Shortcut: WD:PC

Wikidata:Project chat

From Wikidata
(Redirected from Wikidata:Project Chat)
Jump to: navigation, search
Wikidata project chat
Place used to discuss any and all aspects of Wikidata: the project itself, policy and proposals, individual data items, technical issues, etc.
Please take a look at the frequently asked questions to see if your question has already been answered.
Please use {{Q}} or {{P}}, the first time you mention an item, or property, respectively.
Also see status updates to keep up-to-date on important things around Wikidata.
Requests for deletions can be made here.
Merging instructions can be found here.

IRC channel: #wikidata connect
On this page, old discussions are archived. An overview of all archives can be found at this page's archive index. The current archive is located at 2017/02.





for permissions


for deletions


for deletion

for comment

and imports

a query



A tool to batch upload labels for wikidata items[edit]

Hello everyone, As a part of the data donation of translated place names, I have developed a label upload script which helps translators to batch upload a list of labels in a specific language for Wikidata items.

The tool uses pywikibot. The input is a CSV of wikidata Ids and corresponding translated labels to upload. The tool ensures that no duplicates entries are uploaded in the following manner:

Label in wikidata State Action
Present Matches with translation Doesn't get uploaded
Present Doesn't match with translation Uploaded as an alias (If it translation not present as alias)
Not present Uploaded as label

I would like to thank @YuviPanda: ,@Planemad: and folks on #wikidata, #pywikibot to help me out through this. Would love to hear the thoughts and feedbacks from the community on how this could be more useful to help the community of translators.  – The preceding unsigned comment was added by Amishas157 (talk • contribs) at 13:14, 21 December 2016‎ (UTC).

Wikiproject Finnish elections[edit]

1. We are starting to prepare the import of the candidates for the Finnish municipal elections in the new Wikiproject Finnish Elections. While the import of the candidates is the main issue and needs to tackle several challenges, I would like to ask for your assistance in making sense of municipal council items in Wikidata. Currently there is municipal council in Finland (Q191893), which is a subclass of municipal council (Q701632), which is a subclass of city council (Q3154693). Any way that could be sorted out? – Susanna Ånäs (Susannaanas) (talk) 08:15, 3 February 2017 (UTC)

Yes, this could be really difficult. One of the more difficult properties to agree around is office contested (P541). In many elections, you vote for the "seats in the city/municipal council". The mayor and other officies, are thereafter indirectly elected by the council. I do not know how it works in Suomi, but that is how it works here in Ruotsi. P541 is still often used for other things than these "seats".
One challenge I have notified is that we here in Ruotsi are not only voting for individual persons, but also for political parties. That is also difficult to describe. (The sum of all votes therefor exceeds the total # of votes.) And on local level, it is not unusual that persons that are not a member of a political party are representing a party. The lists of potential candidates are often endless, since people can vote on almost anything, including Donald Duck (Q6550) and "free alcohol and wider pavements". -- Innocent bystander (talk) 13:58, 3 February 2017 (UTC)
Really happy that you answered! I am sure we are tackling much the same questions in Finland and Sverige. I am more or less confident that all the parts will fall into place well enough. I am hoping not to create unnecessary connectors, such as municipal council 2012–2017 in Helsinki, even though I already was tempted to. There are still unknown objects, such as constituency association (Q10712411) for elections (do they exist more globally?), no label (Q11902699) coalitions between constituency association (Q10712411), or electoral alliance (Q388602) between parties (should they be declared on a municipal or national level) who all establish candidate lists. I have not yet found an item for the list. In Finland, the votes are counted with D'Hondt method (Q337866), and the correlation between the votes and the elected persons is not direct. A key problem is which items should have instances for different years or municipalities etc. We plan to import all candidates if everything goes well. Votes cast on other than the candidates is an issue that we don't necessarily tackle. Any help regarding any of these question will be warmheartedly welcomed! Perhaps the best place will be the project page. Cheers, Susanna Ånäs (Susannaanas) (talk) 14:56, 3 February 2017 (UTC)
I have planned to use office contested (P541) to express that the candidates are running for the council item rather than the individual representative item. If it poses no computational harm, I propose the scope of the object of the property be extended to cover a legislature as well. – Susanna Ånäs (Susannaanas) (talk) 09:51, 4 February 2017 (UTC)
@Susannaanas: Can you give an example of what this would look like? --Oravrattas (talk) 20:57, 4 February 2017 (UTC)
@Oravrattas: I have created an example on the project page. It should already have the most important features, but generally everything is in the making. – Susanna Ånäs (Susannaanas) (talk) 22:36, 4 February 2017 (UTC)
@Susannaanas:. Great, thanks! Based on this, I don't think that extending the scope of office contested (P541) to allow something like Helsinki City Council (Q4117101) is the right way to go. Instead I would suggest creating a separate "position" entity for "Member of the City Council of Helsinki" (or equivalent), similarly to how this is modelled for members of all national legislatures (see, for example, the list at Wikidata:EveryPolitician) --Oravrattas (talk) 09:48, 5 February 2017 (UTC)
@Oravrattas: Thanks! If you think that this would be the right way to go, I will then add the 311 items that this requires for Finland, plus the historical ones. I would also urgently need discussion on the candidacy in election -proposal. I wish you could spare a thought! – Susanna Ånäs (Susannaanas) (talk) 10:10, 5 February 2017 (UTC)
@Oravrattas:I am not very happy with this but I will follow this. I still need a linking property from the local municipal elections to the local city council. It seems arbitrary to link through the candidate/representative. – Susanna Ånäs (Susannaanas) (talk) 08:51, 9 February 2017 (UTC)
@Susannaanas: I'm not sure I understand. The linking to the body is through the office, not through the candidate:
Note also that this is only needs to be set on the 'parent' concept rather than on each individual election, so in the UK is actually
and is only added to the individual election if it needs to be more specific — e.g. on United Kingdom general election, 2015 (Q3586935), office contested (P541) has a qualfier of parliamentary term (P2937): 56th United Kingdom Parliament (Q21084473) --Oravrattas (talk) 06:31, 10 February 2017 (UTC)
@Oravrattas: Thanks for the answer, and sorry for being unclear. I should have said representative rather than candidate. I have created the following object, misusing the office contested (P541) property. When we locate a suitable property, we will replace this. I hope that these statements would not be removed until replacing ones can be written. I have also create related municipality and municipal council items, as well as the item for the elections nationally. – Susanna Ånäs (Susannaanas) (talk) 14:40, 12 February 2017 (UTC)
@Susannaanas: I've tweaked that slightly to show one potential way you could do it, if you don't want to create a new Item for "Member of Alavus City Council" (though I suspect that creating such items will be better in the long term). --Oravrattas (talk) 12:09, 13 February 2017 (UTC)
@Oravrattas: I like that! The term showing up in the Finnish translation is not a correct one, so I will need to check if it's the right item, but if not I'll just change it. Thanks! – Susanna Ånäs (Susannaanas) (talk) 13:53, 13 February 2017 (UTC)

2. I am realizing that candidates are a property of the elections rather than vice-versa. If we have tens of thousands of candidates that model is not sustainable. Is there an existing model for defining it in reverse or should it be created? – Susanna Ånäs (Susannaanas) (talk) 16:11, 3 February 2017 (UTC)

I have reopened the proposal for an inverse property of candidate (P726), see candidacy in electionSusanna Ånäs (Susannaanas) (talk) 09:51, 4 February 2017 (UTC)

3. I think an inverse property is also needed for number of representatives in an organization/legislature (P1410) to be able to sensibly list all the seats of a municipal council by party rather than list all the seats in all the municipal councils in the party item. If it exists, I would be happy if you let me know! – Cheers, Susanna Ånäs (Susannaanas) (talk) 09:51, 4 February 2017 (UTC)

4. Which property to use for the candidate number? Is a new one needed? It would probably be a qualifier to the candidate (P726) / candidacy property. – Susanna Ånäs (Susannaanas) (talk) 10:08, 4 February 2017 (UTC)

No I do not think you should merge constituency association (Q10712411) with anything. Au contraire, we should probably create more items like this.
What is a "candidate number"? It looks like you in Suomi have to nominate a candidate, and only "approved" candidates can get a valid vote?! (Here you can be elected without a nomination or even have approved it yourself. That has happened to Thomas Wassberg (Q195578) at least two times I think.)
The models for how the seats are distributed are often very complex. The model for Swedish elections, are in many cases not a clean version of any of these models, since constituencies and leveling seats make the math rather complex. -- Innocent bystander (talk) 13:10, 5 February 2017 (UTC)
Yes, parties or constituency association (Q10712411) can nominate candidates, which results in candidate lists for each party / association / alliance in each municipality. The candidate has a number in this list. This is one modelling problem: Which property to use for the number, and should I also make an item for the candidate list(s)? If I use the number already as a qualifier for the candidacy then I have no more means to add the list as a qualifier to the number. – Susanna Ånäs (Susannaanas) (talk) 14:42, 5 February 2017 (UTC)

Thanks for help and discussions! The hackathon at Yle was successful, with over 30 (data) journalists taking part and learning to use Wikidata. We will gather our observations on the project page as well as links to hackathon outcomes and future plans. The data import has still a lot of cleaning and fixing to do, and we hope to be well prepared for this years election data in April. – Susanna Ånäs (Susannaanas) (talk) 14:26, 19 February 2017 (UTC)

Items without articles in wikipedias[edit]

There are Wikidata items about Russian mathematicians without articles in Wikipedias:

Is it normal, or these entries should be deleted? -- Andrew Krizhanovsky (talk) 20:17, 5 February 2017 (UTC)

It is normal. See Wikidata:Notability.--ԱշոտՏՆՂ (talk) 21:08, 5 February 2017 (UTC)
yes, it is normal, the work is not finished. here is translation tool mw:Content translation. Slowking4 (talk) 12:34, 6 February 2017 (UTC)

But in the case of the latter three, how is Notability established? They do not fulfill a structural need and have no sitelinks. I am not sure about the serious and publicly available references. And it is about Living People, and their data is unsourced. I would question whether these should really stay. --Denny (talk) 22:53, 8 February 2017 (UTC)

They each have multiple recognizably important external identifiers: which means that other data sources find this notable. I would call that structurally important: after all Wikidata is serving not only our projects, but connecting other kinds of projects. Sadads (talk) 23:10, 8 February 2017 (UTC)
OK, agreed on the second and fourth - they have good external identifiers, like library of congress. But the third one does not. Both of his external identifiers are going to crowdsourced sites. I am unsure here, and would suggest to consider the case. --Denny (talk) 05:02, 9 February 2017 (UTC)
The Mathematics Genealogy Project is a serious source. It describes the person. ChristianKl (talk) 11:05, 9 February 2017 (UTC)
It is crowdsourced and the data is not being checked beyond crowdsourcing. Whereas I obviously appreciate crowdsourced material, I would refrain from using it blindly as a source. Having it additionally to at least one good source would be fine with me, but we are lacking that for the third person above still. --Denny (talk) 17:56, 10 February 2017 (UTC)
I'm not sure whether calling the Mathematics Genealogy Project crowdsourced is right. They take submitted datasets and integrate them. I see no sign that they don't do sanity checks on the submitted data. Additionally zbmath seems to have peer-review for their entries. ChristianKl (talk) 09:20, 11 February 2017 (UTC)
I just took one example - starting at number 3 above, I continued clicking on the first advisor to the end and came to this last link: - Checking the articles on the advisor and the student, and trying to find any evidence for it for a few minutes, I came up with nothing. Don't misunderstand me, I really like the MGP, but I don't trust the data as a primary source. Regarding zbMath, I tried an edit there, and it seems indeed that edits go through a peer review first, so that seems better. --Denny (talk) 21:03, 11 February 2017 (UTC)
I had no problem with finding by searching for author and advisor. To the extent we can trust the scientific publisher springer the person wrote a thesis with this advisor. ChristianKl (talk) 10:12, 12 February 2017 (UTC)

