 (UTC)

Help with research

Hi there! I hope that I'm in the right place and that I can be helped here. I am looking for a user who has a subscription to "The Telegraph" and who could verify information in an article for me. Could someone help me with that? --Gymnicus (talk) 11:15, 18 January 2021 (UTC)

Wikidata weekly summary #451

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
  • Development
    • Continued working on the Query Builder. We continued to set up tracking for the number of queries that result in an error, timeout, etc. (phab:T269224). We als made it possible to query for Item values (phab:T268814). You can always try the current state on the test system.
    • Finished fixing an issue with whitespace being stripped while typing when editing Lexemes (phab:T250550)
    • Worked on no error being shown when trying to save Lemma with invalid language code (phab:T265783)
    • Working on using Special:MyLanguage feature for more links in the Wikidata Query Service interface to redirect people to documentation in their language (phab:T267656)
    • Continued interviewing people to better understand how to best compare Wikidata's data against other databases in order to find mismatches that need the editors' attention

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

Males carrying female names

Working most of the time on (especially Turkish) given and family names, I noticed that there are male people who carry female names, just like females carrying male names. In the case of females carrying male names we have a solution: "series ordinal: male given name borne by a female". (See the case of Şerif Sezer (Q8082498)...) We must also create a similar "female given name borne by a male" thingy for men carrying female names. I do not know how to do that.

Please note that in my language, Turkish, there are many "unisex" names like Ufuk, Engin, Aydan, Deniz etc but some names, like Elçin is only/normally a female name and I have seen Turkish men (very rare, indeed only one) with that name. (In Azerbaijan Elçin is a male name, but I am speaking about the Turkish name). This is not the only case BTW.

Can someone make a... (how is it called?) for "female given name borne by a male", please? --E4024 (talk) 13:33, 17 January 2021 (UTC)

Actually, series ordinal (P1545) was created for a quite different purpose…--Shlomo (talk) 09:30, 18 January 2021 (UTC)
I don't understand the need. For instance, Alexis (Q1558167) is very "man" for the Frenchs and the Greeks, but very "woman" for the Americans. --Bouzinac 💬●✒️●💛 10:02, 18 January 2021 (UTC)
I see that you don't. I am speaking about "Turkish" names. I discovered a Turkish man who carries a "very female" Turkish name. I am sure that is not the only case. (In parenthesis: Also there is another notable Turkish man who carries a very female -but rarely used in Turkish, if ever- name... The case of "Elçin" above is similar to what you say...) --E4024 (talk) 17:29, 18 January 2021 (UTC)
Sorry, there might be misinterpretation. Could you rewrite "female given name borne by a male" (borne being not understandable, not an english word). Feel free to type in turkish or another languague you know, let's see if google translates works ;) --Bouzinac 💬●✒️●💛 19:49, 18 January 2021 (UTC)
"male given name borne by a female (Q18220911)" exists and was not made by me. If instead of adding smilies you had looked at "the case of Şerif Sezer (Q8082498)" -just as referred to above- you would have seen it. English (and not "english") is not my native tongue but I think I can express my ideas clear enough. We already have a "thingy" for females (if it was written in proper English or not beats me) and I asked for the possibility of making a similar one for males. If you have an intention to help you can make this one in proper English and also correct the other one. I think for some reason Turk bashing is more attractive... (or maybe you wish to criticise the colleague who wrote the "borne" thingy over me?) Whatever. Have a good time. --E4024 (talk) 21:16, 19 January 2021 (UTC)

Phase 2 of the Universal Code of Conduct consultation is starting!

Hello, as you might probably know, Wikimedia Foundation is currently conducting a movement-wide consultation about the new Wikimedia movement’s Universal Code of Conduct (UCoC).

What is the Universal Code of Conduct?

UCoC is an attempt at drawing a basic level of norms and behaviours that everyone is expected to follow, regardless of their background, their level of commitment or way in which such commitment is expressed (i.e. online or offline).

As you will see, it is a collection of very basic principles about how to behave in a collaborative environment, most of which our community already implements and shares as a value. It is, though, an important attempt at strengthening our mutual commitment to protect those users who might suffer from harassment, abuse or threatening behaviours, because of their involvement on the projects.

What is Wikimedia Foundation asking us to do?

In the following months, a movement-wide consultation will be carried on by various facilitators (like me), in order to ask for and collect relevant feedback from all Wikimedia communities. The consultation is scheduled to go on until the end of February 2021.

In particular, we will ask you – through general posts like this or personal requests – your inputs on possible enforcement pathways for UCoC, your overall experience on the projects, and your point of view about how to improve and/or revise the current draft, as well as the current project’s guidelines.

How can I help?

You can help by expressing your opinion, at any given time and in any way you may consider useful (for example, publicly in this page or on my talk page, or privately via email), and to engage periodically on posts about UCoC that we will be sharing in the next months.

As your facilitator, I assure that it will be my responsibility to make sure that your voice will be heard. Please remember that there is no such thing as a “stupid opinion” or “worthless idea”, so be bold and feel free to express yourself. :)

Will also chapters or user groups be involved in the discussion?

Yes, they will, in a separate and parallel consultation. If you’re a member of a chapter or a user group, you may as well speak up your mind in that capacity too. We will let you know as soon as possible when the other consultation will begin.

Do you have any questions or doubts? Please, read the FAQs or ask me anything!

Hope to hear from you all soon! --Sannita (WMF) (talk) 12:43, 18 January 2021 (UTC)

FYL Wikidata item subjecting the consultation is here: Q104896623 --*Youngjin (talk) 12:42, 19 January 2021 (UTC)

Q6219564/Q3180938

Same?--GZWDer (talk) 19:20, 19 January 2021 (UTC)

No. an anonymous IP carelessly conflated the two with this edit on Wikipedia, and then robots/robotic humans imported the errors to Wikidata. If the Virginia militia man was active in 1756, he can't have been the Royal Navy officer who died in 1693. -Animalparty (talk) 20:05, 19 January 2021 (UTC)
I have reverted errors and added some references to w:en:John Ashby (militiaman), and removed the errors from Q6219564. -Animalparty (talk) 21:20, 19 January 2021 (UTC)

Merging too objects

Hello Community! I just noticed that there are 2 objects: d: Q209896 and d: Q11415564 that seem to be related to the same thing, so I believe they have to be merged. Also in enW Honoris causa redirects to Honorary degree.
What do you think?

I welcome your opinions.
LiberDIO (talk) 21:46, 19 January 2021 (UTC)

Are there any honorary degrees which are not doctorates? Ghouston (talk) 21:51, 19 January 2021 (UTC)
Well, there are in some rare cases master's degrees, (EDIT: ignore it, I confused the two meanings) but it's basically one the same. The same article in different languages is tied to languages that do not persist in the other article and vice versa, so I noticed that and when I tried to tie them up with my native language, I saw this (I believe) technical error.
LiberDIO (talk) 22:21, 19 January 2021 (UTC)
According to the Japanese Wikipedia they are different, and it has articles for both. Via Google Translate: "The title of the degree is honored, and the honorary doctor is the representative, but some universities have titles such as honorary master's degree and honorary bachelor's degree according to the degree system." There are also Japanese and Korean articles for "honorary master's degree" linked at honorary master's degree (Q11415560). Ghouston (talk) 22:43, 19 January 2021 (UTC)

Update on the support process review

Hello all,

Here’s an update on the support process review project, that has been ongoing since September last year. We didn’t give a lot of news on this page since the end of the survey, but we did a lot of work internally to analyze your feedback and come up with ideas that could help to solve the mentioned problems!

As a reminder, here are the main areas of improvement that we identified, together with you:

  • Having a better/clearer overview of the open bugs and already reported issues
  • Improving the way tasks are triaged and monitored, as well as the creation of Phabricator tickets
  • Being able to suggest and vote for the most important features and bugs to fix
  • The possibility of communicating in other languages than English
  • Making processes and priorities more transparent
  • Reducing the amount of channels where the development team actively provides support

For each area, we’ve been brainstorming on ways to solve these issues, identify the ideas that are possible to implement, and define feasibility and priority for each of them. We could select some concrete tasks that we can work on in the next months, and others that will be worked on later during the year.

These ideas include for example: communicate more about our development roadmap and our priorities (why we choose tasks over others), improve the contact the development team set of pages to improve the documentation on how to report an issue, encourage reports in languages other than English, improve the content and structure of tasks on Phabricator to help finding existing tasks, evaluate the possibility of joining WMF’s community wishlist and include community requests in our roadmap. These are just a few of them, you can find the full list here.

We will keep you updated on the project page, for example to ask your feedback about the redesigned documentation page that we started working on. If you have any feedback, for example if some of the ideas don’t seem useful to you, or if you feel like something important is missing, feel free to add a comment on the talk page.

Thanks for your attention! Léa and Mohammed - Lea Lacroix (WMDE) (talk) 08:59, 20 January 2021 (UTC)

St. Mary's Hospital (Q100276507): hospital in Scranton, Pennsylvania, United States

Some time ago, I created that as a place of birth (P19)-value. It would be good if more information about it could be found. It seems it no longer exists (or was renamed) and somehow disappears among many other hospitals of the same name, in Pennsylvania and elsewhere. --- Jura 11:03, 20 January 2021 (UTC)

Gender forms of surnames

Is there a fixed rule / policy on Wikidata modelling for surnames, which have different forms for different genders of people? This is a very important issue for Polish names. We've recently had a discussion about it in the main Polish-language Wikimedia related group on Facebook. Most people favoured creation of separate items for all forms and there was also an idea to link those items using surname or patronymic for other sex (P5278). However, FB discussions in one language are obviously non-binding for Wikidata community, so I want to bring it up here. Powerek38 (talk) 17:08, 7 January 2021 (UTC)

Ash Crow
Dereckson
Harmonia Amanda
Hsarrazin
Jura
Чаховіч Уладзіслаў
Joxemai
Place Clichy
Branthecan
Azertus
Jon Harald Søby
PKM
Pmt
Sight Contamination
MaksOttoVonStirlitz
BeatrixBelibaste
Moebeus
Dcflyer
Looniverse
Aya Reyad
Infovarius
Tris T7
Klaas 'Z4us' van B. V
Deborahjay
Bruno Biondi
ZI Jony
Laddo
Da Dapper Don
Data Gamer
Luca favorido
The Sir of Data Analytics
Skim
E4024
JhowieNitnek
Envlh
Susanna Giaccai
Epìdosis
Aluxosm
Dnshitobu
Ruky Wunpini
Balû
★Trekker

Notified participants of WikiProject Names ChristianKl ❪✉❫ 21:55, 7 January 2021 (UTC)

Happens in many Slavic languages. Example: Stepanova in Russia and Ukraine. Klaas `Z4␟` V:  08:20, 8 January 2021 (UTC)
FB is free to do whatever they want there. I don't think this is the place to discuss Facebook. --- Jura 12:31, 8 January 2021 (UTC)
@Jura1: I'm sorry, but I don't think you've understood what I wrote above. I wrote that wikimedians were discussing this matter using FB and now I am bringing it up here. This has nothing to do with Facebook, apart from the fact that it was used as a convenient forum for initial discussion. Powerek38 (talk) 16:45, 8 January 2021 (UTC)
Ok. Wikidata generally doesn't care about FB discussions. Don't some wikis have explicit policies prohibiting offsite discussions? --- Jura 08:00, 9 January 2021 (UTC)
Please try to be less offensive, Jura. This is really unnecessary. If you have a problem with some user's behavior or some Wikidata policy, please bring it up separately.Vojtěch Dostál (talk) 16:03, 11 January 2021 (UTC)
Personally, I find it offensive to conduct policy discussion on another website. Can we know who were the users who participated in this? --- Jura 16:14, 11 January 2021 (UTC)
I'm sorry, but I think this is really getting absurd, apart from being totally off-topic. Wikimedia, in its various shapes and forms (WMF, projects, chapters, user groups) has had a very strong social media presence for years. There are numerous FB groups for different Wikimedia-related topics. Obviously, no binding decisions are made outside one of wikis, but those informal discussions take place all the time. Our main Facebook group in Polish, which is called simply pl.Wikipedia, is very far from being a secret and it has over 500 members and among its moderators you can find WMPL's Vice President and WMPL's Community Support Officer. And that's just Facebook. The we can talk about Discord with many Wikimedia servers etc. If you really consider it harmful, Jura, it's obviously your right to hold that opinion, but please, don't create any more off-tops out of it. If you fell you need to, initiate, just as Vojtech suggested, a separate discussion, probably on Meta, as it's a movement-wide issue. Powerek38 (talk) 18:12, 11 January 2021 (UTC)
It's really a matter you brought up here. If you don't want to address questions about it, don't. If you don't want to participate onwiki, it's unclear why you come here. Redoing onwiki discussions elsewhere doesn't seem very constructive. --- Jura 11:37, 13 January 2021 (UTC)

My view is that -ova feminine forms should only be created as separate items if it is unclear what the masculine version of the surname actually was. This happens in some languages due to grammatic changes in word endings.