I am agree with Denny that only "Mathematics Genealogy Project" is not enough for person's notability. If we will decide that MGP shows a notability of a person, then we can create Wikidata item for every person who wrote math Dissertation. But it is nonsense. -- Andrew Krizhanovsky (talk) 06:24, 12 February 2017 (UTC)

Why is it nonsense? It's information for which there are trustworthy sources. Having the data allows queries to be made that for example compare the average amount of dissertations that a professor oversees per country. Wikidata is structured data. Having data sets linked together in a way that it's possible to query them with computers is valuable. ChristianKl (talk) 10:12, 12 February 2017 (UTC)
But we're talking here about unsourced personal data of every single math PhD in the world - their birth day, place of birth, etc. This should be at least be required to be sourced, and not be added through a game. --Denny (talk) 18:35, 12 February 2017 (UTC)
the source is MGP. i do not think you have a consensus to exclude "game" information. are you gaming wikidata to conform to your wikipedia bias? Slowking4 (talk) 05:51, 15 February 2017 (UTC)
@Denny: There's no birth day filled for Dmitri E. Kvasov (Q21694206). The one that's filled for Viktor Aleksandrovich Gorbunov (Q18558636) comes from which seems to be a nice academic source.
When looking at the history of the people in question I don't see that this data was added through a Wikidata game. What do you mean with the term game in this context? ChristianKl (talk) 14:31, 15 February 2017 (UTC)
Sorry, no, this was not meant in particular for this. I am just worried that by having the notability criteria so low as to say "yes, everyone with a PhD" - between 1 and 2% of the population of some countries - and then have at the same time have games going through items with missing data where people make decisions in a snap second about living people who are undersourced might lead to problems. --Denny (talk) 19:26, 16 February 2017 (UTC)
I don't think the problem of games adding data to living people in a snap decision has much to do with the quantity of data we have. We might simply avoid building games that fill in information without references about "place of birth" and "date of birth". ChristianKl (talk) 20:08, 16 February 2017 (UTC)

Age at event[edit]

I have tried to locate a property or earlier discussion for a property for the age of the subject at the time of an event. This would be needed for example when presenting age information where the exact date of birth would not be published for a living person. Another usage I have encountered are notes of the age of the mother of a born baby in parish registers. Being able to record the actual data of the source would prevent from interpreting the data unnecessarily, if the year of birth was deduced from the age data. If there's any support, I will make the property proposal. – Susanna Ånäs (Susannaanas) (talk) 09:16, 9 February 2017 (UTC)

I agree that such a property would be useful given that quite often the birth year isn't available but the age at an event is. ChristianKl (talk) 09:52, 9 February 2017 (UTC)
Thanks a lot! I will create the proposal! Cheers, Susanna Ånäs (Susannaanas) (talk) 10:19, 9 February 2017 (UTC)
Created! – Susanna Ånäs (Susannaanas) (talk) 11:22, 9 February 2017 (UTC)
Changes made to scope & name based on discussion. Hoping that the property could be created - pressing need for it. One more day until a week of notice. Could it be made already today? – Susanna Ånäs (Susannaanas) (talk) 08:31, 15 February 2017 (UTC)

Alexa rank updating bot[edit]

Hi everyone, I mainly edit on zhwp and recently I created a bot which can update the alexa rank in infoboxes automatically. When I requested for approval on zhwp, a wikipedian suggested me to run the bot here. Is there a similar bot running? If not, I'll be very glad to run one. However, I'm new to Wikidata and don't know where (i.e. which page to read) to start from. Can anyone tell me what to do next? Thanks, WhitePhosphorus (talk) 15:09, 10 February 2017 (UTC)

Our bot policy is at . The next step is to create a new account for the bot and request permission for the bot at . While the individual Wikipedia's mostly delete old data entries Wikidata puts a date on every entry and keeps old data. As a result I would say that running the bot once per year is fine to update the Alexa rankings on Wikidata but I wouldn't run it more often than that. ChristianKl (talk) 20:08, 10 February 2017 (UTC)
Or keep around the last 3 or so values and delete the older ones. For Alexa I believe the old values are not that important. --LydiaPintscher (talk) 14:47, 13 February 2017 (UTC)
I don't see why old values aren't interested. I can imagine myself interested in the historical traffic patterns of a website. ChristianKl (talk) 22:22, 17 February 2017 (UTC)

states with limited recognition[edit]

What is the criteria for state with limited recognition (Q15634554)? This is a complex and politically sensitive area. How many states must not recognise you before you should be given the label "state with limited recognition"? 31 UN member states refuse to recognise Israel–does that make Israel a "state with limited recognition"? 136 UN member states recognise Palestine, the UN itself calls it an "observer state", and it is a member of UNESCO (a UN specialised agency) and the International Criminal Court–is that not enough to remove the label "state with limited recognition" from Palestine? 110 UN member states recognise Kosovo, and it is also a member of the IMF and the World Bank (both also UN specialised agencies)–is that not enough to remove the label "state with limited recognition" from Kosovo? I feel like the current categorisation of states as "state with limited recognition" could be seen by some as politically biased (even if only unintentionally so). SJK (talk) 12:38, 12 February 2017 (UTC)

The general idea is that you add claims with reputable citations. If different sources claim different things multiple claims can be added and statement disputed by (P1310) can be used. Wikidata itself doesn't set the standards. ChristianKl (talk) 13:19, 12 February 2017 (UTC)
There are two problems with that idea (1) current usage of "states with limited recognition" has no references – Palestine (Q219060) has claim P31 Q15634554, but no reference; same claim made on Kosovo (Q1246) but again no references – do we remove these claims until references are provided? (2) it isn't clear what reference is sufficient to justify "states with limited recognition". For example, considering the case of Israel, there is a resolution passed by the US House of Representatives saying (bottom of page 1) "Whereas the following countries still do not recognize Israel’s right to exist: Afghanistan, Algeria, Bahrain, Bangladesh, Bhutan, Brunei, Chad, Comoros, Cuba, Djibouti, Guinea, Indonesia, Iran, Iraq, Kuwait, Lebanon, Libya, Malaysia, Maldives, Mali, Morocco, Niger, North Korea, Oman, Pakistan, Qatar, Saudi Arabia, Somalia, Sudan, Syria, Tunisia, United Arab Emirates, and Yemen" – is that reference enough to call Israel a "state with limited recognition"? Or, this book Tomis Kapitan, "Philosophical Perspectives on the Israeli-Palestinian Conflict", p. 334 says "The limited recognition of Israel's factual existence is enough for pragmatically oriented peace negotiations to begin and perhaps even to reach some result. But continued refusal to affirm de jure the right to Israel's original and continued existence is self-defeating to everyone and is destabilizing concerning any lasting peace"–is that cite enough to label Israel a "state with limited recognition"? Unless we have objective criteria to define what a "state with limited recognition" means, how do we decide whether any given reference is sufficient to justify the claim? SJK (talk) 20:04, 12 February 2017 (UTC)
@SJK: if you have reference, please add them ; (almost) anything is better than nothing. For touchy case like that, the abundance of reference can only be welcomed. As for the signification state with limited recognition (Q15634554), I'd say that any states that in not fully recognise has ipso facto and by definition only a limited recognition. Cdlt, VIGNERON (talk) 22:41, 13 February 2017 (UTC)
Do you think the people who wrote that house resolutions would find it fair to describe their document as stating that Israel is a state with limited recognition? In this case I doubt that's the case. But there's likely a source that you could find that actually thinks that's the case.
Whether a state accepts Israels right to exist and whether it recognises Israel are also two different issues. ChristianKl (talk) 08:25, 14 February 2017 (UTC)
@ChristianKl: Whether a state accepts Israels right to exist and whether it recognises Israel are also two different issues–are they separate issues? I mean, do any states recognise Israel yet deny its right to exist? Or conversely, do any states accept Israel's right to exist yet refuse to recognise it? In both case I believe the answer is "none". Or, to give another example, does the Republic of Serbia recognise the Republic of Kosovo's right to exist? Serbia denies both that the Republic of Kosovo legally exists and that it has a legal right to exist. Or, yet another example, Israel denies the legal existence of the State of Palestine, which surely implies it would also deny that the State of Palestine has any legal "right to exist". Why does state A refuse recognition of state B? Generally it is because state A considers the process by which state B claims to have come into existence to be illegal or illegitimate–if state A believes that about B, they will surely deny that B has any legal or moral "right to exist" as a state. Indeed, the authors of the cited resolution say "the Republic of Iraq continues to refuse to recognize the existence of the State of Israel and urges the Government of Iraq to recognize the right of Israel to exist", implying that they see "recognition of existence" and "recognition of the right to existence" as at the very least closely connected if not largely identical. Do you think the people who wrote that house resolutions would find it fair to describe their document as stating that Israel is a state with limited recognition? Well, the whole point of the document is to complain about the fact that Iraq (along with 32 other states they list) do not recognise Israel–if 33 states do not recognise Israel, how is that not a case of "limited recognition"–"limited" here meaning "less than universal" (and 33 less is a significant number less)–as contrasted with the universal recognition of Israel which the authors of the resolution desire to achieve? The fact that Israel is a state with limited (as opposed to universal) recognition is the very thing they are objecting to and trying to change, so it would be very contradictory for them to deny that fact. SJK (talk) 10:31, 14 February 2017 (UTC)
Recognition can come through some sort of treaty. I'm for example not sure whether you can be a WTO member without formally recognizing other WTO countries. ChristianKl (talk) 11:09, 14 February 2017 (UTC)
@ChristianKl: Generally speaking that isn't true. To quote the Yearbook of the International Law Commission 1998, Volume 2, part 1, p. 251-252, "it is generally accepted that participation in the same multilateral treaty does not signify mutual recognition, even implicit". It makes that comment while discussing treaty reservations lodged by Arab states at the times of ratifying multilateral treaties to which Israel is a party saying that ratification of the treaty is not an implicit recognition of Israel–the ILC argues those these reservations are legally speaking unnecessary, since even without the reservation the ratification is not an implicit recognition of its states parties; as well as Arab non-recognition of Israel, it discusses unrelated cases, e.g. Australia's non-recognition of East Germany, North Korea, North Vietnam and the People's Republic of China at the time of ratifying the Geneva Conventions in 1949. On the same point I can cite "Text of the draft guidelines constituting the Guide to Practice on Reservations to Treaties, with commentaries, as provisionally adopted by the International Law Commission" which says on page 74, again while discussing reservations denying recognition, "in some cases, the author of the statement is simply taking a “precautionary step” by pointing out, in accordance with a well-established practice, that its participation in a treaty to which an entity that it does not recognize as a State is a party does not amount to recognition". Both ILC documents cite, as justification for the proposition that ratification of a multilateral treaty does not imply even implicit mutual recognition, J. Verhoeven, La reconnaissance internationale dans la pratique contemporaine, Paris, Pédone, 1975, pp. 429–431 (I don't have access to that book, and even if I did, I couldn't understand it because it is in French.) (Bilateral treaties are different from multilateral treaties–unlike a multilateral treaty, a bilateral treaty is evidence of mutual recognition.) Now, if that is the general rule for multilateral treaties, it should be the rule to be applied in the case of WTO membership as well, unless there is some good reason to suppose WTO membership should be treated specially compared to other multilateral treaties. SJK (talk) 08:30, 16 February 2017 (UTC)
And why Islamic State of Iraq and the Levant (Q2429253), a terrorist group, is also a state with limited recognition (Q15634554) (at least on Bosnian Wikipedia and Catalan Wikipedia)? --Liuxinyu970226 (talk) 15:38, 15 February 2017 (UTC)
Well, a "state with limited recognition"–"limited" implies at least some recognition not "none", but as far as I am aware no state recognises ISIS. So this is incorrect. I think the question is whether ISIS is a "state with no recognition" or not a state at all, and that's a difficult question, because it depends on a disputed question of international law – the declarative theory of recognition vs. the constitutive theory. The declarative theory says that states don't depend on recognition to exist, so it is possible for a state to be a state even if no other state recognises–so, even if no other state recognises ISIS, ISIS could still be a state if it meets the factual criteria for being one; by contrast, the constitutive theory says that states exist because they are recognised, so if no one recognises you as a state you can't be one–by that theory ISIS can't be a state. Now, the mainstream position says the declarative theory is correct and the constitutive theory has been rejected, so it is at least possible that ISIS is in fact a state recognised by no one but themselves – but while that is the mainstream position in international law I don't think it is universally accepted. SJK (talk) 08:39, 16 February 2017 (UTC)