Tobias1984 Vojtěch Dostál YjM Jklamo Walter Klosse Sintakso Matěj Suchánek JAn Dudík Skim Frettie Jura1913 Mormegil Jedudedek marv1N Sapfan Daniel Baránek Draceane Michal Josef Špaček (WMCZ) The photonaut Hartasek Zelenymuzik Gumruch Shadster Dænča M.Rejha Janek Jan Kameníček Eva Vele Linda.jansova Lukša Packa Fukejs Hugo Xmorave2 J.Broukal Lenkakrizova Steam Flow Pavel Bednařík Sanqui

Notified participants of WikiProject Czech Republic as this has been discussed in the Czech community as well and someone may come up with a good example of it happening. Vojtěch Dostál (talk) 16:03, 11 January 2021 (UTC)

Good question, Powerek38. A very important topic for many Slavic languages. I see two possible ways of modelling this. First, have one item per surname and distinguish the gender forms with male form of label (P3321) and female form of label (P2521). Second, have separate item for each gender variant, connected via surname or patronymic for other sex (P5278) (and I can see now we need a better label in Czech, so it's obvious that this property is symmetric). But most importantly, we should stick to one of these schemes and do not mix them up and make no exceptions. --YjM | dc 17:06, 11 January 2021 (UTC)

The problem is both approaches have their upsides and downsides.
For the scheme (currently my slight preference) of sharing one item for both the male and female forms, noted using female form of label (P2521), it is a surname of the father inherited by the daughter / the name of the husband / the name of the sons of the woman; it is also the version used by/for some women internationally (see e.g. Otilie Suková (Q60467588), with randomly found [12], [13], and also “Photo of Josef and Otilie Suk” at en:Josef Suk (composer)). Basically, both gendered and non-gendered version are the same surname, only in a grammatically differing role. The downside is you cannot distinguish women who normally use the gendered version from the exceptional cases of women using the non-gendered version (e.g. Emma Smetana (Q11878748)); also, other languages would always reconstruct the non-gendered version using given name (P735)+family name (P734) (e.g. in English, Petra Kvitová (Q30812) would get Petra (Q740790)+Kvita (Q62740665)).
On the other hand, the separate-items scheme is basically the reverse: The name of the person is trivially composable with the proper version of the surname already set, but you cannot determine what is the “family name” of the woman, i.e. what was her father named / what is her husband’s name / what is her son’s name. And please note surname or patronymic for other sex (P5278) does not really help with that: There are many surnames which share the gendered version, so the item for the gendered version will have multiple P5278 claims (e.g. Blažková (Q43371396): Blažková might be a daughter of Blažek, Blažke, Blažka, Blažko…). On the other third hand, there are even exceptional cases where there are multiple female versions of a single male surname, depending on the family tradition (e.g. for Jirků (Q50308708), the normal female version is identical, Jirků, however, there are women who use Jirkůová). And, this also means the shared-item scheme has a similar drawback for less known persons: you might have an item about e.g. “Jarmila Blažková”, but without other information, you just don’t know if the family name is Blažek, Blažko, etc.
And… to be honest… the choice depends on the use cases for the properties, which I’m not really sure what they are.
--Mormegil (talk) 09:01, 12 January 2021 (UTC)
Another problem with the first model is, that there are (and will be) trigger-happy bots on hunt for people without surname statement. They won't care about verifiable sources and will fill in the missing piece of information using random choice, or maybe some “heuristical” method: e.g., they will add a family name (P734)Suk (Q20998030) statement for any unspecified Ms Suková, claiming that “Suk” is certainly more common surname than “Suka”. Some of them may even check this assumption in statistics but better don't count on it.
The second scheme is less prone to this kind of corruption since the bot (or even the human editor) is not forced to lie about knowing the male form while adding a surname statement to a woman's item. But as you pointed, the missing link to the male form will be still percieved as a problem and sooner or later somebody will introduce some way to fill it in (via a qualifier, maybe). And then we have the aforementioned army of trigger-happy bots here again. In this case though, it wouldn't be an intrinsic problem of the data model, it's just that people expect Wikidata to content even data that are not available in the real world.--Shlomo (talk) 21:32, 14 January 2021 (UTC)
Oh, don’t be that optimistic about trigger-happy bots, they are happy to say such things like Zuzana Dobiašová (Q81902636)given name (P735)Žužana (Q12808063), nothing will stop them in any model.
I don’t have a real solution. The most thorough solution would be to have multiple instances of “Blažková”, each with a different value of surname or patronymic for other sex (P5278). Plus possibly one more with unknown value there for those people we just don’t know. That would be a “correct” model, IMHO, but it is a lot of work with doubtful benefit. I guess I was primarily just objecting to the “we should stick to one of these schemes and do not mix them up and make no exceptions” statement which took quite a simplistic view of the reality…
--Mormegil (talk) 09:05, 15 January 2021 (UTC)
Good points, @Mormegil:. I would only object that every model is a simplification of reality :-) Do you perceive the implication of your statement "There are many surnames which share the gendered version, so the item for the gendered version will have multiple surname or patronymic for other sex (P5278) claims" (added {{P}}) as problematic? --YjM | dc 23:20, 18 January 2021 (UTC)
Similar to @Mormegil:, I also have a slight preference for the "item sharing" model. When necessary, these could be supplemented by qualifiers - eg. Emma Smetana (Q11878748) - family name (P734): Smetana (Q12192297), subject named as (P1810): "Smetana" (string). Maybe subject named as (P1810) is not the best match, but I guess the essence of my solution is understandable. In cases where we cannot determine the real family name (such as many cases of Blažková (Q43371396)), we can just leave it out until the name of this woman's father or other male relatives is determined (and ideally sourced using inferred from (P3452)). Also @Jklamo, Shlomo, YjM, Powerek38: if we could find agreement on something like that. Vojtěch Dostál (talk) 13:27, 18 January 2021 (UTC)
Polish is the only Slavic language, which I know, but my worry is that at least in Polish many of those "real family names" would be in fact male forms, which are often considered to be the "basic" ones. This can might be quite problematic, because in the last few years that has been a general trend in Poland (especially in more progressive circles) to use female forms of nouns as much as possible, when talking about women. This is seen as part of the broader social effort to stress gender equality. There are even many new forms introduced to the language. For example, when I was studying political science 15 years ago, it was obvious that our future profession was called "politolog" (male form of political scientist), regardless of gender. Now we use more and more also the female form "politolożka". What I'm trying to say is that using male forms as basic and female forms as variants could be quite difficult to accept for many members of our community. Powerek38 (talk) 14:16, 18 January 2021 (UTC)
That's a legitimate concern, Powerek38, but my view is that equal treatment of men and women is actually guaranteed by the usage of one common item for all gender variants. The proposed solution recognizes that people belonging to one family have one common family name and the precise gender form is not so important. Input from other members of your community is welcome, of course.Vojtěch Dostál (talk) 17:40, 19 January 2021 (UTC)
I slightly prefer separate-items scheme, as we for the given names we also have separate items for each variant of the name, even with very little difference (like a different diacritical mark over only one letter). I was a bit skeptical about this given names solution, but over time I like it much more.--Jklamo (talk) 20:51, 20 January 2021 (UTC)

JSTOR

Could anyone pls tag it in simple wikipedia per created as here to Q1420342. One more thing can you tell me when will I be an autoconfirmed user.TTP1233 (talk) 10:45, 20 January 2021 (UTC)

@TTP1233: The sitelink has been added by User:BrokenSegue. ArthurPSmith (talk) 18:50, 20 January 2021 (UTC)

Merge Sacrament and Sacraments_of_the_Catholic_Church

I'm not sure what are the differences between https://www.wikidata.org/wiki/Q3233636 and https://www.wikidata.org/wiki/Q49703 Don't they talk about the same thing?  – The preceding unsigned comment was added by 151.64.12.108 (talk • contribs) at 18:08, January 20, 2021‎ (UTC).

The first is general to all christian churches (and perhaps others?) while the second is specific to one denomination. ArthurPSmith (talk) 18:54, 20 January 2021 (UTC)

Parent organization versus part of

I found that more organizations use part of (P361) than parent organization (P749): 

SELECT (COUNT(DISTINCT ?item) AS ?count) WHERE {
	?item wdt:P31/wdt:P279* wd:Q43229; wdt:P361 [].
}
Try it!

gives 134,808 items, while 

SELECT (COUNT(DISTINCT ?item) AS ?count) WHERE {
	?item wdt:P31/wdt:P279* wd:Q43229; wdt:P749 [].
}
Try it!

gives only 97,790. I think that parent organization (P749) is a subproperty of part of (P361), and all usages such that (some organization)part of (P361)(some other organization) should be converted to use parent organization (P749) (and similar has part(s) (P527) statements to has subsidiary (P355)). I’m a bit afraid of asking bot owners to change over a hundred thousand items (although the above figures are somewhat overestimated, as I didn’t check whether the object is an organization—the queries almost timed out even in this simple version), so I welcome any input and ideas about how to deal with these four properties. —Tacsipacsi (talk) 22:59, 19 January 2021 (UTC)

  • What is a "parent organization"? Can you complete the description of the property? Currently it's defined as the opposite of a "subsidiary". Supposedly P749 can only apply to some organizations. --- Jura 23:42, 19 January 2021 (UTC)
There is, as Jura alludes to, a definitional problem. And then, whereas I don't doubt that there are many relations using part of which should use parent, just because both are organisations does not make it a slam-dunk that parent is the more appropriate. Most of my work at the moment is on UK government organisations, very broadly defined. Very many of these use part of (e.g. NHS Trusts are part of the NHS) and it would be inappropriate to use parent (b/c each organisation is autonomous, albeit constrained by its establishing primary & secondary legislation, so the NHS is not the parent organisation of a Trust). You are, thus, right to be afraid of asking bot owners to change over a hundred thousand items, because it would be a really really really bad idea to undertake such an exercise blind and without analysis and appreciation of context. --Tagishsimon (talk) 23:55, 19 January 2021 (UTC)
It's not clear to me that parent organization (P749) can't be used, if an organization is part of another, or owned by another, even if it's more or less autonomous. Ghouston (talk) 00:30, 20 January 2021 (UTC)
It's as simple as, an organisation can be part of some other organisation without that somthing being a parent. And in such cases, using parent/subsid rather than part of/has part would seem to imply that A is a parent of B (when it is not) and B is a subsid of A (when it is not). I'm unsure a) why that is not clear and b) why anyone would wish to choose what seems to be the less appropriate property pair. It seems like an underpant gnome jump is being made from 'it's two organisations that have a relationship' to 'it must be a parent/subsid relationship' without any intervening analysis or logic. From the top of my head, Trusts and the NHS; CRCs and the Probation Service; private prisons and the Prison Service all have non-parent/subsid relations which are frankly *excellently* described by part of/has part. smh. --Tagishsimon (talk) 03:08, 20 January 2021 (UTC)
Better than that, a private prison or a CRC will almost invariably have a parent/subsid relationship - e.g. with Serco, G4S, &c - and has part/part of relations with the Probation and Prison services. Are we seriously expending electrons wondering whether we should set a blind bot off to screw over predicate choices which humans can find difficult to make, on the basis of an ill-considered syllogism? --Tagishsimon (talk) 03:29, 20 January 2021 (UTC)
The private prison would be neither part of nor a subsidiary of the Probation and Prison services, it would be a contractor. Ghouston (talk) 06:04, 20 January 2021 (UTC)
As services/facilities would it be operator (P137) for operators of prisons and CRCs, and part of (P361) for the service they are part of? NHS trusts can just have P361. Peter James (talk) 14:45, 20 January 2021 (UTC)
@Tagishsimon: b) Because they happened to find that one—the labels are inappropriate/missing in their interface language, they didn’t really understand the distinction, or P749 simply didn’t exist when the statement has been added (P361 is a few months older). I understand that I made a too big generalization and this shouldn’t be done in the way I proposed, maybe even the vast majority of the statements my query finds is right, but I’m sure there are wrong statements. The question is only whether we can find and fix them without making any wrong edits. (Probably we can’t.) —Tacsipacsi (talk) 01:10, 21 January 2021 (UTC)
I also see a problem. I am currently organizing the structure of the Federal Government of the United States (Q48525) and notice there are varying differences in choices between the two properties. For example, the Federal Government of the United States (Q48525) has numerous has subsidiary (P355) statements that are technically part of (P361) executive branch of the U.S. government (Q104907540). I'm going to remove and move them, but I'm conflicted about wheter I should keep the current property structure using parent organization (P749) or switch to part of (P361).
I personally think that parent organization (P749) should be removed, or limited to specific types of organizations like companies or NGOs. part of (P361) would then be used for all other cases.
Lectrician1 (talk) 03:12, 20 January 2021 (UTC)
We cannot remove part of (P361). It is used for conceptual concepts such as groups of people like music bands, members of sports teams, etc.
parent organization (P749) seems like it should be used on actual defined organizations. For example, governments, company structures, nonprofits, etc.
I agree that parent organization (P749) should be a subproperty of (P1647) part of (P361) in addition to the current owned by (P127). That way it is known that there the property indicates that there is actual relationship of ownership and hierarchy structure similar, but different to the other use cases mentioned of part of (P361). This also helps establish a link indicating a difference between owned by (P127) and part of (P361).
Lectrician1 (talk) 03:32, 20 January 2021 (UTC)
The credit rating agencies that are part of the Big Three (Q4906536) could have P361 but not P749. There is also member of (P463) which says it can be used for members of a band or organisation; it's a subproperty of affiliation (P1416) so couldn't be used for linking to Q4906536 or anything similar. Peter James (talk) 14:45, 20 January 2021 (UTC)
@Peter James: As far as I understand, Big Three (Q4906536) isn’t an actual organization, just an umbrella term, so my proposal doesn’t apply here. (In any case, it’s not classified on Wikidata as an organization—it’s not classified at all—, so any bot that would have executed this task would have left these untouched.) —Tacsipacsi (talk) 01:10, 21 January 2021 (UTC)
@Tacsipacsi: I agree with other commenters here that they are not interchangeable. One additional use I've seen which is not mentioned above is to use part of (P361) to link to a top-level organization (for example a particular university department to the university) whereas parent organization (P749) is for the next level in the organization (the university faculty that the department actually belongs to). parent organization (P749) expects the inverse has subsidiary (P355) to also be present, which is reasonable, while part of (P361) does not require a has part(s) (P527) inverse, which allows that sort of many-to-one relation to be encoded without overly burdening the top-level organization with a lot of the inverse one-to-many statements. ArthurPSmith (talk) 18:42, 20 January 2021 (UTC)
@ArthurPSmith: I’ve never seen such usage, and I don’t quite understand its purpose. parent organization (P749) is transitive, isn’t it? The department has parent organization (P749) the faculty, which in turn has parent organization (P749) the university, so it’s transitively already described that the department belongs to the university, no additional part of (P361) statement is needed. —Tacsipacsi (talk) 01:10, 21 January 2021 (UTC)
There is also business division (P199) for parts which aren't separate organizations. Ghouston (talk) 02:29, 21 January 2021 (UTC)