Query regarding the descriptions of famous CEO's[edit]

Hi all,

I was looking at the description of some of the famous CEO's like Marissa Mayer,Tim Cook etc. -> American business executive -> American business executive and engineer.

I guess it would be better to have specific description like CEO of Yahoo! for Marissa Mayer, CEO of Apple Inc for Tim Cook. However Sundar Pichai( has Google CEO as description. Can anyone please let me know why is there's disparity in description for different CEO's?

I guess it would be better to have consistent descriptions for all CEO's.

Thanks in advance.


How to set up a default value for a property[edit]

User:Discanto asked me a question that I find useful, but I have no idea what to answer. Is it possible to make a default value appear in a box when we are inserting a property? For example if an Italian user want to insert the "country" property can he find "italy" as a default value? Can it be customized in any easy way? That would be very useful to speed up the work for some of the users I follow. Thank you in advance.--Alexmar983 (talk) 12:34, 13 February 2017 (UTC)

I'm not aware of a direct way. When adding a lot of data like this it can however be useful to use QuickStatements and preformat the data with Excel/LibreOffice Calc. ChristianKl (talk) 12:06, 14 February 2017 (UTC)
There is MediaWiki:Gadget-currentDate.js available as a gadget at Special:Preferences#mw-prefsection-gadgets. It puts the current date into retrieved (P813) by default. —MisterSynergy (talk) 12:11, 14 February 2017 (UTC)

Property for board and card game creators[edit]

Hey everyone,

I've noticed, that there's no consistent way of stating the creators of board and card games.

The two mainly used properties are:

  • Author(P50) which I think is wrong since this is the property for creators of a written work according to the definition. I can see how this happens al lot though, since in German it's called the "Spieleautor" so people resort to P50. Example
  • Developer(P178) which isn't technically wrong since it's intended for creators of software, games etc. but it's mainly used for software/digital products so it might be confusing to use this one. Example

And then of course there's the generic Creator(P170) which can be used for anything but doesn't seem to be used for board games much.

Many items don't have anyone stated as their creator even when there are a handful of other statements there. Don't know if it is directly connected to this issue though.

I think we should decide on which property is the right one to use for board and card games.

My questions are:

  • Should it be the same one used for computer games?
  • Should game authors get a separate property?
  • ...

Any ideas/opinions would be very welcome! Incabell (talk) 16:16, 13 February 2017 (UTC)

-- just saw that BoardGameGeek calls them designers. Incabell (talk) 16:37, 13 February 2017 (UTC)

I agree "designer" is the right term for board games, e.g. it's echoed on Wikipedia as well, and there's often a dedicated credit for that role (even to the point of being a selling point for board games). Does "designed by" property therefore work? Then, the designer can have "game designer" assigned to their item entry, as appropriate? Video games, on the other hand, have more varied and elaborate credits, particularly for larger projects. For example, Sid Meier is listed as "director" of Civilization V, yet according to MobyGames, he's technically the "Creative Director" while there are also 3-4 other people in different "director" roles on the game, plus another person yet (not Sid) who's listed as "designer." Should all these credits be modeled? Maybe, maybe not; and complicating things further, this mix of roles is probably different between studios or even projects. So, "designed by" for board games, more discussion for video games? Argotechnica (talk) 02:16, 14 February 2017 (UTC)
I think it's not that easy: The term "game designer" is common in the englisch language, in German the same person is named "Spieleautor" while the designer is that person who is responsible for the artwork (artist in english board game descriptions). Therefore I would say that "Spieleautor" = "game designer" should be a property by it's own to avoid the problems described above. -- Achim Raschka (talk) 09:29, 14 February 2017 (UTC)
Good point - and someone listed on a game under the "game designer" property may not necessarily be considered to have the "occupation" property of game designer. So - new property, "game designer"? Argotechnica (talk) 18:13, 14 February 2017 (UTC)
I like that idea. Just for clarification, would this then be exclusively for non-computer games i.e. board- and card games? If we make the German translation of "game designer" -"Spieleautor" this will not be usable for computer games at least from the German language side. Otherweise it would probably need to be something like "Spieleentwickler". Incabell (talk) 11:07, 15 February 2017 (UTC)
IMO it would be better to use quite a generic property - eg creator (P170) or contributor(s) to the subject (P767), and then a qualifier to indicate the exact nature of the role. That is how film and music articles typically work at the moment, using as (P794) with a value to give more detail on the nature of the contribution; eg:
stated as (P1932) can be used to give even more information:
The film industry has evolved fairly defined terms for particular roles; for video games things may be a lot more fluid, so more generic properties like creator (P170) or author (P50) may be a good idea. The advantage of using something fairly generic, and then qualifying it, is that the grouping together makes searching a lot more straightforward, so that doesn't have to search separately for "director", "supervising director", "sequence director" etc; but one can still extract any of those particular roles out if one wants.
There is a general preference on Wikidata not to multiply the number of properties unless really necessary: to prefer properties to be generic, and specify details using items. This principle will probably be argued to you in opposition, if you propose a new highly specific property without a very good case. Jheald (talk) 14:22, 15 February 2017 (UTC)
I should add that as (P794) is currently being discussed, and may be divided into two new qualifier properties, "subject has role" and "object has role"; but the principle stands. Jheald (talk) 14:30, 15 February 2017 (UTC)
That's a very fair point you have. I just want there to be a consensus on how to do it, so the statements are the structured the same way. I've created Wikidata:WikiProject_Board_Games to have a central place to collect and discuss how to structe board game items and so forth. I suggest to move all further discussion to the talk page there. Incabell (talk) 10:53, 16 February 2017 (UTC)

Property:P952 ISCO code - string instead of identifier[edit]

Why that? Looks like nonsense. 00:02, 14 February 2017 (UTC)

See Wikidata:Identifier migration/0. Sjoerd de Bruin (talk) 12:38, 14 February 2017 (UTC)

2 days past since looking, and it is still a string. 19:47, 16 February 2017 (UTC)


Please merge Q20997623 and Q2094670, they both are disambiguation pages, one with links to wp:de, wp:en and wp:es, the other one to wp:it. --Gejotape (talk) 07:45, 14 February 2017 (UTC)

Pignatelli (Q20997623) is no disambiguation page but an item for a last name. The Italian page for Pignatelli (Q2094670) is about more than just a last name and also list architecture. It's debateable to which of those two Wikidata items the English Wikipedia page should go, but there's no basis for merging the Wikidata items. ChristianKl (talk) 08:20, 14 February 2017 (UTC)
The english page is about surname, the others are real disambiguation. I moved the english page. --ValterVB (talk) 17:54, 14 February 2017 (UTC)

Wikidata weekly summary #247[edit]

Looking for qualifier to use when calculating date of birth[edit]

Can someone please point me to the qualifier pair to use where you can calculate a year of birth when you have a date in another record at a point in time with an age. I think that there could be some value in adding some batches of qualifiers against the lists that we generate for properties, eg. against "date of birth" Having to dig for these is a PITA.  — billinghurst sDrewth 10:06, 14 February 2017 (UTC)

I think the nearest we get is the generic sourcing circumstances (P1480) -- one could qualify with "circa" and maybe a new value "calculated date"
Susanna Anas has a proposal in for a qualifier "age at event" which, once adopted, could be used to record the age in the record with the date. Jheald (talk) 19:59, 14 February 2017 (UTC)
Yes, I saw that later when hunting and probing through wikidata: ns links to date of birth. "Age at event" records data at the event, and it needs a reciprocated qualifier for the evidence (my comment at the proposal). I am comfortable with "sourcing circumstances" and something like "estimated" or "calculated" though don't see the need for the additional specification of "... date", though *if* we are making the jump, the reference itself is not sufficient, we need to tie in the event, it needs two factor authentication.  — billinghurst sDrewth 21:48, 14 February 2017 (UTC)
There is also inferred from (P3452), which lives in the references section for a statement, currently being used all of 10 times to indicate a total of 4 distinct sources.
I think it was originally intended that its value should be a specific Wikidata item would specifically contain the information needed to verify the statement: so from the ownership details of Oriental Coin Cabinet Jena (Q824791) one can infer that each of those owners was a coin collector; or from the inclusion of an individual in List of coin collectors (Q19893121).
But it might be an acceptable use for its object to instead indicate the nature of the source of the information, eg "age at death". Jheald (talk) 10:30, 15 February 2017 (UTC)
A neat way to do this might be to propose a new property "inferred from property", with the constraint check that that property should also appear on the item. So one might have e.g.:
... etc. Jheald (talk) 11:17, 15 February 2017 (UTC)

meaning of has parts of the class (P2670): related to "has part" (composition)? Or more related to "has subclass"/"has superclass" (generalization/specialization)?[edit]

Which of the following is most correct about the meaning of has parts of the class (P2670) "has parts of the class"?

  1. has parts of the class (P2670) is a property that tells us something about the "parts"/composition of a class (or its instances), something like "has part" does (without having exactly the same meaning as "has part" of course)
  2. has parts of the class (P2670) is a property that helps define generalization or specialization of membership in this class, i.e. tells us something about whether certain things may or might not be instances of the class, something like "subclass of" (or its implicit inverse "superclass of") does (without having exactly the same meaning as "subclass of" or "superclass of", of course)
  3. has parts of the class (P2670) can mean either of 1 or 2 (ambiguous?)
  4. Neither 1 nor 2 is correct at all
  5. Undefined answer / meaningless question??

Some of the property's aliases seem suggestive of 1 or 2, but these are completely different meanings of the property. DavRosen (talk) 14:01, 14 February 2017 (UTC)

  • 1 is the intended meaning - this is a specialization of "has part" to cover the cases where one object has many subcomponents of a common class that may possibly be counted (the canonical example was "unvierse" and "stars"). Some folks seem to have edited the property labels, descriptions, and examples in the last few weeks to confuse this, particularly the "article" example seems to be your (2), and I believe is wrong according to the intended purpose of this. It was never intended to be an inverse relation to subclass of (P279). @TomT0m: as proposer - any additional comments on this? ArthurPSmith (talk) 14:22, 14 February 2017 (UTC)
    This is not a specialization. «Has part» is supposedly linking, on first approximation, instances of objects to other instances of objects, that is «my car» has part «my car’s rear left wheel». But «has part» is not supposed to express stuffs like «my car» has part «wheel». Wheel is a class of objects, «my car» is just an object. But we suppose it’s OK to link «car» to «wheel» with «has part» because they are both classes. In general, «has part» should probably only be used for classes which ar eof the same order (see Help:Classification and Metaclass_(Semantic_Web) for more what are classes and class order). --author  TomT0m / talk page 16:02, 14 February 2017 (UTC)
  • This query: most connected classes for has parts of the class (P2670) I found quite interesting.
The results should be interpretated as saying there are uses of the property of the form
<instance of> column 1 has parts of the class (P2670) <sub-type of> column 2.
The only one might look a bit odd from that query might be the line:
( <instance of> States-Provincial (Q2097101) ) has parts of the class (P2670) ( <sub-type of> position (Q4164871))
but looking at an actual concrete example,
Provincial Executive of Utrecht (Q27650837) has parts of the class (P2670) member of the Provincial Executive of Utrecht (Q27650845)
I suppose that does make sense. Jheald (talk) 14:42, 14 February 2017 (UTC)
  • So it could be described as "has part that is an instance of" or "has one or more parts that are instances of"? Perhaps someone could edit the description and aliases to make them less ambiguous? DavRosen (talk) 16:36, 14 February 2017 (UTC)
I edited the 'en' description and removed the bad example; other languages may also need some attention. ArthurPSmith (talk) 16:56, 14 February 2017 (UTC)
Okay, your new description (1 below) does look much better to me than the old one. But compare these as sentences (the way many of us "read" property relations):
  1. Joe's car contains items that are instances of the target class wheel.
  2. Joe's car has parts that are instances of the class wheel.
  3. Joe's car has parts that are instances of wheel.
  4. Joe's car has parts (some constituents) that are instances of wheel.
In my opinion, (1) makes an awkward sentence, and "contains items" could easily be misinterpreted to mean "has instances", rather than "has part(s)". Also the other properties don't usually have the word "target" which not everyone might understand. But "class" could be misinterpreted as refering to the subject item having this property, and (2) is still a bit awkward, so I would favor someting more like (3), or, better yet, (4). DavRosen (talk) 17:25, 14 February 2017 (UTC) - edit; added (4) DavRosen (talk) 17:49, 14 February 2017 (UTC)
@DavRosen: well the label of the property still says "has parts of the class"; between label and description the idea is to be clear. I don't actually see a way to translate your (3) or (4) into description text, do you have something specific in mind? And if you do have something concrete perhaps you should just edit it yourself? ArthurPSmith (talk) 19:25, 14 February 2017 (UTC)
@Ogoorcs: you have done some editing of this property's description in other languages (eg. Italian) - and using it elsewhere - please ensure what you are doing is consistent with what we have discussed here, or contribute to the discussion here, thanks! ArthurPSmith (talk) 15:06, 15 February 2017 (UTC)
@Ogoorcs: check your Italian description (& aliases?) again; it might not be consistent with the English description which I just updated again to help make it less ambiguous.DavRosen (talk) 17:58, 15 February 2017 (UTC)
@DavRosen: @ArthurPSmith: I'm sorry for delay, I'll add my contribution to this topic and check aliases and property as soon as I can. --Ogoorcs (talk) 18:13, 15 February 2017 (UTC)

Continuing this topic in a new section down below. DavRosen (talk) 17:52, 15 February 2017 (UTC)

The point of having three instances of names?[edit]


I noticed that the language links for the Wikipedia article of Felix was not linking correctly has the languages had been split into two groups. One group that used Felix (given name) and another group that used Felix (name). To resolve this I moved every language to Felix name and it now works.

However I don't see the point of having this setup with three instances of the same name and wonder why this is the case. Can someone explain? Almost every name can be used as both a given name and a surname and there is no reason to have separate pages for them. In the case of Felix the instances are:

Q971545 Felix (name) Q18177136 Felix (given name) Q18325135 Felix (surname)

Wouldn't it be enough to have just Felix (name)?

--Immunmotbluescreen (talk) 17:58, 14 February 2017 (UTC)

Exist a Wiki project for the Name. You can report there your doubts. Meanwhile I restore the item. --ValterVB (talk) 18:22, 14 February 2017 (UTC)
I don't really want to work on Wikidata, but from a Wikipedia perspective you vandalize. You should be able to reach the English version from other versions. The only way to achieve this is to place everything on name. You link did not specify clearly what to do in case of Wikidata:WikiProject_Names#For_a_combined_family_name_with_given_name. This way seem to be the same as the example of Alonso so it is correct? I reverted again manually.
--Immunmotbluescreen (talk) 21:20, 14 February 2017 (UTC)
Having only one name page instead of three would work, if Wikipedia does not have different pages for given names and family names, e.g. en:John (given name) and en:John (surname). Because Wikipedia does sometimes this distinction, we do it on Wikidata always. A mix of using sometimes one, sometimes two and sometimes three pages is not structured data. Moreover, there are certain properties, e.g. name day (P1750), which do only apply to given names but are wrong for family names. --Pasleim (talk) 10:48, 15 February 2017 (UTC)
Wikidata items don't exist primarily for interwiki links. Merging items together just to have more interwiki links is generally not desired. If you are not happy with the interwiki links that Wikidata provides you can manually add the ones you want on the Wikipedia side. ChristianKl (talk) 09:58, 17 February 2017 (UTC)

Election hackathon import[edit]

Hi, there is regarding the Finlands municipal elections 2012 and 2017 and 17. feb hackaton for the elections and plan was to upload all (total 37125 candidates with 9674 elected ones) candidates from 2012 elections to Wikidata, but because licensing issue is still under discussion and also least i think that there are some privacy issues for uploading all candidates it has been delayed. (in project page there is more information ) However hackaton is at friday so it would be nice to get elected ones online for the hacking so i would like to push them if it is ok  ? Older discussions for this are here and here.

Plan for this is next
  1. create new wikidata items for councilmembers which clearly doesn't exist in wikidata already (example Heli Piirainen (Q28777227)). (~7000 candidates)
  2. Handle clear cases of existing ones by hand (~1000 candidates eg just approve by hand the updating of the existing wikidata items)
  3. Do something tho the rest (manual work is needed)

Also implementation least for creating the new items is using quickstatement if it works nicely and if it is not then pywikibot. --Zache (talk) 09:42, 15 February 2017 (UTC)

Split "Holy See–Italy relations" and "Apostolic Nunciature to Italy"?[edit]

Italy has an embassy in Vatican, and Vatican has an embassy (called nunciature) in Italy.

So I am worried to see that Holy See–Italy relations (Q3932303) is marked as an instance of Apostolic Nunciature (Q620045) (a subclass of embassy (Q3917681)).

Ideally I believe there should be one Wikidata item for every single embassy. But I guess providing "Apostolska nunciatura v Italiji" visitors an interwiki link is considered better than nothing?

Looking forward to hearing your advice about this, thanks! Syced (talk) 13:31, 15 February 2017 (UTC)

Your thinking is absolutely correct; the English and Italian links should be split off into their own item since they cover relations and not the locations through which these relations happen. Mahir256 (talk) 14:40, 15 February 2017 (UTC)
Thanks for your feedback! Done: Apostolic Nunciature to Italy (Q28788295) Syced (talk) 07:44, 16 February 2017 (UTC)

Making Flow the default for new accounts[edit]

Given that this country prefers to use flow, as the recent poll suggests how about enabling flow as the default way for discussion on new user pages? ChristianKl (talk) 13:36, 15 February 2017 (UTC)

Semi-OT: can you please link to “the recent poll” results? I’m curious about it … Thanks, MisterSynergy (talk) 13:49, 15 February 2017 (UTC)
meta:Collaboration/Flow_satisfaction_survey/Report#Overall_satisfaction_based_on_project_where_Flow_is_mainly_used --Pasleim (talk) 13:55, 15 February 2017 (UTC)
Symbol support vote.svg Support Flow pages are much more clear. Btw, how can I change my user talk page to Flow?--Micru (talk) 15:40, 15 February 2017 (UTC)
You can just enable it in your Beta settings. (Different from other features there, enabling and disabling results in some actual edits being performed by some bot or service, so one should try to avoid toggling it on and off every second day.) --YMS (talk) 15:50, 15 February 2017 (UTC)

Symbol oppose vote oversat.svg Strong oppose Please, no. Please leave this weak forum impersonation in the dustbin, where it belongs. Make this crap at maximum opt-in for those, who explicitly want it, not default for defenceless newbies. Sänger (talk) 16:56, 15 February 2017 (UTC) As long as there is no way to write on a proper, flexible talkpage besides a dumb Flow one, there should be none of these dumbed down pet project junk software extensions installed. Sänger (talk) 22:30, 16 February 2017 (UTC)

I also oppose the move.--Ymblanter (talk) 20:42, 15 February 2017 (UTC)
Yikes. That was unnecessarily harsh. (Symbol oppose vote.svg Oppose.) --Yair rand (talk) 02:18, 16 February 2017 (UTC)
Symbol oppose vote.svg Oppose This extension with an uncertain future and that has significant flaws should not even be enabled on this wiki, in my opinion. --Rschen7754 02:32, 16 February 2017 (UTC)
Symbol oppose vote oversat.svg Stöder absolut inte Every present talk_page with Flow installed should instead have a mandatory Flow free alternative! -- Innocent bystander (talk) 09:32, 16 February 2017 (UTC)
Symbol oppose vote.svg Oppose Current opt-in state is appropriate.--Jklamo (talk) 09:37, 16 February 2017 (UTC)
Symbol oppose vote.svg Oppose One problem with Flow is that it makes it harder than with a conventional talk page to follow the full sequence of interactions that the user has had. This may be a particularly bad negative for new users who may be just finding their feet here -- if one is thinking how to write or reply to them to help them up some next step of the learning curve here, one wants to be able to see as easily as possible the sequence of what advice they may have already have been given. As well as that, coming new to Wikidata is already quite a big step into the unfamiliar for experienced Wikipedians; adding on top of that what may well be an entirely new and unfamiliar talk system to them as well feels like putting an additional unnecessary obstacle in their way. Jheald (talk) 09:39, 16 February 2017 (UTC)
  • The mechanism for the proposed change has not been mentioned, which makes it a little ahrder to assess. As users are able to turn it on if they wish, I don't see the need to force it. Also, how would users go through the reverse process? I don't see the case made for the change.  — billinghurst sDrewth 09:48, 16 February 2017 (UTC)
Symbol support vote.svg Support We use Flow on the French-speaking project chat and there is no problem. I think instead that it is easier for new users. They should not have to learn wikicode. Tubezlob (🙋) 13:16, 16 February 2017 (UTC)
@Tubezlob:It's interesting how the French project chat is much more active than the German project chat even through German has more translated labels. Do you know whether this also was the case before you switched to flow? ChristianKl (talk) 13:13, 21 February 2017 (UTC)
Symbol oppose vote oversat.svg Strong oppose it's a beta feature with a unknown prospect. --Succu (talk) 22:41, 16 February 2017 (UTC)
Symbol oppose vote oversat.svg Strong oppose.--Arbnos (talk) 23:42, 16 February 2017 (UTC)

Equivalent of QuickStatements to remove statements?[edit]

I have found many cheese varieties whose item is an instance of cheese, whereas I believe they should be a sub-class of cheese.