Honestly I think people often use "part of"/"owned by" instead of "parent org" because the latter asks for a symmetric statement which is annoying. BrokenSegue (talk) 18:46, 20 January 2021 (UTC)

Petition for a merger

Hi, I would like to merge the following items but I don't understand how: 1080 recetas de cocina (Q84022085) and 1080 recetas de cocina (Q84022082). Can I ask someone to do it for me? Thanks and regards--El Mono Español (talk) 00:31, 21 January 2021 (UTC)

No. One is an item for the literary work; the other an item for an edition of that work. There is more information on the preferred structure for items pertaining to books at Wikidata:WikiProject Books; in particular the Functional Requirements for Bibliographic Records (FRBR) model which differentiates between works and editions. In any event, because one points to the other - Q84022085#P629, they will not merge. --Tagishsimon (talk) 00:47, 21 January 2021 (UTC)

Best way to deal with link rot on "Official Website"

Solved with Template Infobox OS website parameter special value hide

Hi,

What is the best way to deal with an official website URL (P856) Property_talk:P856, which is now invalid?

The item in question is https://www.wikidata.org/w/index.php?title=Q2665351&oldid=733581269

Happily, archive.org got some snapshots of the URL, including:

https://web.archive.org/web/20080506045212/http://www.nec.de/hpc/software/super-ux/index.html

This is the last valid archive for the URL.


For the now invalid valid, should any of the following be set?

Also the URL appears to have been invalid when imported into Wikidata on 14 March 2017 from the Russian Wikipedia, see https://www.wikidata.org/w/index.php?title=Q2665351&oldid=466252341

Does this alter the qualifier properties that should ideally be set?

The Wikidata information is used by the https://en.wikipedia.org/wiki/Template:Infobox_OS of the page https://en.wikipedia.org/w/index.php?title=SUPER-UX&oldid=999525330 .

Thanks! --Lent (talk) 06:29, 14 January 2021 (UTC)

OK, I think I did it reasonably. Added an end time (P582) of 6 May 2008, a reference URL (P854) of https://web.archive.org/web/20080506045212/http://www.nec.de/hpc/software/super-ux/index.html with a retrieved (P813) of 15 January 2021. Anything I should still do, or should have done? See: SUPER-UX (Q2665351) . Thanks! --Lent (talk) 06:24, 15 January 2021 (UTC)
Solved with Template Infobox OS website parameter special value hide

✓ Done Lent (talk) 03:56, 22 January 2021 (UTC)

Program for Cooperative Cataloging Wikidata Pilot

The Identity Management Task Group of the Program for Cooperative Cataloging has launched a Wikidata pilot that will run for at least a year. The Program for Cooperative Cataloging is an international cooperative effort aimed at expanding access to library collections through cataloging that meets mutually-accepted standards of libraries around the world. The pilot is an opportunity for catalogers to begin or continue to work in Wikidata in their everyday work and further integrate contributions to Wikidata into regular metadata workflows in libraries. 70+ institutions are participating in the pilot and are editing and creating items related to their institutions and collections. The work is documented in the WikiProject PCC Wikidata Pilot. Some institutions are using on focus list of Wikimedia project (P5008) to keep track of items they have added when not easily retrieved with a query, or if they intend to go back and further enhance them. Check out the project page if you are interested in further information about the pilot.--Chicagohil (talk) 19:19, 21 January 2021 (UTC)

Two posible values

How do we deal with two posible values for the property date postally canceled (P9052)?Smiley.toerist (talk) 23:27, 21 January 2021 (UTC)

Maybe change the date to "1900s", with earliest date and latest date qualifiers of 1903 and 1905. This doesn't exclude 1904 as a possibility though. Ghouston (talk) 01:42, 22 January 2021 (UTC)

I wil scan the backside and upload it. On second examination I tend to favor '5'. Only the lower part of the digit is visible, however the upper part of the lower semicircle turns downwards. Bij a three it would be horizontal. However the principle still stands: What happens when there are two posibilties? I would favor a broader range 1903-1905. One can always put a comment in the talk page. One should not try to define every posible rare variant.Smiley.toerist (talk) 13:07, 22 January 2021 (UTC)

Smiley.toerist (talk) 17:16, 22 January 2021 (UTC)

Collections split between two (or more) museums

Hello, hoping someone can help. Wondering if there's a better / best way to model a collection of artefacts that has been split between two or more museums? Working with the National Museum of Scotland on some images of the Lewis Chessmen (Q217796), a collection that's split between the National Museum of Scotland, the British Museum, and the Museum nan Eilean (latter is a long term loan from the BM.) Have added NMS as an additional location, was going to add MnE but just in case there's a better way to do this? Advice appreciated. Also for ref, this has come about due to trying to get the object template to display the correct locations on commons (eg here. Sara Thomas (WMUK) (talk) 13:38, 21 January 2021 (UTC)

How do I indicate that values belong to different versions of the item?

I'm updating Ithkuil (Q35846), and I've run into a problem: different versions of the language have wildly different grammars. For example, versions 1 to 3 have fusional grammars, but the WIP version 4 is almost completely agglutinative. Same with the native names: there's a different one for each version. Is there a qualifier I can use to indicate that a value belongs to a certain version or revision?

If they differ so much, and you want to describe a particular version, perhaps a separate item for each version would be helpful. Ghouston (talk) 06:12, 23 January 2021 (UTC)

Items to merge

My script suspects that the items below need to be merged. I already merged dozens, but I have no time left for a while. Would someone please be kind enough to check and merge them? Please write "done" afterwards. Thanks so much! :-)

to merge
* Apostolic Nunciature to Djibouti (Q85742264) and Apostolic Nunciature to Djibouti (Q12784991)

Syced (talk) 09:51, 23 January 2021 (UTC)

Done, with the exception of the US & UN missions, which are both in the US but different things. I've also corrected the operator which, according to enwiki, is the Holy See and not the Vatican. --Matthias Winkelmann (talk) 01:08, 24 January 2021 (UTC)

Alexander Waibel

I have just edited the object for the German article Alexander Waibel. While editing, I noticed that there was another object with the name Alexander Waibel. The two objects appear to be identical in my opinion. I would ask the IT project ( WikiProject Informatics has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.): Do you think that Alexander Waibel (Q14918223) and Alexander Waibel (Q102123856) should be merged? --Gymnicus (talk) 00:07, 30 January 2021 (UTC)

Merged. It's a recent Mathematics Genealogy Project import of 250k or so people. Expect to come across many maths bio duplicates. Be bold and merge them yourself; see Preferences / Gadgets / Merge if you have not got that tool checked. --Tagishsimon (talk) 00:20, 30 January 2021 (UTC)
I know how to merge objects. But I wanted to be on the safe side in this case and wanted to ask you for help. Thanks for quick help. --Gymnicus (talk) 00:32, 30 January 2021 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. --Gymnicus (talk) 00:32, 30 January 2021 (UTC)

Hierarchy of rivers and streams

I know that we have various tools for displaying the family tree of a group of humans; do we have something similar for the tributaries of waterways? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:29, 22 January 2021 (UTC)

@Pigsonthewing: https://www.entitree.com/en/mouth_of_the_watercourse/Danube --GZWDer (talk) 16:51, 23 January 2021 (UTC)
Thank you, but that's more of a linear representation, rather than a tree. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:53, 24 January 2021 (UTC)

QID on mw:Extension:CiteThisPage

Is there a good way we could suggest to interested projects to include the QID for some formats generally output by that extension?

The sample mentioned there is w:Train where the "cite this page" gives https://en.wikipedia.org/wiki/Special:CiteThisPage/Train?page=Train&wpFormIdentifier=titleform

Currently, no format at enwiki includes the QID. Note this is not about the use of the extension on Wikidata itself.


Maybe this interests The Source MetaData WikiProject does not exist. Please correct the name.

The Source MetaData/More WikiProject does not exist. Please correct the name. --- Jura 13:26, 24 January 2021 (UTC)

@Jura1: I think you wish to propose that interested other projects, such as individual Wikipedias for various languages, return the QID in this extension which generates citation metadata.
Ideally decisions about how mw:Extension:CiteThisPage operates could happen at MediaWiki. If I understand correctly, you think or know that the change you propose would require permission from the local volunteer communities at individual Wikimedia projects.
I think you are asking if anyone has ideas for getting this support from various communities.
If that is the case, such broad community discussion over this feature seems unlikely because of its abstractness. Can you say more about what you want your result to be, and what your idea is for getting to that result? Blue Rasberry (talk) 14:57, 24 January 2021 (UTC)
  • Technically, I think it's probably already possible to output the QID.
    The question is more what would be the best way to include the QID in various citation formats. Contributors at Wikidata might be more likely to know about that or interested in figuring it out.
    Once it's clear, we can suggest it here and there. Supposedly each wiki maintains its own MediaWiki:Citethispage-content --- Jura 15:03, 24 January 2021 (UTC)

Are Holocaust-Transports P2632 or P793

Hello,

are Holocaust-Transports in Personitems better as P2632 or P793.

Examples:

Q76220 - here the transport is in P2632

Q100152745 - her it is in P793

What do you think and how can the wrong ones mass-changed?


Thanks --McSearch (talk) 16:14, 24 January 2021 (UTC)

significant event (P793) is used in 77,415 cases, place of detention (P2632) in 503 cases. Changing these would tend to involve some SPARQL to get data for the 503, and quickstatements or wikibase-cli to remove P2632 data and add P793.
SELECT ?propertyItemLabel (count(?propertyItem) as ?count) WHERE {
  ?item ?propertyRel ?transport.
  ?transport wdt:P31 wd:Q61927259.
  ?propertyItem wikibase:directClaim ?propertyRel  
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }  
} group by ?propertyItemLabel
Try it!
--Tagishsimon (talk) 16:59, 24 January 2021 (UTC)

ISO 639-3 language code issues

While working with lists of ISO 639-3 codes from wikidata entities I found Kituba (Q35746) that refers to the Kituba language in general which is accompanied by Kituba (Democratic Republic of Congo) (Q63283489) (Democratic Republic of Congo) and Kituba (Congo) (Q12953639) (Republic of the Congo). Each of these last two has its ISO code set correctly, (ktu and mkw respectively) but the general one has both codes. I noticed the property has single and distinct value constraints which raises warnings in the first entity I mentioned. My question is: since there are entities for each of the languages, shouldn't we remove the ISO codes from the general entity?.
Thanks for your attention, josecurioso ❯❯❯ Tell me! 21:45, 23 January 2021 (UTC)

Hello @josecurioso: I agree that the code should be removed from the general item. While I have not looked much into languages, a similar situation exists for units of measurement, where we sometimes have one general item and various more specific versions (say, for US and UK). Usually the latter ones have conversion factors and external identifiers pointing to other ontologies. Toni 001 (talk) 09:18, 25 January 2021 (UTC)

Wikidata weekly summary #452

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Events
    • Past: Wikidata+Wikibase office hour happened on January 21st. See notes here (English).
    • Past: 2021 OpenStreetMap x Wikidata Beginner Teaching Workshop (in Chinese). Replay on YouTube 1, 2, 3, 4
    • Upcoming: Next Linked Data for Libraries LD4 Wikidata Affinity Group call: Wikibase, the Integrated Authority File (GND) pilot at the German National Library, and the WikiLibrary Manifesto, 26 January. Agenda
    • Upcoming: SPARQL queries and OpenRefine live on Twitch and in French by Vigneron, Januray 26 at 18:00 CET
    • Upcoming video: LIVE Wikidata editing #29 - YouTube, Facebook, 30 January
    • Upcoming: Online Wikidata meetup in Swedish #47, 31 January
    • Upcoming: The Open Festival of Persistent Identifiers starts from 27 January 2021. See schedule here.
  • Tool of the week
    • TAViewer is a web-based anatomy atlas viewer that cross-references Terminologia Anatomica terms with data from Wikipedia and Wikidata.
  • Development
    • Continued working on the Query Builder: made it possible to limit the number of results in the result set, made it possible to include results from subclasses when querying for Item values, worked on negating queries and started on sharing of queries via URL. As always you can try the current state on the test system.
    • Making links in the Query Service redirect to the proper translations of their pages on-wiki by using Special:MyLanguage redirects (phab:T267656)
    • Working on fixing an issue with displaying large numbers for quantity values (phab:T268456)
    • Fixed a case where no error message was shown when trying to save a Lemma with an invalid language code (phab:T265783)
    • Talked to a few more people about checking Wikidata's data against other databases to find potential errors
    • Continuing the Release Strategy project for Wikibase, with the following progress:
      • Created a “release pipeline” for Wikibase extension (phab:T267627)
      • Created MediaWiki docker images based off release branches (phab:T270133)
      • Defined format and contents of Wikibase suite components not maintained by Wikimedia Germany (phab:T271194)