Question: How to remove these many instance of (P31) statements quickly and easily?

Unfortunately QuickStatements does not seem to have that feature yet. Thanks! Syced (talk) 13:37, 15 February 2017 (UTC)

Petscan can perform SPARQL queries and add/remove claims on the results set once you’ve authorized Widar. —MisterSynergy (talk) 13:47, 15 February 2017 (UTC)
PetScan works for this indeed, thanks! Memo: Tab "Other sources" (not "Wikidata"), enter SPARQL query, run, ignore error messages, authenticate WiDar, enter "-P31:Q10943" and execute. Syced (talk) 04:52, 16 February 2017 (UTC)
The new version of QuickStatements does include it. Matěj Suchánek (talk) 14:54, 15 February 2017 (UTC)
That version of QuickStatement says "Try it on the Sandbox item only, for now", and actually I can't figure out how to load items in it. Syced (talk) 03:34, 16 February 2017 (UTC)

NoBounds constraint for quantity-type properties[edit]

We have ~340 properties with quantity type. Can we have a new constraint template for this property type to indicate that a particular property typically should not or must not be used with bounds (i.e. amount only)? The concept of uncertainty does not apply for all quantities, thus bounds have been made optional recently. Right now we have plenty of bounds in our claims just because it was inevitable to add them, it might be useful to tidy them via covi pages. —MisterSynergy (talk) 15:37, 15 February 2017 (UTC)

To push this once: who is working in the field of constraint definitions? Alternatively to a new constraint type, we could also use Complex Constraints with a somewhat simple SPARQL query such as this one for VAT-rate (P2855):
SELECT ?item { ?item p:P2855 [ psv:P2855 [ wikibase:quantityLowerBound ?lowerBound; wikibase:quantityUpperBound ?upperBound ] ] . }
Try it! I think it is important to distinguish between quantities for which there simply is no uncertainty (which are mostly not determined by physical measurements, VAT-rate (P2855) is an example) and properties with (possible) uncertainty and therefore bounds (which might be 0). We also distinguish between no value and concept of no-value in Wikibase (Q19798647) (or other similar NULL-like custom values). —MisterSynergy (talk) 09:35, 19 February 2017 (UTC)


How do I state that somebody is a refugee? Thanks, GerardM (talk) 16:42, 15 February 2017 (UTC)

Don't. Nikkimaria (talk) 02:36, 16 February 2017 (UTC)
I don't support labelling someone a refugee. MechQuester (talk) 02:40, 16 February 2017 (UTC)
When people make a point of BEING a refugee, it is important to be able to register it. Mind you I stopped adding ethnicity and mostly nationality because it is so problematic. But denying that refugees exist is worse than our registering that people have been in jail. Thanks, GerardM (talk) 06:19, 16 February 2017 (UTC)
Given, for example, the individuals listed under en:Category:Refugees, it would seem reasonable to describe at least some subjects as being, or having been, refugees. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:48, 16 February 2017 (UTC)
Use 'significant event', with properties like "applied for asylum", "asylum granted". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:48, 16 February 2017 (UTC)

What property would you use for a "refugee"? MechQuester (talk) 06:29, 16 February 2017 (UTC)

There is refugee (Q131572) and asylum seeker (Q564811). Thanks GerardM (talk) 07:21, 16 February 2017 (UTC)

has parts of the class (P2670) (description and aliases)[edit]

It appears that the description and aliases are still being mistranslated (presumably misinterpreted) even as #1 below, so I tried to make it even more explicit as #2, and then added an additional part of the description that makes its close relationship to "has part" (P527) clearer (this is #4 --edit: I had #3 but it was too long--).

  1. class having instances (members) that are asserted to be components (parts) of the subject (whole)
  2. class having some instance(s) that are each asserted to be a component «part of (P361)» (NOT an instance of) the subject instance(s)
  3. class having instance(s) that are asserted to be component «parts of (P361)» (NOT instances of) the subject. The subject (if it is not a class), or each instance of the subject (if a class), «has part(s) (P527)» (not instances) that are each an instance of this target object (class)
  4. class having instance(s) that are asserted to be component «parts of (P361)» (NOT instances of) the subject. The subject (or each of its instances if a class) «has part(s) (P527)» (not instances) that are each an instance of this object.

I know it's long but I don't know how else to prevent misinterpretation. I think the problem is that we haven't found an English word for component that can't be (mis)interpreted to mean component of a set or instance of a class. Other suggestions?
Also I'm tweaking the aliases so that some of them might translate better even if some might sound awkward in English, and adding some property references to them. If they are too long or whatever, let me know (or trim them down).
DavRosen (talk) 17:51, 15 February 2017 (UTC)

  • Update: I added a new one (#4) above that I actually used, because #3 was too long :-) DavRosen (talk) 18:13, 15 February 2017 (UTC)
  • I'm pinging the Ontology project folks here. In your #4 I think you can leave out the 2nd and 3rd parenthetical statements - "or each of its instances if a class" and "not instances", a description doesn't have to be totally comprehensive and I think it's clear enough without those. Further details could be provided as Wikidata usage instructions but I'm not sure that's really needed here. ArthurPSmith (talk) 21:52, 15 February 2017 (UTC)

--Micru (talk) 21:46, 24 August 2014 (UTC) Tobias1984 (talk) TomT0m (talk) Genewiki123 (talk) Emw (talk) 03:09, 9 September 2014 (UTC) —Ruud 16:15, 9 December 2014 (UTC) Emitraka (talk) 14:32, 14 October 2015 (UTC) Bovlb (talk) 19:10, 21 October 2015 (UTC) Peter F. Patel-Schneider (talk) 22:21, 23 October 2015 (UTC) ArthurPSmith (talk) 15:51, 5 November 2015 (UTC) --Daniel Mietchen (talk) 20:53, 3 January 2016 (UTC) --Harmonia Amanda (talk) 22:00, 27 February 2016 (UTC) --Lechatpito (talk) --Andrawaag (talk) 14:42, 13 April 2016 (UTC) --ChristianKl (talk) 16:22, 6 July 2016 (UTC) --Cmungall Cmungall (talk) 13:49, 8 July 2016 (UTC) Cord Wiljes (talk) 16:53, 28 September 2016 (UTC) DavRosen (talk) 23:07, 15 February 2017 (UTC)

Pictogram voting comment.svg Notified participants of WikiProject Ontology

  • Okay. I, too, am interested to hear any input from Ontology proj. Based on my limited experience with wikidata, my thinking was, I've seen large numbers of divergent uses of the most basic properties, so why not try to cover all the basic use cases in the description (within 250 chars) so that one can check it without going to a faq or a help page repeatedly. One of the most basic questions that I and many others still aren't clear on is *precisely* what a given class-class relationship means (i.e. exactly what it implies about the instances of subj vs instances of obj), so I thought we should spell it out for a particular property -- I don't think it's obvious to everyone. Also this particular property apparently hasn't been added to the basic membership properties page :-) DavRosen (talk) 23:01, 15 February 2017 (UTC)
  • BTW, I tried to follow the pattern that "most" of the property descriptions do (I looked at dozens of them): they are intentionally written as direct descriptions of the target object (without explicitly saying so), while the subject must be referred to as such explicitly. (But if there is a second sentence in the description this doesn't need to follow the same rule.) Confusingly, among the notable exceptions to the point-of-view-of-the-object rule are the current descriptions of "part of" and "subclass of" :-) DavRosen (talk) 23:48, 15 February 2017 (UTC)

A new Labs Tool to visually explore etymological relationships extracted from the English Wiktionary and an RDF database of etymological relationships[edit]

Hi all! I have developed a tool to visualize etymologies. Please check it out at My work is funded by an IEG grant. Please leave your feedback here. It will help improve it.

As a first release, it's is impressive how well automatic extraction of data works (with some bugs of course...). This is because Etymology Sections are written using well defined standars. I would like to get some feedback about some difficulties I have encountered while extracting data and some ideas I have about new templates. I wrote some notes here. Please add your comment there if you have any. Some additional notes follow:

  1. I could not use trees as in the nicer demo because there are loops between words that cannot be fit in trees (in trees branches don't merge). Loops are conflicting etymologies. Many are due to simple inconsistencies that users can easily fix, others are real conflicting etymologies and should be represented using multiple trees. I will work on this in a future release.
  2. Etymology Sections rarely link to words and their sense/pos, generally only link to the lemma. This is a problem for homographs, cause they generally have different etymological trees which get mixed up in this current implementation. See for example the discussion on the Etymology Scriptorium. It would be nice to have more precise links in etymology sections that link to the correct word.
  3. I am not plotting all derived words as of now to clean up a bit the visualization.

Looking forward to your comments! Epantaleo (talk) 18:29, 15 February 2017 (UTC)

a screenshot of the graph for word coffee
Looks great! Regarding your Suggestions on how to make Etymology Sections easy to parse, did you present them at wiktionary yet (maybe even at bot requests)? Thank you for this awesome work! --Atlasowa (talk) 21:01, 15 February 2017 (UTC)
@User:Atlasowa: Thanks for your feedback! I have posted something in the beer parlour, the etymology scriptorium and the grease pit. Do you have more specific ideas? I'll try with mailing lists maybe. Epantaleo (talk) 00:12, 16 February 2017 (UTC)

GS1 product codes[edit]

There are Wikidata properties for GS1 country codes and GS1 manufacturer codes. I was thinking of proposing adding a property for GS1 product category codes. Is there any problem in adding this kind of property? Peter F. Patel-Schneider (talk) 21:12, 15 February 2017 (UTC)

@Peter F. Patel-Schneider: I see no reason against it. You could create a property proposal at ChristianKl (talk) 13:37, 17 February 2017 (UTC)
I expect that this is the route I will go, probably with some other exploration first. Peter F. Patel-Schneider (talk) 15:18, 17 February 2017 (UTC)

Review of initial updates on Wikimedia movement strategy process[edit]

Note: Apologies for cross-posting and sending in English. Message is available for translation on Meta-Wiki.