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

Lexeme compound word

Hi all! I'm new to Lexemes. I was wondering how to properly denote that the following lexeme words are compound words:

Is there currently a model on how to do this? (t) Josve05a (c) 22:57, 21 January 2021 (UTC)

@Josve05a: Based on the property examples, it looks like combines lexemes (P5238) does what you want; I'm not sure how you would encode the "-s-", but unless that has meaning in Swedish, that might not be an issue. Vahurzpu (talk) 01:34, 22 January 2021 (UTC)
@Vahurzpu: In this specific case, I would describe -s- as a linking morpheme (Q1472909). The construction "landkamp" does not fit the mouth very well. By adding -s-, the parts become joinable. Sometimes, they exists in the spoken language, but are missing in the written language. Where to use them and where to not use them, requires an academic degree to master. 62 etc (talk) 17:38, 22 January 2021 (UTC)
Exactly, see https://www.wikidata.org/wiki/Lexeme:L60603 for an example --So9q (talk) 23:00, 24 January 2021 (UTC)
@Josve05a: An example that is a bit more complete is smörgåsbord (L54463). Ainali (talk) 22:43, 25 January 2021 (UTC)

I blanked neurodiversity movement (Q104773840). Can now be deleted or re-used

I thought the Wikiversity project "The Neurodiversity Movement" needed its own Wikidata item. After consulting another user it was suggested to link to the general Wikidata item neurodiversity movement (Q56296949). Now I've done that and the previous Wikidata item is no longer needed so I blanked it. Was that proper or should I do anything else next time? LotsofTheories (talk) 05:15, 31 January 2021 (UTC)

I have merged it so it's practically deleted. --SCIdude (talk) 08:04, 31 January 2021 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. SCIdude (talk) 08:04, 31 January 2021 (UTC)

Twinkle?

Do we have that here? --Derpdart56 (talk) 14:59, 22 January 2021 (UTC)

I don't know we need Twinkle or not but I have never used it in this project. Can you give some examples of how it can be used here? Alphama (talk) 07:39, 25 January 2021 (UTC)

@Alphama: See en:Wikipedia:Twinkle. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:59, 25 January 2021 (UTC)
@Pigsonthewing: I don't mean that. I need examples how Twinkle can work on Wikidata. I did research Twinkle before. Alphama (talk) 11:20, 26 January 2021 (UTC)

We don't have twinkle, but these few can help:

‐‐1997kB (talk) 13:08, 26 January 2021 (UTC)

Do all courses on Wikiversity ie. Python Concepts and Python Programming get Wikidata items?

Python Python (Q28865) is an article(is this a "resource"?) on Wikiversity which has two courses:

My question is, should all courses on Wikiversity have Wikidata items and if yes how should they be tagged? Will all Wikiversity courses have instance of (P31) and subclass of (P279) values?

If you wish you can help me by analysing the v:Python(to help me if anything listed there could need its own Wikidata item) ie. we also have v:Python/pip (package manager) but in my mind that page "belongs to Python" making me assume that it should not have a Wikidata item because v:Python which it belongs to already has a Wikidata item. LotsofTheories (talk) 06:19, 26 January 2021 (UTC)

Best practice / lesson learned for External Identifiers

I did this list User:Salgo60/ExternalIdentifiers gathering problems I have seen when connecting Wikidata with an external source.... Please add/comment.... - Salgo60 (talk) 07:47, 26 January 2021 (UTC)

Jorge Bernal mixup

We have Jorge Bernal (Q26897903) and Jorge Bernal (Q26897903), which both seem to describe the same soccer player (but it could be also two different players?), and in the latter one, a TV presenter is mixed up also. Can somebody please clean up those items? Steak (talk) 10:17, 25 January 2021 (UTC)

Jorge Bernal (Q6277748) is the original item; the English article, originally about the footballer, had been replaced with content about a television presenter and was moved to a new item Q26897903, linking it to es:Jorge Bernal (presentador de televisión). When the changes were reverted the link wasn't moved back, so content about the footballer was added there. The Arabic pages are primarily about the footballer, but mix information about both people. Peter James (talk) 13:38, 25 January 2021 (UTC)
Q6277748 is the footballer, and Q26897903 is the television presenter. I moved the English and Arabic links to Q6277748. Peter James (talk) 18:33, 26 January 2021 (UTC)

How do I request revdel?

Can't find a page describing how to request a revdel. Also, if there's an email link as well, I might prefer to do it that way. Thanks, Mathglot (talk) 21:14, 26 January 2021 (UTC)

@Mathglot: Our living people policy describes how to request removal in https://www.wikidata.org/wiki/Wikidata:Living_people#Requests_for_the_removal_of_private_information privacy@wikidata.org is the relevant email address. In the spirit of making things easier for the next position in your situation, where did you search for this information and expected to find it? ChristianKl ❪✉❫ 21:33, 26 January 2021 (UTC)

Consultation for Universal Code of Conduct - first round of questions is up!

Hello everybody! It's time to kick-start the discussion about the Universal Code of Conduct and how to implement it. So let's get started with the first round of questions, about the pathways for UCoC enforcement.

The three questions that will be part of this round are:

  1. What are the best paths for delicate issues?
    For example, consider a user belonging to a "vulnerable group" (in the broadest, most comprehensive meaning of the term), who is a target of actual (not perceived) harassment and/or credible threats of violence through our channels. In your opinion, how this case can be addressed? What would be the best path on the project for this user to report what is happening?
  2. How do we create better reporting pathways for people who are targets of harassment?
    In other words, do you think our current ways of dealing with non-welcome behaviours is the best way to report and deal with harassment too? Or do you think a separate or a complementary procedure should be set up? If yes, how?
  3. How do we deal with incidents that take place beyond the Wikimedia projects, but are directly related to them?
    For example, consider a Wikimedia-related discussion between Wikimedians that takes place on a non-Wikimedia platform, such as Telegram or Facebook or a mailing list, that degenerates in public harassment or other non-welcome behaviours. How would you deal with it? Should it have repercussions also on the projects?

Bear in mind that this is only the first round of questions, and that there are other aspects that will be taken into consideration, but if you think something's missing, please do bring it up in the discussion!

Be bold and have your say at the Consultation talk page! --Sannita (WMF) (talk) 21:23, 26 January 2021 (UTC)

Sovereign states, not countries

What a constantly changing list... Now we have additional hundred of them: https://w.wiki/uxM. More work of adding heads and prime-ministers and ambassies... --Infovarius (talk) 22:01, 19 January 2021 (UTC)

GIGO. federally recognized Native American tribe in the United States (Q7840353) does not fit the definition of a sovereign state, which according to en wiki is: "International law defines sovereign states as having a permanent population, defined territory, one government and the capacity to enter into relations with other sovereign states.[1] It is also normally understood that a sovereign state is neither dependent on nor subjected to any other power or state." So removing that from the SS class tree (diff) will probably sort out the issue, until some bright spark re-adds it. --Tagishsimon (talk) 00:00, 20 January 2021 (UTC)
Federally recognized Indian tribes do fit that definition completely. They have a permanent population, defined territory (they are geographical entities in the Library of Congress name authority files), they have their own government that is sovereign over that territory, and they have the capacity to enter into relations with other sovereign state (they have treaties between their government and the government of the United States and some other countries as well). All four of these criteria are discussed in this article: Sovereignty and Self-Determination: The Rights of Native Americans under International Law. Buffalo Law Review, volume 27, no. 4 (1978), pages 669-714. See pages at 673-679 at https://core.ac.uk/download/pdf/236350558.pdf. From page 712: "Supreme Court decisions repeatedly refer to Indian tribes as semi-sovereign states, recognizing at least a degree of their independence, and therefore debunking the theory that Indians are merely a part of the United States." Ultimately, these tribes are (dependent) sovereign states. That makes including them in Wikidata as subclass of sovereign state completely correct.

If you look at the Art & Architecture Thesaurus, among other controlled vocabularies, you see that the term tribal nations is a subclass of the term sovereign states (http://vocab.getty.edu/page/aat/300403959):

................ political administrative bodies
.................... <political administrative bodies by degree of independence>
........................ sovereign states
............................ nations
................................ tribal nations

Incidentally, countries in the AAT are also a subclass of sovereign states, that is, they are not the same thing as sovereign states: http://vocab.getty.edu/page/aat/300387506

........................ <political administrative bodies by degree of independence>
............................ sovereign states
................................ city-states
.................................... altepemeh
................................ countries (sovereign states)
................................ empires (sovereign states)
................................ nations
.................................... developing countries
.................................... island nations
.................................... republics
.................................... tribal nations
................................ polities
................................ suzerains (states)

See also additional information that I posted on the Snoqualmie Indian Tribe talk page: https://www.wikidata.org/wiki/Talk:Q141646 UWashPrincipalCataloger (talk) 02:21, 20 January 2021 (UTC)

@UWashPrincipalCataloger: A better link for the hierarchy is http://vocab.getty.edu/hier/aat/300387506 --00:13, 27 January 2021 (UTC)
That may or may not be a good way of arranging entries in a thesaurus. Wikidata is not a thesaurus. As you note, TNs are not the same thing as sovereign states. As you note, they are (dependent) sovereign states. (which != sovereign states). As you note, the expressed legal position is "Supreme Court decisions repeatedly refer to Indian tribes as semi-sovereign states". (which != sovereign states)
In set theory terms, the definition for a subclass is: given classes A and B, A is a subclass of B if and only if every member of A is also a member of B. Self evidently semi / dependant / not-the-same-thing things of type A are not members of class B b/c they lack attributes required for membership of B.
My understanding of subclasses in Wikidata is that they are bound by the above definiton and share entirely the characteristics of their superclass (and can be differentiated from the superclass by some other measures). Semi / dependant / not-the-same-thing things do not share entirely the characteristics of their superclass. If we accept Getty, then countries like England, Wales, are sovereign states b/c country is a subclass of sovereign state. England is not a sovereign state. Wales is not a sovereign state. Wikidata is not a thesaurus.
The State class tree would be more useful if along the lines of:
State
Sovereign State
Semi-Sovereign State
Country
than
State
Sovereign State
Semi-Sovereign State
Country
b/c the first reflects the differences between the three state subclasses and respects basic set theory, and the second suggests, wholly incorrectly, that Semi-Sovereign States and Countries have the characteristics of Sovereign States. I don't know what the goals of a thesaurus cataloger are, but it seems not to be the same as a wikidata ontologist. --Tagishsimon (talk) 19:44, 22 January 2021 (UTC)
And for the avoidance of doubt, the "more useful if along the lines of" example, above, is drawn on the fly to illustrate issues with the AAT/Getty thesaurus hierarchies, rather than as properly analysed proposals for wikidata's future. The main point it seeks to make is that Sovereign States and Semi-Sovereign States are two distinct things, both members of a common superclass. --Tagishsimon (talk) 19:51, 22 January 2021 (UTC)
I've changed federally recognized Native American tribe in the United States (Q7840353) to be a subclass of state (Q7275). This is set theory 101. --Tagishsimon (talk) 00:41, 23 January 2021 (UTC)


Might be better querying that way, with some rules to "clean" data. (wonder why Q104906060 still appears in the results)
SELECT distinct ?item ?itemLabel WHERE {
  ?item wdt:P31/wdt:P279* wd:Q3624078.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
 MINUS {?item wdt:P576 [].} 
  MINUS {
    ?item wdt:P31/wdt:P279* ?without.
  VALUES ?without { 
               wd:Q133311 # tribes
               wd:Q13417114 # houses
             }
  } 
  }
Try it!
Bouzinac 💬●✒️●💛 08:24, 22 January 2021 (UTC)

Let's reach consensus on descriptions for our scientific article descriptions

Hi, recently we discussed in the telegram channel how to describe these by bot, because no description on a lot of items yields a pretty awful UI experience in the wbsearchentities API (upper right search box).

I have 1 proposal, feel free to add your own (or your dogs):

  • scientific article published in {label of journal} on {date} by {count} authors"

Proposal by Nikki:

  • scientific article published in {label of journal} in {year}"

--So9q (talk) 16:32, 26 January 2021 (UTC)

It should be as terse as needed, and only be expanded if there are more than one articles with the same name in the same year in the same publication. The problem is the display box for description is about 80 characters, and you will only see the first 80 characters when typing. --RAN (talk) 18:56, 26 January 2021 (UTC)
I don't think this is a good place for this discussion. https://www.wikidata.org/wiki/Wikidata:WikiProject_Source_MetaData seems to be a better place. ChristianKl ❪✉❫ 21:47, 26 January 2021 (UTC)
Thanks for the proposal, see https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Source_MetaData#Discussion_about_descriptions_of_scientific_articles and please go there to discuss further.--So9q (talk) 22:30, 26 January 2021 (UTC)

Multiple start and end dates within a single property

So I have an question regarding the best way to handle multiple start and end dates within a single property. The specific example I'm having issues with is on Martin Mayhew's entry (Q6776149) under the "member of sports team" property, as he formerly played for the Washington NFL team in the early 1990s and was just hired as the team's GM last week. Due to this, I can not have separate "stints" for him as a player and now GM within the same team/label, at least not cleanly. Any suggestions or options? Maybe creating a new "executive of sports team" label to go along with the "member of sports team" and "coach of sports team" ones? - Dissident93 (talk) 22:23, 26 January 2021 (UTC)

@Dissident93: are GM's members of a sports team? maybe use employer (P108)? we also have a coach property. If you must have multiple stints you can do it like we do politicians serving multiple (non-consecutiive) terms in the same position. just list the position/team multiple times with different qualifiers specifying the time ranges. BrokenSegue (talk) 23:22, 26 January 2021 (UTC)
@Dissident93: I see, I had not thought to check politicians for examples. Failing that, employer (P108) would work. Thanks. - Dissident93 (talk) 00:15, 27 January 2021 (UTC)

parliamentary group (P4100) for Angus King (Q544464) and Bernie Sanders (Q359442)

Currently this is set to independent politician (Q327591). Which is probably correct in terms of the "party" they were elected for.

As the qualifier is for "parliamentary group" during the term, shouldn't this be Democratic Party (Q29552)? --- Jura 20:50, 21 January 2021 (UTC)

I'd assume you have to be a member of the party to be part of their political faction. Otherwise, you've just got some kind of political alliance or coalition. Ghouston (talk) 21:41, 21 January 2021 (UTC)
Some affiliation is there, otherwise there wouldn't be a 50:50 split. In Wikidata, it's currently 50R, 48D, 2?. Maybe the value shouldn't be Democratic Party (Q29552), but Democratic caucus or even Democratic caucus of the US senate? Possibly, in the US that difference doesn't matter that much and we could stay with Democratic Party (Q29552).
The entire point of having a qualifier for the parliamentary group is that we can added that grouping.
BTW list of current United States Senators (Q2297935) has two queries for the current composition. They currently do list 100 members! --- Jura 21:52, 21 January 2021 (UTC)
Then I'd say it comes down to external sources: is there documentation for the idea that these politicians are part of the Democratic Party, or the Democratic Party caucus, or whatever it is we are saying that they are part of? Ghouston (talk) 22:59, 21 January 2021 (UTC)
This from the Wikipedia article on Angus King: "For committee assignment purposes, he caucuses with the Democratic Party. He is one of two independents currently serving in the Senate, the other being Bernie Sanders of Vermont, who also caucuses with the Democrats." UWashPrincipalCataloger (talk) 07:06, 22 January 2021 (UTC)
* A similar wording can be found on the senate website. --- Jura 10:28, 22 January 2021 (UTC)
You can also have a 50:50 split between a party and its allies and another party and it's allies (if the Republicans have anything similar). Ghouston (talk) 23:01, 21 January 2021 (UTC)
I think the standard terminology is that Bernie Sanders "caucuses with" the Democratic Party. I'm not sure how this would/should be represented. Gettinwikiwidit (talk) 07:06, 22 January 2021 (UTC)
I'm not 100% sure what you're trying to capture. Are you looking to search for times in which the Senate was split 50/50? I don't agree that which party a senator "tends to vote with" is the primary information we're trying to capture with parliamentary group (P4100). Gettinwikiwidit (talk) 07:10, 22 January 2021 (UTC)
@Andrew Gray, Oravrattas: who have previously thought about these issues, per Wikidata:Property_proposal/parliamentary_group. FWIW the relation of King/Sanders goes beyond "tends to vote with" and is closer to "members in all but name", in that they regularly attend party events, go on retreats with them and, you know, sometimes run for their candidacy for President. Matthias Winkelmann (talk) 10:20, 22 January 2021 (UTC)
@Matthias Winkelmann: Party events and retreats sound far from anything official. Even if they voted 100% with Democrats, the relationship isn't binding nor registered in any way. Gettinwikiwidit (talk) 12:51, 22 January 2021 (UTC)
  • @Dipsacus fullonum: who proposed the qualifier. --- Jura 10:28, 22 January 2021 (UTC)
    Thank for the ping. I proposed the property "parliamentary group" to express who are working together in a parliament. The parliamentary group can be but doesn't have to be members of a certain political party. I don't know enough about the US Senate to say if the 48 democrats and 2 independents together is a parliamentary group, but if these 50 persons do have joint meetings where they discuss and coordinate their politics, I would say yes. Finally I think it is a mistake to use independent politician (Q327591) for a parliamentary group. That isn't a parliamentary group because different independent members of a parliament don't necessarily have any kind of collaboration. I propose to instead use the special value "no value" if a member of parliament doesn't belong to any parliamentary group. --Dipsacus fullonum (talk) 15:36, 22 January 2021 (UTC)
  • Maybe it should be Senate Democratic Caucus (Q3117832)? ChristianKl ❪✉❫ 17:18, 22 January 2021 (UTC)
  • I think this is one of those cases where you might reasonably expect either approach depending on what question you're thinking of - if it's "tell me the exact party affiliation of Senators" then you probably expect to have these two listed as some kind of independent, but if it's "tell me the overall voting split" you'd probably expect just to get 50:50. So I'm not sure there's a "perfect" answer here - either approach has flaws, and neither is obviously right or obviously wrong.
I wonder if a good solution here would be to give them a parliamentary group (P4100) that is a "flavour" of independent, something like "independent caucusing with the Democrats". It means you'll still return Sanders and King in a different group from Leahy and Schumer, but make it clear that they are aligned with one side rather than another. Andrew Gray (talk) 21:25, 22 January 2021 (UTC)
@Andrew Gray: Why do you think the Democratic caucus isn't a parliamentary group and we should make up a new parliamentary group? ChristianKl ❪✉❫ 22:15, 22 January 2021 (UTC)
@ChristianKl: I think using "Democratic caucus" has the problem that it makes it a little difficult to obviously see the difference between Sanders and "full Democrats", though I'd agree it's better than saying Sanders should be listed as "Democratic Party". Personally I think the type-of-independent approach seems a bit clearer. I don't think it's really "making up" a group, though. Andrew Gray (talk) 22:26, 22 January 2021 (UTC)
I'm not sure that this is valuable to capture. For real questions you're likely looking for other information, like votes on a specific topic or chairs of specific chairs. In my book it's better to add more information than obscure information which exists to cover more ground. Gettinwikiwidit (talk) 22:43, 22 January 2021 (UTC)
But if we're after clarity, then parliamentary group (P4100) probably isn't appropriate for indicating party elected on because the Senate is not a parliament. It was used only to maintain a facsimile of consistency across nations. Gettinwikiwidit (talk) 22:57, 22 January 2021 (UTC)
I think a parliamentary group is the kind of thing that regulations thus as the regulation that govers committee appointsments and sometimes speaking slots care about. The term is not about the party ticket on which a person got their seat.
Sorry. I'm having trouble parsing this. I'm not sure what you mean by "governs". Do you have specific rules you can cite? I'd rather have facts which can be referenced than try to capture people's sense of how things should work. Gettinwikiwidit (talk) 23:50, 22 January 2021 (UTC)
As far as the senate being a parliament, the main definition of parliament that Wikipedia uses seems to be that it's a legislative body. I think the senate qualifies under such a broad definition (and it's reasonable to use it here). ChristianKl ❪✉❫ 23:16, 22 January 2021 (UTC)
I looked at this previously, and it seemed like a "parliament" is a just a legislature (Q11204) that happens to include "parliament" in it's English name: Talk:Q35749. Interestingly, Parliament of the United Kingdom (Q11010) isn't an instance of parliament (Q35749), or any subclass of it. Ghouston (talk) 00:40, 23 January 2021 (UTC)
  • I think it's fairly clear that this goes beyond mere voting preference. For parliaments with more than one or two parties, it's not uncommon that there is a difference between party membership and the parliamentary group of its members. Not too long ago, we discussed how to reflect that two Canadian minister had been excluded from their group in parliament, but not from the party. There P4100 could easily be used to reflect this.
    For the two persons above, Senate Democratic Caucus (Q3117832) mentioned by @ChristianKl: seems fairly close to the purpose of parliamentary group (P4100). There is a similar item for Republicans. @Dipsacus fullonum: what do you think? Accordingly, we could update the P4100 values to that. For actual party membership, there is a different property. --- Jura 09:20, 23 January 2021 (UTC)
Whatever you choose to do, please make sure it's consistent for all historical entries. Regards, Gettinwikiwidit (talk)
Changing it means that templates used for pages like Wikidata:WikiProject_every_politician/United_States_of_America will not work as expected. Did we ever get an answer why we think this is a change worth making? Gettinwikiwidit (talk)
@Gettinwikiwidit: parliamentary group (P4100) is by name about parliamentary groups. Senate Democratic Caucus (Q3117832) happens to be the parliamentary group to which Bernie Sanders (Q359442) belongs. Putting in valid values into statements is part of what Wikidata is about and it wouldn't be a change in our datamodel. Allowing both parliamentary group (Q848197) and political party (Q7278) in parliamentary group (P4100) produces some problems. If we want consistent modelling then it would make sense to get rid of political party (Q7278) as valid value for parliamentary group (P4100). ChristianKl ❪✉❫ 16:36, 25 January 2021 (UTC)
@ChristianKl: I don't disagree. The reason this model was chosen is explained above. Regards, Gettinwikiwidit (talk) 23:04, 25 January 2021 (UTC)
However, I do question who is going to find which caucuses senators have belonged to historically. Do you have a good resource? Or is the plan just to fill it in going forward? If this is going to be extremely spotty information I think it fair to ask if it's worth it. I.e. I think there are projects for more deserving of time. Including info about US senators. Gettinwikiwidit (talk) 23:07, 25 January 2021 (UTC)
For clarity, I will restate my request that whatever course is chosen that it be consistent across all historical entries. Regards, Gettinwikiwidit (talk) 10:04, 27 January 2021 (UTC)

No Wikidata Cleanup

Problem

I notice that Wikidata has no way of indicating that an item needs cleanup. There is no Wikidata Cleanup, categories, or templates.

Examples of things that need cleanup structures

  • Marcus Money (Q83897822) requires a lot of attention. It is an item for a music artist that has identifiers for every one of the artist's releases using YouTube video ID (P1651), Spotify track ID (P2207), and more. These releases should have their own items.
  • I've seen on some property documentation templates that the property's author will put "TODO" for constraints that they have not added or conflicts the property has with certain items they have not addressed.
  • Circularism/Broader problems: Wikidata has some general "systems/structures" for certain classes of items. For example the linking Wikiproject classes and instances of Wikidata property about Wikimedia entities (Q51118821) create a massive structure for linked categories, templates, articles, lists, and more between Wikiprojects.
    One current problem of this system is that Wikimedia list article (Q13406463) runs into some circularism issues with its has list (P2354) Wikimedia list of lists (Q33532284) statement.
    This is just one example of a circular issue, but I'm sure there are plenty of other broad issues about systems that I have not come across yet and need to be addressed and documented in a cleanup effort. Broader problems should not be limited to one entity's talk page, or only be documented on Wikidata:Project_chat, where they usually are and where they'll eventually be archived.

Current cleanup structure

Thankfully we have Template:Interwiki_conflict which puts their pages under Category:Interwiki conflicts, but that only addresses property/interwiki conflicts. That's about the only "cleanup" category we have right now.

Addressing the problem

Should I:

  • Startup the Wikiproject page Wikidata:Cleanup
  • Create a set of templates and categories for entities that need cleanup

For example, a Template:Item cleanup could be:

  • Placed at the top of an item's Talk page.
  • Gives general description that there is a problem
  • Provides parameters for certain statements of the item and their properties/values that need addressing.
  • Adds cleanup categories: Items that need cleanup, Items of (property) that need cleanup for every property parameter in the template, and items with value (value) that need cleanup for every value parameter in the template.

So... and help needed

If you think this is needed, please indicate so and I'm happy to start right away!

I'm a relatively new Wikidatian/Wikimedian so I only have a 50% grasp on creating Templates and have never ran/been part of a Wikiproject before, so I'll need some help.

--Lectrician1 (talk) 22:38, 25 January 2021 (UTC)

  •  Support We have a number of more systematic approaches with lists and WDQS queries, but something item-specific like this would be nice too, thanks! ArthurPSmith (talk) 18:41, 26 January 2021 (UTC)
  •  Oppose as described. We have constraints to mark items that need improvement. Marcus Money (Q83897822) shows a bunch of different constraint violations and is on lists for those constraint violations. I see no reason why we would need another way to classify this item. For broader problems that span multiple items, putting a tracking category on individual item, doesn't seem to be a good way to describe the problem. Having a discussion on the relevant Wikiproject is likely better. ChristianKl ❪✉❫ 12:28, 27 January 2021 (UTC)
  •  Comment No offense, but most of Wikidata is in dire need of a major cleanup, if you compare it to most Wikipedias and a lot of databases … --Emu (talk) 13:04, 27 January 2021 (UTC)

Improper images

Here Q1547571.