The Wikimedia movement is beginning a movement-wide strategy discussion, a process which will run throughout 2017. For 15 years, Wikimedians have worked together to build the largest free knowledge resource in human history. During this time, we've grown from a small group of editors to a diverse network of editors, developers, affiliates, readers, donors, and partners. Today, we are more than a group of websites. We are a movement rooted in values and a powerful vision: all knowledge for all people. As a movement, we have an opportunity to decide where we go from here.

This movement strategy discussion will focus on the future of our movement: where we want to go together, and what we want to achieve. We hope to design an inclusive process that makes space for everyone: editors, community leaders, affiliates, developers, readers, donors, technology platforms, institutional partners, and people we have yet to reach. There will be multiple ways to participate including on-wiki, in private spaces, and in-person meetings. You are warmly invited to join and make your voice heard.

The immediate goal is to have a strategic direction by Wikimania 2017 to help frame a discussion on how we work together toward that strategic direction.

Regular updates are being sent to the Wikimedia-l mailing list, and posted on Meta-Wiki. Beginning with this message, monthly reviews of these updates will be sent to this page as well. Sign up to receive future announcements and monthly highlights of strategy updates on your user talk page.

Here is a review of the updates that have been sent so far:

More information about the movement strategy is available on the Meta-Wiki 2017 Wikimedia movement strategy portal.

Posted by MediaWiki message delivery on behalf of the Wikimedia Foundation, 20:30, 15 February 2017 (UTC) • Please help translate to your languageGet help

Order of presentation for a person[edit]

Hoi, I think that having won an award is more relevant than being nominated for an award. So I prefer it when the order is reversed and "award received" comes first. Thanks, GerardM (talk) 11:13, 16 February 2017 (UTC)

I agree. Maybe it makes sense to create a phabricator ticket? ChristianKl (talk) 21:15, 19 February 2017 (UTC)

Proportion of female voters and votes[edit]

I have used female population (P1539) to express the number of women voters of all voters. Since we are querying human (Q5) it seems to suit well. However for the proportion of votes cast by women, I used a different structure with applies to part (P518), because we are querying votes rather than humans. This is horrible to query. How would you do this? Here's an example. – Susanna Ånäs (Susannaanas) (talk) 12:28, 16 February 2017 (UTC)

I think you are right. If this is put on the main "ballots cast" statement, it will really mess up querying, because the most common queries would be for the total ballots cast.
Instead, I think a separate property "breakdown of ballots cast" would be useful, with a mandatory "applies to part" qualifier. That would keep this data away from the total ballots cast, thus removing the possibility of confusion, and be more efficient for queries. Queries for total ballots cast wouldn't have to filter to avoid the qualifier; and queries for partial ballot counts wouldn't have to avoid the total ballot counts. Jheald (talk) 14:41, 16 February 2017 (UTC)


Hello. I have just created the following pages, to help us find people interested in specific English-speaking countries:

Please don't hesitate to add your name to the list so that you can get pings whenever the subject might be of interest to you. Thierry Caro (talk) 12:55, 16 February 2017 (UTC)

Linking to these pages are property proposals where your vote would already be useful. Thierry Caro (talk) 13:04, 16 February 2017 (UTC)


Yesterday i tried to use Pywikibot to create statement with claims and i was unable to figure out how it should be working. It seems that least these examples 1 and 2 are out of date and it is possible that pywikibot:s code is too. However currently i am out of my skills with that so i ask if somebody with better skills can show me how to create claim with one qualifier and with source with information enough to be rendered properly at someday. In example property position held (P39) in item Vuokko Vakkuri (Q28789936) is something what i am trying to do using pywikibots claim.addSources() and claim.addQualifier() functions. --Zache (talk) 13:45, 16 February 2017 (UTC)

There is Wikidata:Pywikibot - Python 3 Tutorial. --Molarus 14:33, 16 February 2017 (UTC)
I believe these examples are ok.
# assume you've got an 'item' you want to modify and the repository as 'repo'
claim = pywikibot.Claim(repo, 'P39') # you initialize the claim, though it's a stub object now
target = pywikibot.ItemPage(repo, 'Q18694248') # this is the value
claim.setTarget(target) # now the claim is connected to the target but hasn't been saved yet
item.addClaim(claim) # this saves the claim to the item

qualifier = pywikibot.Claim(repo, 'P2715', isQualifier=True) # qualifiers are claims as well
qtarget = pywikibot.ItemPage(repo, 'Q28753630') # target for the qualifier
qualifier.setTarget(qtarget) # just like above, not saved yet
claim.addQualifier(qualifier) # 'claim' is already connected to the item, so we can attach a qualifier to it

date = pywikibot.Claim(repo, 'P813', isReference=True) # references are claims as well
date.setTarget(pywikibot.WbTime(year=2017, month=2, day=14))
publisher = pywikibot.Claim(repo, 'P123')
publisher.setTarget(pywikibot.ItemPage(repo, 'Q54718'))
title = pywikibot.Claim(repo, 'P1476')
title.setTarget(pywikibot.WbMonolingualText('Ylitornio: Ehdokkaat vertauslukujärjestyksessä', 'fi')) # yep, this is nowhere in the examples
url = pywikibot.Claim(repo, 'P854')
claim.addSources([date, publisher, title, url])
Tested on the sandbox. Matěj Suchánek (talk) 16:22, 16 February 2017 (UTC)
You can find another sample here: SV-berg but I didn't run this one for about a year. Edoderoo (talk) 22:26, 16 February 2017 (UTC)

Ethnic communities and diasporas[edit]

Italian American (Q974693) and similar items are not really ethnic groups, so I've made this an instance of the newly-created ethnic community (Q28790362). On another level, I have also been developing Overseas Chinese (Q779191) and African diaspora (Q385967) as instances of diaspora (Q133004) items. What do folks think of this approach of distinguishing between ethnic groups, national/regional ethnic communities, and diasporas?--Pharos (talk) 15:43, 16 February 2017 (UTC)

@MechQuester, Billinghurst, TomT0m, Jc3s5h, Sjoerddebruin: Realizing now this is a bit of a sequel to Wikidata:Project chat/Archive/2017/01#Irish-American - is it an instance of ethnic group?.--Pharos (talk) 18:55, 16 February 2017 (UTC)
There was another related discussion here at WD:PC (archived here). Let’s put it like that: we know that the situation about ethnic group items and property ethnic group (P172) is not good, but right now there is unfortunately nobody willing and/or capable to solve this problem. If you’d like to work in this field, I think nobody will complain. You also might find a suitable WikiProject at Wikidata:WikiProjects (or found a new one, to provide documentation of the results). —MisterSynergy (talk) 20:18, 16 February 2017 (UTC)
For wikidata to try to make sense of people's sense of community, self, ethnicity, ... is going to be a challenge as it is full of opinion, prejudice, history, ... In a world of air flights, refugees, asylum-seeking, immigration, one wonders whether it is an old world view if we are applying it to the now; though probably realistic for dealing with the old world. All I can say is good luck.  — billinghurst sDrewth 22:46, 16 February 2017 (UTC)

For proposed data, I don't mind. In fact, itis good. In this world, I agree with billinghurst. There will be massively different opinions on what defines self-identity. MechQuester (talk) 18:34, 18 February 2017 (UTC)

I've made attempt to start a guideline at Wikidata:WikiProject Ethnicity, your input would be welcomed. I see this as more figuring out how to treat types of ethnicity in a meta way, rather than trying to categorize individuals.--Pharos (talk) 05:20, 19 February 2017 (UTC)

Single value constraint[edit]

The property Thailand central administrative unit code (P1067) has as one of its constraints the single value constraint (Q19474404). However, what it actually would need is a constraint which checks that there is only one value in the highest used rank - there can be only one current id for each administrative subdivision, however there can be several old discontinued ids. For example, Mueang Bueng Kan (Q476000) has the current id 3801 and until 2011 had the id 4303, and as the constraint checks over all ranks this is listed as one of the constraint violations. Any ideas what to do, other than ignoring the violations or removing the constraint completely? Ahoerstemeier (talk) 17:46, 16 February 2017 (UTC)

+1. I can't offer a general solution, but I do agree that constraint checks (especially single value constraint (Q19474404)) should ignore deprecated statements. (Or at least that should be their default behaviour). I don't know how to fix the central constraint checker, but it should be easy enough to write a SPARQL query that only looks at the wdt: values. Jheald (talk) 23:34, 16 February 2017 (UTC)
For example, like this query. (Currently doesn't find any).
I think you can register a SPARQL query as a "complex constraint" Jheald (talk) 23:42, 16 February 2017 (UTC)
Deprecated statements really should always be ignored, as this is the way to include "false" statements. However, that'd not help in this case, as old values aren't deprecated, they are just outdated, thus having a end time and rank "normal", whereas the current value has highest rank. Thus there can be more than a single value in the lower ranks, as an example I added to Bung Khla (Q17616122) all old values. Ahoerstemeier (talk) 14:38, 17 February 2017 (UTC)

what to do about user edit-warring based on trying to force the item label to exactly match the linked enwp article title (and then on sarcasm)?[edit]

Where/how can I deal with this? New user Succu has twice renamed galaxy group or cluster (Q17512228) to "galaxy groups and clusters", first saying it was because it needed to match the linked enwp article title and that I should change that instead if I "don't like it", and then when I explained why (I believe) it doesn't work that way, he reverted me again saying "now it does". DavRosen (talk) 23:37, 16 February 2017 (UTC)

@DavRosen: User:Succu is by no means new to Wikidata since his talk page archive goes back four years; you are most certainly the newer user (for the most part) given your ~3.5 year hiatus in editing and lack of discussion on your talk page. These differences notwithstanding, I am inclined to support the singular formation you suggested since it makes statements like instance of (P31) "galaxy group or cluster" more elegant. This need not be a complaint about the behavior of other users when it could just be a discussion of plural forms in labels for items representing classes. Mahir256 (talk) 05:31, 17 February 2017 (UTC)
Succu is plain wrong. We do not keep an incorrect item just because there is an article in any language. When the article is singular, we do not need an item. Thanks, GerardM (talk) 05:51, 17 February 2017 (UTC)
Anyway, what does this article add to the Wikidata data model? Can't we just use group of galaxies (Q1491746) and galaxy cluster (Q204107) to represent what we need? instance of (P31) galaxy group or cluster (Q17512228) does not look correct anyway, be it singular or plural.Koxinga (talk) 06:49, 17 February 2017 (UTC)
I simply wasn't sure what to do next because I didn't want to engage in an edit war by reverting again, and after I pointed out that we don't need to change the en wp article title in order to change the wd label because that's not the way WD works, he did one more reversion with "That's now how it works..." which I presumed did not really mean that a policy dicussion or change had actually occurred recently, so it was by definition a sarcastic (or at least ironic) statement, and in any case gave the strong impression that he was reverting it repeatedly to force it to be the same as the article title no matter whether it was a good practice or not. That suggested to me that there was no point in further discussion with Succu on the merits of the article title, which is why I wasn't sure what to do next.
@Mahir256:, I simply mistakenly thought (in good faith) that Succu was a new user (I'd thought I'd checked) and I only mentioned it out of consideration for Succu so that none of us would "bite a newcomer", not to be dismissive of Succu's standing to disagree with us or me, nor to claim or assert how long I myself might have been editing. So there is really no need to assume bad faith on my part and so to discuss how long I, personally, have been or have not been editing WD (compared to Succu or anyone else), unless you're concerned yourself in good faith that I myself might be vulnerable to "bites", in which case... then I do appreciate the consideration.
In any case, I think we can now keep the focus solely on the question about galaxy cluster (Q204107): whether it should exist at all, and in general or in this case whether it is good WD modeling practice to use a singular label for a class of singular instances (each being a galaxy in this case), and incidentally whether to use the the "or" in the label at least when there is a union statement making it clear that any instance of this is an instance of one or the other subclass.
I'm going to revert the label back to the singular, based on the discussion above, and a third or fourth party can always revert it back again if we're mistaken, or delete it if that were the consensus.
BTW, I'm just curious, in the event that the item is ever deleted, would it be a good idea in such a case to switch the wp links to one or the other of the remaining subclasses appearing in the union statement (or perhaps to an appropriate superclass if there is one?), or just let the links disappear (which would of course remove the other-language links on the WP article as well as the wd label appearing at the top of each article in the mobile wp app) ? I have no idea what's usually done in such a situation.
DavRosen (talk) 14:14, 17 February 2017 (UTC)
As general rules:
  • No rule can force to use the article title of a defined WP. WD is not WP and has different rules (granularity, notability,...)
  • This is a general rule in ontology to use singular form for labels. Unless the item is representing a concept which can be used only in the plural form like Immortals (Q213165). Snipre (talk) 15:55, 17 February 2017 (UTC)
  • Snipre: Exactly. (BTW is there a help doc I can point others to that says this?) And do you also agree that "or" should be used in an item's label if it is made up of names of two subclasses of which the item is a union? In this case "galaxy group or cluster" Q17512228 contains a claim stating (correctly) that it is the union of "group of galaxies" with "galaxy cluster", which means it should not be "galaxy group and cluster".DavRosen (talk) 16:41, 17 February 2017 (UTC)
DavRosen External reference can help but WD should create its own set of rules based on best practises and according to its own targets as ontology. But a reference can be "Building Ontologies with Basic Formal Ontology", R. Arp, B. Smith, A. D. Spear, MIT Press, 2015, p. 61-62 (see here).
You can cite this document too, in page 12, where it is specified "A common modeling mistake is to include both a singular and a plural version of the sameconcept in the hierarchy making the former a subclass of the latter." See section 6 for more discussion.
Using only singular or plural prevent the creation of 2 items for the same concept, one in the singular form and the other in the plural form. Snipre (talk) 21:31, 17 February 2017 (UTC)
Snipre, Help:Label states nothing about naming conventions for items we regard as classes. --Succu (talk) 20:10, 17 February 2017 (UTC)
Succu, just apply the principle of heredity
* Panthera is a subclass of Pantherinae
* Pantherinae is a subclass of Felidae
* Felidae is a subclass of Feliformia
* Chordata is a subclass of Animal
So Bagheera (Q1860030) is an instance of Panthera, an instance of Pantherinae, an instance of Felidae, an instance of Feliformia,..., an instance of Chordata and an instance of animal. Snipre (talk) 21:17, 17 February 2017 (UTC)
This is not an answer to my question (naming conventions for items we regard as classes) and BTW wrong. --Succu (talk) 21:20, 17 February 2017 (UTC)
Succu What was the question ? I didn't see any question mark in your last coment. Your sentence was an affirmation and even if something is not explicitly defined it doesn't mean it's wrong.
What's wrong ? You as human are not a mammal ? You are an instance of human and an instance of Homo sapiens, an instance of mammal, an instance of animal,... and an instance of plenty of other things. Snipre (talk) 21:42, 17 February 2017 (UTC)
Really? Sure I'm an ape, but your answer did not match my observation that „Help:Label states nothing about naming conventions for items we regard as classes.“ Bagheera (Q1860030) is an instance of fictional panther (Q21009855) and we can't model taxonomic hierarchies as a monohierarchy--Succu (talk) 21:53, 17 February 2017 (UTC)
Succu Ok, so try to show what is wrong if you replace Bagheera (Q1860030) with Leopard of Rudraprayag (Q3393547) ? Is it still wrong ? And by the way can you show me how you can define yourself as ape ? You are an instance of Homo sapiens and ape are instances of Pan paniscus (for example) so you aren't an ape. But you are an instance of Hominidae and ape too. Snipre (talk) 22:14, 18 February 2017 (UTC)
So again you omitted my Help:Label problem which is responsible for this threat, not very helpful.
What happens when some new research finds out that leopard (Q34706) should be place in the new genus Neopanthera? Is it still correct to state that the individual Leopard of Rudraprayag (Q3393547) is an instance of Panthera pardus?
The usage of common names is allways problematic as your other question shows. So how do you define the term „ape“ (=„Affe“)? Should it applied to simian (Q5452918) or Hominoidea (Q102470)? I think declaring me to be an instance of (individual, particular) bonobo (Q19537) would be ridiculous.
--Succu (talk) 23:27, 18 February 2017 (UTC)
I think Snipre is right about the principle, and Succu is right that it would be very helpful if this were documented in the Help. Something like this for starters?
  • The label of a class should be written as if it described a singular generic instance of the class being labeled. E.g. the class of all stars should be called "star" as its label. And if it were desirable to have a class describing all beeps as well as all burps then it could be called "beep or burp" because each instance is indeed either a beep or a burp.
DavRosen (talk) 21:37, 17 February 2017 (UTC)
If the item in question is a merge (Q17853087) / union (Q185359) why should that labeled with or which denotes that one part is optional? --Succu (talk) 22:02, 17 February 2017 (UTC)
That's not what "or" denotes here. "galaxy group or cluster" is is literally nothing but a label that correctly applies to every instance of the class individually. Every individual instance of the class is a galaxy group or a galaxy cluster (perhaps both but that's a separate question of whether it's a distinct union or merely a union). There might be no instances (class might be empty) if each instance was required to be a galaxy group AND a galaxy cluster.DavRosen (talk) 22:17, 17 February 2017 (UTC)

WikiProject Germany has now a participant list[edit]

I created a participant list for WikiProject Germany following the example of Thierry Caro creating lists for the Anglophone countries.

I think it would be worthwhile if we have Wikiproject participant lists for most countries to be able to ping relevant people. Especially if you speak languages that aren't well represented, I encourage you to create a Wikiproject. ChristianKl (talk) 09:10, 17 February 2017 (UTC)

Yes, this would be helpful. Thierry Caro (talk) 11:14, 19 February 2017 (UTC)

Property for annual rainfall[edit]

Hi, is there a property for the annual and average rainfall of a location? The closest I could find was precipitation height (P3036), but the description reads like it's meant for storms or weather events rather than locations. Icebob99 (talk) 14:32, 17 February 2017 (UTC)

I don't think we have a property for that. It would need to be proposed. ChristianKl (talk) 07:28, 18 February 2017 (UTC)

Tools to add multiple instances of the same claim[edit]

I have a problem. I'm wanting to add a series of claims like this to various items:

They need to be separated like this so that the dates and constituencies line up - as we can't nest qualifiers, this is currently the only way to do it. Adding them manually works, but I'll probably want to do this for ~10k items so manually won't scale.

Unfortunately, if I use QuickStatements, it skips the second set of qualifiers. QuickStatements2 adds all the qualifiers to the first P39 instance, which also doesn't help much. Are there any other tools out there that can do this? Andrew Gray (talk) 20:20, 17 February 2017 (UTC)

#Pywikibot without a doubt. Multichill (talk) 21:07, 17 February 2017 (UTC)
Excellent, thanks - I'll look into this :-). Andrew Gray (talk) 12:28, 18 February 2017 (UTC)

How to handle apocryphal people and biographies, similarly apocryphal events and their accounts[edit]

Through transcribing A biographical dictionary of eminent Scotsmen (Q19020079) it is evident that some of the articles are doubtful for actual people. I am looking for guidance how we should be dealing with the subjects of such articles. For people they should not be "human", as they did not exist, we could call them "fictional character" though that doesn't ring true to me either. As apocryphal works can be about people (apocryphal person), or events (apocryphal event), or we could amalgamate and have "apocryphal subject", "doubtful subject" or something someone else can think to cover this matter. Also, as these biographical materials given detail about the supposed person, how much do we see that we could or should have? Thanks for your thoughts.

The first set of biographical articles will be added when I finish the first volume (soon), and I feel that I need to have a "main subject" link to something, especially where some of these identified had the same issue in the DNB00.  — billinghurst sDrewth 02:44, 18 February 2017 (UTC)

How about simply creating a new "apocryphal person" item? ChristianKl (talk) 07:17, 18 February 2017 (UTC)
definitely an option considered, though that would lead to people adding dates, and other classifying items, and that may not be the communities wish, hence why the issue is being mentioned.  — billinghurst sDrewth 07:37, 18 February 2017 (UTC)
There is human who may be fictional (Q21070568), which may be what you need? It is already used for characters in the Bible for example.Koxinga (talk) 08:48, 18 February 2017 (UTC)
Thanks. So different item, otherwise treat the same as a human.  — billinghurst sDrewth 09:42, 18 February 2017 (UTC)
Ah. Duplicate of Koxinga's comment. Sorry. Jheald (talk) 12:38, 18 February 2017 (UTC)

@billinghurst: I have created a new item: apocryphal biographical entry (Q28801354), and made it a subclass of both fictitious entry (Q1417174) and human who may be fictional (Q21070568). Hopefully this can be of some use here!--Pharos (talk) 04:03, 19 February 2017 (UTC)