There are some of his works, not depicting himself. Attached as images. What should be done? Or nothing? Please, sort it out and/or answer here. Longbowman (talk) 08:26, 27 January 2021 (UTC)

They get deleted. Done. --Tagishsimon (talk) 09:15, 27 January 2021 (UTC)

Authority over coin minting

I raised this issue in the Numismatics Project page but was advised to bring here for broader discussion. I was wondering about the best way to model authorities in WD. For example, if you look at this coin Tetradrachm of Caracalla, Roman Emperor, from Cyrrhus, Yale University Art Gallery, inv. 1938.6000.56 (Q100477694), the coin was produced in the regnal years of Caracalla, under the authority of Caracalla. See here for more info. What would be the best way to model this? depicts (P180) is not always a good property in these scenarios because the emperor is not always depicted on the coins produced in their era and commissioned by (P88) is perhaps too strong a claim (the emperor may have oversight of use of his imperial image and the minting of coins but the processes of commission can vary dramatically in different parts of the empire with local boards responsible for minting). I see that one user has used authority (P797) with this coin Caliph standing (coin face) (Q66372570) for a similar situation. There was some ambivalence about the use of this project for coins. I was wondering if it might be suitable to create a property like "regnal period of" or something which could also be applied. What do people think would be the best way to model? Valeriummaximum (talk) 16:56, 25 January 2021 (UTC)

@Valeriummaximum: The British Museum uses role terms like Issuer, Moneyer etc: see https://confluence.ontotext.com/display/ResearchSpace/BM+Association+Mapping. So I agree that a new prop is needed. Call it eg "minting authority" or "moneying authority" and not "regnal period of" --Vladimir Alexiev (talk) 00:36, 27 January 2021 (UTC)

@Vladimir Alexiev: I think the issue is that it is not always clear whether the emperor is the issuer or minting authority in the ancient world. Many local cities had their own boards or officials for minting coins and the emperor's authority over the coins was less direct. I notice that the British Museum has a field for "authority ruler" [ https://www.britishmuseum.org/collection/object/C_BNK-R-26 see here], so I am wondering about a property like regnal period or authority that is broad and vague and doesn't make too substantive claims about the role of the emperor in issuing/minting the coin. Valeriummaximum (talk) 22:02, 27 January 2021 (UTC)

Moving Wikimania 2021 to a Virtual Event

Wikimania's logo.

Hello. Apologies if you are not reading this message in your native language. Please help translate to your language. Thank you!

Wikimania will be a virtual event this year, and hosted by a wide group of community members. Whenever the next in-person large gathering is possible again, the ESEAP Core Organizing Team will be in charge of it. Stay tuned for more information about how you can get involved in the planning process and other aspects of the event. Please read the longer version of this announcement on wikimedia-l.

ESEAP Core Organizing Team, Wikimania Steering Committee, Wikimedia Foundation Events Team, 15:15, 27 January 2021 (UTC)

What qualifier can I add to someone born after the death of the father

What qualifier can I add to someone born after the death of the father to qualify the "contemporary constraint" flag. I want to create a Qid called "born after the father's death", but what property should it be attached to as a qualifier? I need something like "cause=born after the father's death". Any ideas? This way every I see it at Johan Christopher Ruuth I (Q50346736) I will not have to look to see if I made an error. --RAN (talk) 19:11, 26 January 2021 (UTC)

Project Grant Open Call

This is the announcement for the Project Grants program open call that started on January 11, with the submission deadline of February 10, 2021.
This first open call will be focussed on Community Organizing proposals. A second open call focused on research and software proposals is scheduled from February 15 with a submission deadline of March 16, 2021.

For the Round 1 open call, we invite you to propose grant applications that fall under community development and organizing (offline and online) categories. Project Grant funds are available to support individuals, groups, and organizations to implement new experiments and proven ideas, from organizing a better process on your wiki, coordinating a campaign or editathon series to providing other support for community building. We offer the following resources to help you plan your project and complete a grant proposal:

Program officers are also available to offer individualized proposal support upon request. Contact us if you would like feedback or more information.

We are excited to see your grant ideas that will support our community and make an impact on the future of Wikimedia projects. Put your idea into motion, and submit your proposal by February 10, 2021!

Please feel free to get in touch with questions about getting started with your grant application, or about serving on the Project Grants Committee. Contact us at projectgrants@wikimedia.org. Please help us translate this message to your local language. MediaWiki message delivery (talk) 08:00, 28 January 2021 (UTC)

New online event: Wikidata bug triage hour

Hello all,

As part of the improvements on the support process, we would like to find ways to involve better the community in the development of Wikidata. This is why, on top of the Wikidata and Wikibase office hour, we are experimenting with a new format: the Wikidata bug triage hour. It's an online event where Lydia, the product manager of Wikidata, publicly works on triaging development tasks (typically on Phabricator), improving their descriptions, defining their priority, and collecting the wishes and needs from the participants.

The first session will take place on February 16th and will be about Lexicographical Data. Join us and bring your favorite Phabricator task! You can find more details here, feel free to reach out to me if you have any questions. We're looking forward to chat with some of you there!

Cheers, Lea Lacroix (WMDE) (talk) 10:06, 28 January 2021 (UTC)

Duplicates, to merge

Looks like Mro (Q75919253) and Mro (Q104837502) are duplicates. Their ISO 15924 numeric code (P2620) (should be unique) are equal. @Pmepepnoute, Nikki: -DePiep (talk) 12:00, 28 January 2021 (UTC)

Aspect of history vs. list

Hi all. If history of the bicycle (Q2059831) was instance of (P31)=Wikimedia list article (Q13406463), then I wouldn't hesitate to use category related to list (P1754) to link it to Category:Bicycles (Q7025398) (and list related to category (P1753) to link back again). However, it is instance of (P31)=aspect of history (Q17524420). I have quite a few examples of this, as it's important for linking 'History of' articles to the relevant Commons category on enwp, but before I try to sort them out systematically, I'd like some input. Would it make sense to use category related to list (P1754) anyway, or do we need a new pair of properties for these links, or something else? Thanks. Mike Peel (talk) 15:48, 25 January 2021 (UTC)

@Mike Peel: We have history of topic (P2184) for the topic → history direction, and I think facet of (P1269) works pretty well for the history → topic direction. I don't know if we need a new property. Vahurzpu (talk) 16:09, 25 January 2021 (UTC)
Thanks, but I'm particularly interested in the history → category aspect of this (linking enwp to commons via wikidata). Thanks. Mike Peel (talk) 16:13, 25 January 2021 (UTC)

Horse-drawn trams or mule-drawn trams?

We do not have a data item specific for mule trams. Do we create one or broaden the definition of Q832003? There is a Commons category: 'Mule-drawn trams'.Smiley.toerist (talk) 12:48, 28 January 2021 (UTC)

I say create a new one. Rarely should we expand a definition. BrokenSegue (talk) 13:02, 28 January 2021 (UTC)
They are in scope at en:Horsecar: "an animal-powered (usually horse)", with some mentions of mules. Ghouston (talk) 21:30, 28 January 2021 (UTC)

Why does a WD search not find Category:Nobel laureates

I know this page exists here as Category:Nobel laureates (Q6635159), but when I do a search on "Category:Nobel laureates" the search comes back empty handed. What am I missing? Thanks in advance, sorry cannot sign my keybard seems broken ̴

Wrap it in quotes, if using the full text search. At a wild guess Category: is a reserved word on cirrussearch. Escaping the colon with a forwardslash also seems to work. --Tagishsimon (talk) 23:37, 27 January 2021 (UTC)
Thank you, that did it. Now comes my real question: Is there a way to find a category that I suspect exists on WD, but I don't know exactly what it is called. Example: "Category: Franz Kafka award winners" Ottawahitech (talk) 23:46, 27 January 2021 (UTC)
Shame about your keyboard. Seems like a tilde is your friend. Your example doesn't work so well since there isn't a cat for the Franz Kafka Prize. If we take instead Category:Prix Louis Guilloux winners (Q26973369) a seach for Category Prid~ Louis Guilloux seems to work ... tidle is a fuzzy search modifier. More details at Help:CirrusSearch --Tagishsimon (talk) 00:35, 28 January 2021 (UTC)
Yes it appears that there is no Category:Nobel laureates on WD. But Category:Nobel laureates (Q6635159) does exist? Ottawahitech (talk) 16:46, 28 January 2021 (UTC)
That is an item (or page in namespace "0") of Wikidata, not a "category". Compare with a search for Category:WikiProjects (which exists as a category at Wikidata). But, explained by Tagishsimon, it's a feature of CirrusSearch to prefer one over the other. Personally, I think we could do away with it. Maybe Wikidata:Contact_the_development_team/Query_Service_and_search can help with it. --- Jura 16:54, 28 January 2021 (UTC)
I can do a successful search on enwiki using the search term "Category:Nobel laureates". Actually I am presented with pages and pages of search results. I guess WD uses a different search engine? Ottawahitech (talk) 17:04, 28 January 2021 (UTC)
Same search engine. Category: has meaning on wikipedia, since that has categories. Wikidata has no categories, but the search engine still recgonises Category: as meaning categories, looks for same on wikidata, does not find them, hence the result. Wikidata has items that have information about category pages on wikipedias ... that is not the same as having categories. --Tagishsimon (talk) 17:20, 28 January 2021 (UTC)
I suspect what the OP really wanted was Q6635159#P3921 --SCIdude (talk) 08:44, 29 January 2021 (UTC)

Recommend another property for conflicts with property constraint

This was copiedish from Discord for convenience and where it was originally asked.

I've come across a lot of properties that I think should have Help:Property_constraints_portal/Conflicts_with but what I want to have is the flag that pops up when you use this property with an item to recommend using another but similar property.

For example, an item government agency has the statement: part of: government

Let's say we have this statement in part of though: 

property constraint: conflicts with constraint 
  property: subclass of 
  item of property constraint: organization

And because government is a subclass of: organization, it gets a flag:

part of should not be used on an item of subclass organization


But what I want the flag to say is:

part of should not be used on a item of subclass organization. member of should be used instead

So the original constraint statement should look something like this: 

property constraint: conflicts with constraint
  property: subclass of
  item of property constraint: organization
  replacement property: member of


Soooo I'm pretty sure this system doesn't exist.

To make it exist:

  • Should I make a discussion on the constraint page?
  • Could this "replacement" apply to other constraints (values)?
  • Should I make a Phabricator task?

--Lectrician1 (talk) 05:42, 29 January 2021 (UTC)

Fighting link rot

We have lots of link rot on wikidata. I think it would make sense to periodically try to hit those URLs and make sure they at least still resolve and don't return an error. The issue is what would we do with this information? Problems:

  • A link being dead now doesn't mean it won't be alive later (intermittent failure)
  • Links can be live but no longer related to the item
  • We shouldn't add end time (P582) unless we are relatively sure it's permanently dead

I'm thinking we could add a new qualifier indicating that a link failed at some point. Then we could wait for say three failures across a few months before marking it with something like end time (P582). Does this seem reasonable or are there better ideas for dealing with link rot. I'm imagining such a bot could also ensure the link is archived for when it does go dead (I think enwiki has efforts in that direction). BrokenSegue (talk) 17:56, 19 January 2021 (UTC)

All URLs used on wikidata are archived on the Wayback Machine, according to its keeper. --Tagishsimon (talk) 20:51, 19 January 2021 (UTC)
@Tagishsimon: hm, do you have information on how that works? links for external identifiers are not being backed up reliably and it would good if they were. BrokenSegue (talk) 23:26, 19 January 2021 (UTC)
Don't know. Have asked. --Tagishsimon (talk) 23:48, 19 January 2021 (UTC)
Links on external-ids are being generated in the web UI by javascript, I think. They are not written down hardcoded anywhere, including the externallinks table. I'm pretty sure they only look at what's in that table. —MisterSynergy (talk) 23:52, 19 January 2021 (UTC)
Answer I get back is: "yes". --Tagishsimon (talk) 00:07, 20 January 2021 (UTC)
Thanks. I'm still sure "no" is what actually happens, based on how these links are being generated and the observation that plenty of identifiers, added at various periods of time to Wikidata, are not archived by the Internet Archive. —MisterSynergy (talk) 01:04, 20 January 2021 (UTC)
I'm also pretty sure identifiers links are currently not archived. That's why I made a proposal at Community Wishlist Survey which ends in the top ten and should be implemented this year. Ayack (talk) 14:05, 22 January 2021 (UTC)
Internet Archive uses EventStream, not SQL tables. See http://blog.archive.org/2018/10/01/more-than-9-million-broken-links-on-wikipedia-are-now-rescued/ --Snaevar (talk) 00:11, 20 January 2021 (UTC)
The last time I looked into the issue, there are a few tens of thousands of links on Wikidata that got backuped by the bot from the Wayback machine but the amount of those is relatively small. ChristianKl ❪✉❫ 08:36, 20 January 2021 (UTC)
@Snaevar: looking at the stream entries when I edit an item I'm not seeing the property's URL. Is there a link to the code that does this? Is it owned by the archive or wikimedia? BrokenSegue (talk) 00:31, 20 January 2021 (UTC) And looking at some of my few day old edits to properties I'm not seeing them in the waybackmachine archive (though I have no idea how long it would take to show up). BrokenSegue (talk) 00:33, 20 January 2021 (UTC)
What other ways might these types of URLs be detectable by automated processes? -- GreenC (talk) 15:51, 20 January 2021 (UTC)
@GreenC: not sure what you mean? what type of URLs? non-archived ones? BrokenSegue (talk) 13:52, 22 January 2021 (UTC)
@BrokenSegue: Yes, Internet Archive monitors Events Streams API. When it sees (via Events Stream) a new URL was just added somewhere on a WMF project, it triggers this URL to be saved at the Wayback Machine. In this way, every (or most) URLs added to a WMF site also have an archive available (on the Wayback Machine). If the links discussed above are not visible via Events Stream, what other way could Internet Archive monitor for their addition on Wikidata, such that they can automatically be saved into the Wayback Machine? -- GreenC (talk) 14:01, 22 January 2021 (UTC)
@GreenC: They could periodically pull the list of external id properties, look for additions of them and automatically fetch the URL using the associated formatter URL. BrokenSegue (talk) 14:05, 22 January 2021 (UTC)
(BTW I work for IA but not directly on this project). How do you pull the list of external id properties? Any more detailed information would help. -- GreenC (talk) 14:09, 22 January 2021 (UTC)
Sounds like something that would need a custom export possibly built on what is available for WQS replication or Wikipedia updates. Wikidata:Contact the development team might help. --- Jura 15:20, 22 January 2021 (UTC)
@GreenC: sorry only just saw this comment. I'd be glad to help you directly (message me on my talk page) if you want/need more info to investigate this. BrokenSegue (talk) 20:54, 29 January 2021 (UTC)

Literary award dates

Some literary awards, such as Nebula Award (Q194285), are presented for books published in a particular year, and the winner is announced in the following year. I assume that when point in time (P585) is used to give the date of the award, it should be either the date when the winner is announced, or the date when the award is presented. However, in many cases, the year of publication has been used instead. What should it be? See for example, [15], where the "2019 Nebula awards" finalists were announced in 2020. Ghouston (talk) 10:36, 28 January 2021 (UTC)

So you can say for example that Sarah Pinsker (Q28870552) won the 2019 Nebula Award for Best Novel, and you can also say that she won the Nebula Award for Best Novel in 2020. The item Sarah Pinsker (Q28870552) has point in time (P585) set to 2019, but setting it to the date when the award was announced, in 2020, seems more consistent with the way other awards are treated. Ghouston (talk) 10:50, 28 January 2021 (UTC)

  • The date of the award seems to be 2019, even if it was announced the year before, during the year, or the year afterwards. The actual date of the ceremony or publication could be captured differently. For some, the year is notoriously different, for some it's the same, for others that sometimes changes. --- Jura 12:55, 28 January 2021 (UTC)
  • I would agree with Jura - go with whatever point in time the award presents itself as. If it is called the "2020 Award for X" it's point-in-time 2020, whether it's "announced in 2020 based on work from the last three years" or "for work published in 2020 but not announced until 2021". I don't think we should require consistency here, as it will inevitably mean we get out of sync with how some prizes describe themselves - eg the Nebula describes itself by publication year, but the Booker describes itself by award year, so a one-size-fits-all approach would get confusing. Andrew Gray (talk) 21:13, 28 January 2021 (UTC)

Classes

Could you tell me if there are any classes on the South Shore of Massachusetts?

Invalid indirect subclasses

Is there some validation tool to protect against adding invalid indirect sublasses?

For example National Historical Park (type of U.S. protected area) [16] should be never an - even indirect - subclass of behavior.

Right now it is, via following chain:

National Historical Park (type of U.S. protected area) [17]
heritage site in the USA [18]
heritage site (general term for a site of cultural heritage for a specific country (please avoid as a P31 value except for large sites containing multiple entities)) [19]
cultural heritage (physical artifact or intangible attribute of a society inherited from past generations) [20]
heritage designation (act or effect of recognizing the historical, artistic or cultural value of a property) [21] (I removed this as cultural heritage is not a sublass of heritage designation)
heritage (property, custom, object inherited from previous generations and of importance for future generations) [22]
preservation (set of activities aimed at prolonging the life of a record or object) [23]
heritage (cultural values and traditions inherited from one's ancestors) [24]
transfer (granting subject with ownership or communality) [25]
move [26]
behavior (way that one acts in different situations) [27]

Is there some way to list such invalid sublasses so that Wikidata editors would be warned about them? I am trying to use Wikidata and made mistake of reimporting my data. It turns out that multiple subclasses are again broken. Is there some way to prevent this? Mateusz Konieczny (talk) 19:02, 29 January 2021 (UTC)

@Mateusz Konieczny: When is a subclass invalid, i.e. where is the point of defining the conflict? (We want classes of behavior, just not this one.) Warnings are only created through property constraints, and we definitely don't have a conflict with value type constraint. If it existed, it could in this case fire on a heritage identifier when the item is a behavior. If you add a Template:Complex constraint on a talk page of a property a complex constraint report will be created though. I would support a new constraint type for this, if this is what would help here. --CamelCaseNick (talk) 19:50, 29 January 2021 (UTC)

Geolocation of Europe

The article for Europe, Q46, features a "geographic location" before listing the geographic extrema. The location entered here is on the swiss border to france. I tried to understand the rationale of this location but none of the reference links told me more about it's calculation method or source. I understand that it must differ from the geographic centre of the EU. Still it seems far too much in the south-west. How can i find out the source and correct the data, please?

Blausand (talk) 02:03, 30 January 2021 (UTC)

Help with a warning tag

I asked this before but cannot find the answer. At Hans Bernhard Schwerin (Q59164080) at "Commons Creator page" I am getting a warning that "Commons link should exist." Which on the 4 possible commons links is it looking for and is it really necessary to have the warning? If we must have the warning, can the message be written to be less vague? --RAN (talk) 17:58, 29 January 2021 (UTC)

I was bold and removed the constraint myself, even the examples on the property page had the error message, so the creator wasn't even sure why it was added. --RAN (talk) 03:30, 30 January 2021 (UTC)
RAN: I reverted your change, as the constraint you removed isn't the one causing the issue (it's hard to see because property constraints are confusing, but the one you modified just saying that Commons pages should have a Commons category (P373) statement, whereas the problematic constraint is about whether a page exists on Commons). The constraint generating the error message is Property:P1472#P1472$ff3a4195-471d-87e1-a601-1a9cf7d948c9, which is correctly specified, thought the implementation is buggy (see phab:T237920 for details). Vahurzpu (talk) 07:47, 30 January 2021 (UTC)
Thanks! I realized the warning tag was still there, but wanted to wait till today to restore it, in case it was a caching issue. --RAN (talk) 17:48, 30 January 2021 (UTC)

subscriber (Q10638511) is the subject item of number of subscribers (P3744). Is this correct in regards to YouTube subscribers?

There is no English Wikipedia article about subscriber (Q10638511). Is this item about the same kind of "subscriber" that a "YouTube subscriber" is about or do we need a new item for that? What I read ie. in the Swedish Wikipedia article, my interpretation is that it looks like this is only about paying in advance for a subscription, although on YouTube to subscribe to somebody you don't need to pay anyone, you just click a red colored button with the word "SUBSCRIBE" on it all in capital letters and it works. After you've clicked you are then subscribed to a user, no big deal. LotsofTheories (talk) 09:08, 30 January 2021 (UTC)

Question about what to do

Hi. There is a magazine called Revista Marítima Brasileira Q20050031, published by the Brazilian Navy. I want to make a citation of one of their publications, a version from 2009. How do I link that publication, that I will create a QID, with QQ20050031 ? Tetizeraz (talk) 15:54, 30 January 2021 (UTC)

where and in what form do you want this citation? BrokenSegue (talk) 15:56, 30 January 2021 (UTC)
Sorry for not being clear BrokenSegue. I want to create a QID of one of their publications and use Cite Q, in Portuguese Wikipedia, just like I have been doing with QIDs like Q105047567 and Q105047566. It's not about how to use citations on Wikipedia, that I know how to do, but what I should do on Wikidata. Tetizeraz (talk) 16:03, 30 January 2021 (UTC)
By the way, I would add this link as a future QID. Tetizeraz (talk) 16:03, 30 January 2021 (UTC)
Create an item and follow the pattern at e.g. The Atlantic Monthly Number 108 (Q38060062) --Tagishsimon (talk) 16:27, 30 January 2021 (UTC)
Thanks, Tagishsimon. I'll do that. Tetizeraz (talk) 18:04, 30 January 2021 (UTC)

Changing the fallback languages

I've been trying to change the fallback languages for labels, but seem to only be able to do it on the item pages and in the watchlist. Is there any way to change them everywhere and not just in some places? --Sabelöga (talk) 21:40, 30 January 2021 (UTC)

Heavy problem in Q9068 (Voltaire)

Some 250 or even more years ago, some nasty and medieval cleric made a smear campaign by naming Voltaire a "demon" - one of the greatest philosophers of all time. Today, Wikidata features and - much worse - conservates forever this barbaric smear action and stupidity in Voltaire's wikidata item [text: incarnation of: demon"], shamefully masking it with the (hardly findable) attribute "refuted for superstition" or similar. I made a long statement there on the talk page, but was not heard yet. I am a bit tired to discuss about something which I find hard to discuss about at all in the first place, so this will be my last edit concerning this matter. May the enlightenment prevail ;-) Pittigrilli (talk) 18:53, 29 January 2021 (UTC)

Wondering where you stand on the Schleswig-Holstein Question. --Tagishsimon (talk) 15:17, 30 January 2021 (UTC)
I am sorry guys, but sometimes one can be blind. That is you. What is the purpose/the use case/the benefit of writing a piece of (ahem) information which was even refuted at the time that it was made by thinking people in a today's database on facts? If your answer is sth like "historically interesting", then I answer: Maybe. But then it should in the Wikipedia articles, and not in Wikidata. The fact stands there broad and bold, and then there is a little addition "Oh yeah, by the way, this is wrong." The Wikidata is the wrong place for a discussion of superstition and prejudices of medieval quality in about 150 characters. That is at least confusing to most people seeing this. But do as you like. btw, I am sure Martin Luther was also called the devil or demon by some catholic cleric - why don't go search and put in Wikidata? Have fun. Pittigrilli (talk) 16:08, 30 January 2021 (UTC)
You miss the point. Wikidata is not the place to arbitrate the accusation that was made, but to record the fact of it being made. As Jura points out, the claim is deprecated, which is appropriate handling for an incorrect claim which nevertheless was made. --Tagishsimon (talk) 16:24, 30 January 2021 (UTC)
Wikidata isn't just a database of facts. It's also a database of claims made by various sources. The particular book that's the source Dictionnaire Infernal is notable enough to have a Wikipedia article in 16 different languages. ChristianKl ❪✉❫ 16:25, 30 January 2021 (UTC)
If you think it should be in Wikipedia, I don't see why it shouldn't be here. Usually, the question is different: should Wikipedia include some (possibly fringe) view in their neutral point of view of today. And then endless Wikipedia discussions follow. Wikidata just notes it and then moves on. --- Jura 12:22, 31 January 2021 (UTC)
Ok, as all of you are wrong (joke) - no, ok, I see the point. I agree to disagree and accept the majority view. Thanks for the discussion, Pittigrilli (talk) 17:08, 31 January 2021 (UTC)

Proposal

Moved from my talk page:Smiley.toerist (talk) 10:11, 31 January 2021 (UTC)

Hi. Any chance you'd be interested in doing a proposal for "location canceled" to go along with the date canceled property? Thanks :) --Adamant1 (talk) 22:34, 26 January 2021 (UTC)

Can you give an concrete example?Smiley.toerist (talk) 23:38, 26 January 2021 (UTC)
Like with File:1898 circa Zander & Labisch, Berlin, AK Grußkarte Ewige Lampe, Hannover, Inhaber Julius Bockhold, Adressseite.jpg how it gives Hannover for the location inside the circular cancel above the date. --Adamant1 (talk) 05:01, 27 January 2021 (UTC)
It is unclear. By 'date canceled' there is always a location implicit. In this example you have two dates (send date / from post office posted) and the (arrived at) at the post office of the destination. Both are cancel dates. Maybe it is usefull to allow for two values for 'canceled property'. I would not needlesly complicate things by adding a new property. After all it nearly always only used for dating purposes of the postcard. High precision is not needed.Smiley.toerist (talk) 13:53, 27 January 2021 (UTC)
You don't think it would be useful to know what that location is though? If there could be two values for the 'canceled property' that would be helpful to. I think it's helpful to know where the postcard or envelope is sent from or goes to though. A postcard or envelope sent from say Nazi Germany during WW2 is completely different then one sent from Argentina during the same time. Just knowing that it was sent in 1943 doesn't help with that though. Or for instance mail sent by balloon from Przemyśl, Poland during World War 1. "1915" doesn't convey anything in that case. What about "mailed from" and "mailed to" properties instead? Although they would not allow for intermittent stops and I think there is value in confining it to what is inside the cancels. --Adamant1 (talk) 00:37, 31 January 2021 (UTC)
I think this proposal can best be moved to 'project chat', to see what for solutions there are. It is premature to start a concrete property proposal. By the way, I am new to this to, this was my first proposal in Wikidata.Smiley.toerist (talk) 10:08, 31 January 2021 (UTC)
I have added to the Structured Data in the example File:1898 circa Zander & Labisch, Berlin, AK Grußkarte Ewige Lampe, Hannover, Inhaber Julius Bockhold, Adressseite.jpg. It is posible to add a location to Q1194406 but not a date. There are several warnings.Smiley.toerist (talk) 17:36, 31 January 2021 (UTC)

Qualifier to use for two official websites

Does it make sense to use the qualifier applies to part (P518) with the Qid for a university department when a professor has academic appointments in more than one department and therefore has two or more official websites? I'm looking for the best way to overcome the constraint violation on official websites. For a real life example: Phillip Thurtle (Q105086473). Have a look at the official website statement in that item. https://history.washington.edu/people/phillip-thurtle is his website in the UW Department of History and https://chid.washington.edu/people/phillip-thurtle is his website in the UW Department of Comparative History of Ideas. Does "applies to part" work as a qualifier, or is there something better that I should use? UWashPrincipalCataloger (talk) 21:52, 30 January 2021 (UTC)

@UWashPrincipalCataloger: I think applies to part (P518) is almost exclusively used as a qualifier. And yeah I think your examples are reasonable ways to include multiple official websites. You could also consider applies to jurisdiction (P1001) or intended public (P2360) or publisher (P123). BrokenSegue (talk) 22:25, 30 January 2021 (UTC)
Thanks @BrokenSegue:. The description of applies to jurisdiction (P1001) implies that this is only used for political jurisdictions. Is it really ok to use it with any kind of organization? Right now it seems to be limited to geographic places/political jurisdictions. All of the examples in the property are for places. UWashPrincipalCataloger (talk) 00:43, 1 February 2021 (UTC)
@UWashPrincipalCataloger: yeah I think it's only ok for jurisdictions. It's probably not the right choice but none of them really fit perfectly. Probably publisher is the closest. Or you can just include multiple websites. BrokenSegue (talk) 00:58, 1 February 2021 (UTC)
What we really need is just a qualifier property for "applies to". That would be generic and perfect for many situations. UWashPrincipalCataloger (talk) 01:01, 1 February 2021 (UTC)
@UWashPrincipalCataloger: Perhaps, of (P642)? BrokenSegue (talk) 01:32, 1 February 2021 (UTC)
Yes, I ended up changing my qualifier to that. It already had the alias "applies to". UWashPrincipalCataloger (talk) 07:02, 1 February 2021 (UTC)

Showcase Queries

Some time ago I began a discussion about the idea of creating a new community award/recognition - for "Showcase Queries". This would be process to identify, praise, and share to the world Wikidata's true superpower: the combination of statements from many different items – the ability uncover knowledge spread across the database.

People seemed generally supportive of the concept so I wrote a proposed description and criteria for this award at Wikidata:Showcase Queries.

Various people have given feedback on the proposal on the talkpage but, ultimately, it's an abstract idea that needs real-world testing to turn from a nice idea into a community process. And so, User:Joalpe and User:Ederporto have been kind to make a formal proposal - using the draft criteria - as a request to be recognised as the first "showcase query"! You can see their nomination here: Wikidata:Showcase Queries/List of people killed by and disappeared during the Brazilian military dictatorship.

This is, therefore, a "test-case" for the proposed community recognition. A practical example for everyone to discuss, whether:

  1. the proposed award is a good idea,
  2. the criteria are appropriate, and
  3. this specific nomination meets those criteria.

If there is a consensus to support all three of these points, we could recognise the first 'official' showcase query and then create a proper nomination template etc for future nominations.
I do not think there is any official process we have to make a proposal like this, other than to test it. So, I ask you all to please comment on the talkpage of the award proposal, and on the specific nomination page. I am open to suggestions for how people would recommend moving this suggestion from idea to reality (e.g. should I publish this as a Wikidata:Requests for comment)? Wittylama (talk) 11:59, 31 January 2021 (UTC)

  • Maybe you could give a few sample queries you think meet the criteria. Supposedly you have had some in mind when writing this. --- Jura 12:16, 31 January 2021 (UTC)
    • Please see the query and associated nomination document that I provided in the above message. Wittylama (talk) 12:41, 31 January 2021 (UTC)
      • I meant queries you had in mind when writing the "showcase query" page. The nomination seems to be posterior. --- Jura 12:58, 31 January 2021 (UTC)
        • I did not have a list of pre-existing queries which I wanted to get recognition for. I assume many people have a query they’re particularly fond of, that they’ve put a lot of effort into curating, that they would like to nominate. And, at least at the outset while there is no precedents, it would be a debate between whether the criteria should adjust to suit the nomination, or the nomination should adjust to suit the criteria. Hence this first test-case nomination - it’s an example where we ought to debate the appropriateness/applicability of both the query and the criteria. Wittylama (talk) 13:08, 31 January 2021 (UTC)
          • I didn't ask if you had queries you wanted to get recognition for, but if there were some you had in mind when writing it. Maybe they only partially fulfill the criteria you wrote down. Supposedly, you didn't write it without any basis in actually existing or working queries. --- Jura 13:13, 31 January 2021 (UTC)
  • @Wittylama: I like this idea, and the specific proposal seems great. However, I'm wondering if a showcase query should have such specific output - in this case it looks like it's specially formatted for input into a Portuguese Wikipedia page. Maybe output should be a little more generic or flexible for showcase queries? Other than that, I agree it's an excellent well-documented example query that should meet "showcase" level criteria. ArthurPSmith (talk) 18:56, 1 February 2021 (UTC)
    • thanks for the support ArthurPSmith. The question you raise is exactly the kind of thing that I would hope the community should determine over the course of reviewing nominations over the years. As has occurred in English Wikipedia featured article discussions (and probably other languages too) the criteria tend to become more stringently applied over the years... Wittylama (talk) 18:02, 2 February 2021 (UTC)

2 Questions

I am a newbie in the Wikidata project and I just discovered the potential of Wikidata a few days ago when I was creating the still missing articles about the japanese nengo in the german Wikipedia project.

Unfortunately I started adding Wikidata information (label and description) to the nengo articles before I read the Wikidata manuals and I guess now that I made a mistake. The description I added looks like "Japanische Ära year bis year". I learned that 1. the description label should start with a small character and 2. that the year should not be part of the description label. Should I change the description manually or can this be done automatically? It's around 100 articles so far.

Since several years I am collecting data about Japanese writers. The list is far from beeing perfect but it consists of around 1000 entries and contains the year of completion / publishing, the name of the author (Romaji and Kanji), the title of the work (Romaji and Kanji and translation if available), small description of the work (content, No. of volumes). The entries range from the very first work, the Kojiki, until 2007. It is based on a list of a Japanese literature dictionary. In case this list is useful for wikipedia, is it possible to import the data and use it like the sample timeline done in the Historiopedia? -- Elmo rainy day (talk) 12:39, 31 January 2021 (UTC)

Wikidata:Forum might be the better place to ask about the best way to write descriptions in German. Use or not of caps is language specific. Help:Description/de can differ from Help:Description.
For time periods, I don't really see a problem with the inclusion of years. --- Jura 13:02, 31 January 2021 (UTC)
Hello & Welcome. The minor mistake regarding capitalisation is rather easy to make in this case, because the English descriptions all start with with a capitalised "Japanese"–nationalities being among the few categories that are capitalised in English but not German.
Anyway, I took the liberty to fix the capitalisation and also used your descriptions to add missing start time (P580) and end time (P582) statements. Using the dates, I linked consecutive eras with follows (P155) and followed by (P156). There's a tool called Quckstatements for mass-editing that allowed doing this in just a few minutes. But don't despair if that looks complicated–the manual part you've already been doing is the more valuable work exactly because it cannot be automated.
Anyway, this is a query showing all eras, sorted by start date, with their respective links to previous and next era: Query. It looks fairly complete, with a few missed connections where dates didn't quite agree.
Regarding the data you've collected: well, that's one of those "more complicated" cases. There is a tool called OpenRefine that allows doing such imports using a graphical user interface. It's pretty cool and only slightly buggy and it helps with the main difficulty, which is matching the data you have to data already existing in Wikidata. Have a look and see if that might work for you. If not, feel free to post on my talk page and we'll figure something out. Matthias Winkelmann (talk) 23:30, 31 January 2021 (UTC)
thanks for your detailed answers and for fixing the issue regarding capitalisation. I think the quickstatement and the SPRQL sourcecode is pretty well understandable and a good sample to dive into the matter. I already started to check the OpenRefine tool. This will take a few days, but I will come back to you for sure. Much appreciated -- Elmo rainy day (talk) 20:41, 2 February 2021 (UTC)

Internal links to property anchors

It seems I keep seeing links like

  • [[Property:P279#P31]]

look like

Something that generally only happens in edit summaries. Tests:

Is it just me or some change in the code? Tests here seem ok. --- Jura 11:00, 30 January 2021 (UTC)

In the usual page view, older version of the page and in edit preview, I can see what the link is supposed to be.
When I display a diff of this page, I can see something like what edit summary or {{P}} would produce. Maybe it's a bug? --Matěj Suchánek (talk) 09:33, 3 February 2021 (UTC)

New tool: Ranker

You can edit the rank of the following flag image statements on Mississippi: (buttons to select all/none/invert) Flag of Mississippi.svg (preferred rank), start time: 11 January 2021; Flag of Mississippi (2001–2020).svg, start time: 7 February 1894, end time: 390 June 2020; no value (deprecated rank), start time: 1 July 2020, end time: 10 January 2021; Flag of Mississippi (1861-1865).svg, start time: 1861, end time: 1865; edit summary (optional): [blank input]; buttons to set to preferred/normal/deprecated rank or increment rank
screenshot of the tool

Hi all! This is just a quick note that I wrote a new tool, called Ranker (documentation), which lets you edit the rank of multiple statements at once. You can find a usage example in this tweet or this video, or just look at this edit as the result :) cheers! --Lucas Werkmeister (talk) 17:19, 30 January 2021 (UTC)

Hello -Lucas, thanks for your work on this! I think it's high time we had some nice tools for editing ranks. This is really useful, although I often want to mass-edit statements in more than one item. It would be great if I could load more statements using a SPARQL query, such as this one:
select ?item ?id ?statement where {

  ?item wdt:P762 ?id .
  ?item p:P1435 ?statement .
  ?statement ps:P1435 wd:Q385405 ; wikibase:rank ?rank filter(?rank != wikibase:DeprecatedRank) .  

} limit 10
Try it!

... and the tool would load all statements displayed in the ?statement variable and enable me to change their rank. Do you think this would be feasible? Vojtěch Dostál (talk) 15:35, 3 February 2021 (UTC)

Problem with items for 'USB Flash Drive' and 'USB mass storage device' - how to proceed?

I put a lot of work today in USB mass storage device and USB Flash Drive, which were some hours ago (imho) both strongly improvable (irony). They were partly (from a technical point of view) mixed with each other, partly wrong, partly the english description wrong AND contrary to the German description or vice versa.

However, I arrived at a dead end for my Wikidata skills - I realized only now that, e.g., the identificators in the "mass storage" item are (nearly) all for the USB flash drive, the interwiki links as well.

I already put in a lot more work than I intended, and I have the feeling that completion would take me alone some more hours and would probably lead to some errors due to my lack of experience with the rest of the tasks tbd. Hence my suggestion - either I redo all I did completely (see for yourself how the state was before I started), or some other contributors here help me out. I am fine with either version. Please advise. Pittigrilli (talk) 17:26, 31 January 2021 (UTC)

I agree. I originally thought it would be a good idea to have a generic item for "USB mass storage device", but when one looks at the taxonomy and inner logic of Wikidata, it is just a mix-up and redundant. The "class" item USB mass storage device class (Q1155870) is a perfect higher class for all other, inluding USB flash drive, USB hard disc, USB card reader, etc. Pittigrilli (talk) 10:18, 1 February 2021 (UTC)
As I do not see anyone disagreeing, I would say we follow your proposed course. The only bigger task I see is that nearly all identificators and interwikis (at least 15 each) in the to-be defunct USB mass storage device are actually wrong where they are now, because they all correctly belong to USB Flash Drive, which now has almost none. How do we solve that elegantly? Merge both items? To be honest, your proposal above exceeded my Wikidata skills. Pittigrilli (talk) 18:39, 1 February 2021 (UTC)
I'll look into it. I think a lot of the site links can be moved from USB storage conflation (Q4432) to USB flash drive (Q1647694), for a start. Ghouston (talk) 01:28, 2 February 2021 (UTC)
Looks very good. I made some more minor edits in USB flash drive (Q1647694), I think it is now satisfactorily. I will create an item for USB Hard Disk soon. Thanks for the good cooperation, Pittigrilli (talk) 16:55, 2 February 2021 (UTC)
Depends if the type of device, and not the interface, is what is needed in a statement somewhere. Peter James (talk) 17:14, 2 February 2021 (UTC)
The Deed Is Done (Q7729535) (joke): Here is the new USB hard disk (Q105232187). The status is 'in progress'. Pittigrilli (talk) 17:59, 2 February 2021 (UTC)
Your category c:Category:USB hard disks is a duplicate of c:Category:USB hard disk drives, isn't it? Ghouston (talk) 21:22, 2 February 2021 (UTC)
No. It appears at first sight like that, but no. When I created yesterday the new category c:Category:USB hard disks, I first looked around and found exactly the one you mentioned: c:Category:USB hard disk drives. I then wondered why it only had two subcategories, and, even more strange, each of those has only pics of 'blank' hard disk drives without enclosure - hence, they are strictly not at all USB, but might be produced for use in USB hard disks. After some confusion on my side I realized that this category (strangely categorized very low in the tree as well) is only meant for blank hard disk drives (the inner thing without enclosure etc.) meant for use in USB hard disks. There you go. And it has some logic to it. Hence, I left the whole thing as it was and created the new category c:Category:USB hard disks where it belongs, directly below c:Category:USB devices. I hope I could make myself clear ;-) I found that old category a bit strange alltogether, maybe it could be considered to change/delete it. Pittigrilli (talk) 14:38, 3 February 2021 (UTC)
The names are synonyms, and a functional drive with a USB interface is still a USB hard drive, even if it lacks an enclosure. Maybe there could be a category for disassembled USB hard drives. Ghouston (talk) 22:47, 3 February 2021 (UTC)
@Ghouston:I had a closer look at c:Category:USB hard disk drives. The raw hard drives there have a directly attached USB receptacle, in comparison to the normal SATA interface. Thus, they are indeed "USB hard disk drives" and quite special, I assume. Pittigrilli (talk) 11:07, 5 February 2021 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2021/01&oldid=1360116723"