@Pharos: I am not certain that it is helpful. While I am happy to label the underlying person as apocryphal as that is the evidence from later analysis by scholars. The article itself is still a biographical article in a biographical work, and I do not see that it is our call to make a differentiation of a published work. Maybe, (and I am not convinced) we could qualify the item for "biographical article", putting something else is another matter, and not one that I wish to do.  — billinghurst sDrewth 04:28, 19 February 2017 (UTC)
@billinghurst: I'm confused, are you creating Wikidata items for every individual entry in the biographical dictionary, rather than an item for the dictionary or for the people themselves? Can you link an example?--Pharos (talk) 04:56, 19 February 2017 (UTC)
I will be creating an item for each entry, as each entry is a subpage at enWS. This has been the practice for many biographical and encyclopaedic works for numbers of WSes. The work has its own item. Picking an existing biographical article at random as an example Church, Richard William (Q19037309).  — billinghurst sDrewth 05:17, 19 February 2017 (UTC)
@billinghurst: Thanks for the information on Wikisource subpages, wasn't fully aware of that system. Wouldn't a questionable entry in that case just be an instance of fictitious entry (Q1417174) (in addition to being biographical article (Q19389637))?--Pharos (talk) 00:22, 20 February 2017 (UTC)
Fictitious would be a deliberate act, that is a big call. That modern day sources, or lack of access to older sources, are unable to confirm a person's existence doesn't make an entry fictitious. Am I, a humble transcriber of other people's work, not an authority, able to make that judgement call? The article is the article as it was printed at that time, it is what it is. Maybe I could follow your process if we have a process to dispute any published article recorded within WD where it has been authoritatively proved wrong (sourcing required), or other interpretations have followed (sourcing required), it still is opinion/interpretation that you are putting on against an item, and that is not related to the item. It is a slippery slope to do that to articles. The person item is a different matter, that is an amalgam of information that has a chronology, and is able to show and represent disputative information.  — billinghurst sDrewth 04:10, 20 February 2017 (UTC)

How many new users get auto-patrolled every month?[edit]

I'm thinking about the suggestion made by @Alexmar983: to manually give users auto-patrolled status along with a personalized human-written welcome message. Does anybody know how many people get this status every month? (To estimate the amount of work of giving the status along with a manual welcome message. ChristianKl (talk) 18:40, 18 February 2017 (UTC)

Per Wikidata:Requests for comment/Restructuring of the "minor" user rights, all and only autoconfirmed and confirmed users are autopatrolled. Autoconfirmed users are users registered for 4 days and had 50 edits. Admins may give any user confirmed flags if trusted.--GZWDer (talk) 18:48, 18 February 2017 (UTC)
the problem is that you are giving autopatrolled flag to people who simply have 99% (95%? 90%? i don't how how to crunch the number...) of the time simple "interlink" edits. Those people are sometimes unaware of the fact that wikidata has IDs, labels and so one. I have problems with long-term users that criticize me when I start to explain wikidata from its role of centralizations of links but that's the only aspect they usually know. When they start to make other changes (for example they try to understand how a local infobox connected to wikidata works) who is there to teach them? I am no stressing the "they can make a mistake" part. I also did a stupid mistake of a image few weeks ago and I corrected after few hours (and noone noticed BTW). I am sorry about the opportunity of learning from mistakes that is lost. They could do so much more if someone see them. They can understand that their work in ns0 somewhere else is on that item too. And once they understand they really care. One of the reason why I have been introducing wikidata with the image maintenance tools is mainly because it was something effective (it is a property like the other ones) and not so risky if you make a mistake (an improper image is not like a partially wrong data, it is much easier to spot). Still, many of the newbies I found despite being on a wikidata-centric platform (itwikipedia makes massive use of wikidata) and showing 100s edits here simply knew nothing. We are not a platform that should assign an autopatrolled right based on simple edit count, IMHO. --Alexmar983 (talk) 04:31, 19 February 2017 (UTC)
To know your numbers you are going to need to track Special:Statistics, and I don't know if you can (back)track when rights are allocated, something might show through the API.

To amend the autopatrol cut-over it take a consensus discussion here. I see that the previous change request for rights change is is phab:T58203. Presumably the defaults can be changed though it is not evident that any wikis for editing use anything but the defaults of time/no. of edits. I see that Commons restricts uploads to different parameters for autopatrols. Rather than automatic, you might prefer that administrators are able to assign the right, it is how we manage it at enWS. At enWS as we have raised the bar to edits, we patrol with a bot utilising pywikibot's; though having a traditional setup of namespaces, and use of Special:PrefixIndex gives us a good degree of control.  — billinghurst sDrewth 06:48, 19 February 2017 (UTC)

There are two distinct issues:
(1) Does our present system set people on autoconfirmed, who don't know Wikidata well enough for their edits not getting the patrol flag.
(2) Could we personally welcome users when we give them the autoconfirmed status in a way that's makes it more likely for the user to become are more active user of Wikidata.
I personally don't have a firm opinion on (1) but (2) seems promising to me. A personal greeting that informed by the edits the user made and that gives helpful suggestions could potentially increase our user engagement positively. Especially given that we have 2% of the users doing 98% of the edits, getting more engagement is very valuable. ChristianKl (talk) 21:31, 19 February 2017 (UTC)
I made some tests with non it-N users not with images but with labels. I told them "look, you know language X. We don't have a lot people who knows X, when you add a link here what if you open the item and add the label X yourself?" Of course it is like the images... we really don't need labels and aliases as we need properties, but it is something easy, and they get the idea. So yes you can welcome users with only "interlink edits" and propose simple tasks. And see how they react to the proposal and than later give the autopatrolled. But in any case I strongly advice to put a much higher automatic threshold for the autopatrolled flag. I am sorry for the stress of patrollers but those people if they only edit connecting articles from local platform are not aware of what wikidata is. But the good news is... they can learn fast.--Alexmar983 (talk) 03:02, 20 February 2017 (UTC)
It sounds great that you do this. When it comes to changing the way the actual autopatrol flag works we likely need a request for comments to do it. It might be worth to have a Wikiproject for sharing best practices for greeting newcomers. How about starting a Wikiproject Welcome? @Lea_Lacroix_(WMDE): as Project Manager Community Communication for Wikidata. This topic might also be worth getting investigated by the WMDE UX team. ChristianKl (talk) 09:22, 20 February 2017 (UTC)
Thanks for the idea, it's something I already thought about. I can definitely provide some help and support, also from the rest of the team, but it's better if the project is started by volunteers :) Lea Lacroix (WMDE) (talk) 09:26, 20 February 2017 (UTC)
I do think translated labels are valuable. In small languages having translated labels means that the Article Placeholder tool can serve pages. ChristianKl (talk) 09:33, 20 February 2017 (UTC)
Pictogram voting comment.svg Comment running a bot to add welcome messages is done on numbers of wikis, and as can be expected there are upsides and downsides to a process, one of the downsides is welcoming spammers, abusers and vandals. At enWS we do it manually though we fortunately only have to worry about one language, and it looks to be appreciated by users as we regularly get 'thanks' messages, or even personal messages. At metawiki it is done with a bot, and it is a multilingual message, no idea if the bot gets thanked. Adding welcomes takes a little time, though you can make it easy with either a gadget or use of m:TemplateScript. What sort of impression and sense of community are you wishing to achieve?  — billinghurst sDrewth 13:11, 20 February 2017 (UTC)
I don't think that a bot who posts a general message that's independent of the user gives the user a feeling of being welcomed. I personally feel annoyed when I get automated bot welcome messages from 50 different Wiki's most of whose languages I don't know. It might be very good to give different people different messages. A person who just adds interwiki links encouraging them to translate labels and tell them that it helps making the Article Placeholder work (if their language enables the feature) might be a lot better than a general message. On the other hand, if you have a user who adds data about football players it might be worthwhile to encourage them to register for the Wikiproject Football.
As far as language goes, we could have a list where new users who get a minimum of 50 edits get listed by the language of their account. ChristianKl (talk) 12:55, 21 February 2017 (UTC)

Citing Oxford DNB[edit]

I'm struggling with sourcing statements to Oxford DNB online. If the person has an Oxford DNB number as an authority, I assume individual statements still need sources. What's the best practice here? Should those be "reference URL" to the online site, or "stated in" Oxford DNB, or what? I have a bunch I did wrong so I want to fix them correctly. Thanks! - PKM (talk) 22:18, 18 February 2017 (UTC)

  • I believe the preference is to give both. (Though I'm happy to be corrected). Jheald (talk) 23:04, 18 February 2017 (UTC)
  • Help:Sources describes different source models, of which Database and Web Page are relevant here. One would typically choose the Database approach if there is a database (as in this case), but there is no obligation to do so. Mixed sources which fit both models are also possible, but somewhat expensive to add. —MisterSynergy (talk) 23:08, 18 February 2017 (UTC)
Thanks. I read that and looked at some examples. It seems that "stated in" Oxford Biography Index should work. - PKM (talk) 03:58, 19 February 2017 (UTC)
Please also add all the other information to each individual reference, as Help:Sources suggests. There are some gadgets available which help you to work in this field a little more efficiently (DuplicateReferences, CurrentDate, …) —MisterSynergy (talk) 06:17, 19 February 2017 (UTC)
I'm sorry, I don't see how to add qualifiers to a reference. Can you point me to an example of a reference to a database that is done in the manner recommended? - PKM (talk) 07:25, 19 February 2017 (UTC)
Q24005810#P569. There are three different referenced values for the date of birth of this person. Two are referenced by three different databases (which therefore usse the database source model), and another one has a reference with the web page model. The order of the claims within each reference does not matter, it is just important that all important information is there within each reference individually.
If you were to retrieve data including a reference from Wikidata, e.g. in a Wikipedia module or template, you’d need to rely on the fact that all the important source information as defined by some standard (here: Help:Sources) is found in each reference. It was very difficult for Wikipedia to build a useful source if Wikidata just said “stated in: some database”, without giving a database ID, a retrieval date, and an external database entry title (the latter is relevant since there is not necessarily a 1:1 mapping between Wikidata items and external database entries). A simple “reference URL: some URL” would be not as bad, but still clearly below Wikipedia standard. —MisterSynergy (talk) 07:47, 19 February 2017 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── @MisterSynergy: Thank you, that is just what I needed. Now I know how to proceed! - PKM (talk) 20:13, 19 February 2017 (UTC)

Called to the bar[edit]

Do we have a property for "called to the bar"? The value would be a single date. - PKM (talk) 22:21, 18 February 2017 (UTC)

You could look in to using significant event (P793) with the value call to the bar (Q5021687) and date as a qualifier. Danrok (talk) 02:24, 19 February 2017 (UTC)
Perfect, thank you. - PKM (talk) 03:01, 19 February 2017 (UTC)
@PKM: I have usually been adding them as "member of" then adding the requisite Inn with a start date.  — billinghurst sDrewth 04:35, 19 February 2017 (UTC)
@billinghurst:Would "member of" an Inn of Court imply having been called to the bar rather than just studying there? This isn't my area of specialization, would love to know. That sounds like a good solution if so. - PKM (talk) 06:56, 19 February 2017 (UTC)
@PKM: They were usually firstly students at an Inn, then members, before being called to the bar. It was/is a process where being called is the end of the 'apprenticeship' and qualified. Poke at some of the finished transcriptions in s:Men-at-the-Bar (1885) to see numbers of (historical) examples.  — billinghurst sDrewth 07:24, 19 February 2017 (UTC)
@billinghurst: Thanks, will do! - PKM (talk) 07:28, 19 February 2017 (UTC)

described at URL (P973)[edit]

Should this property be used for any URL that discusses the item? For example, in the case of a person, should all online biographies of that person be included in this property? Or should it be applied more narrowly? Nikkimaria (talk) 01:03, 19 February 2017 (UTC)

the description on the property's talk page says "This is to be used to provide links to external resources that are not the item's official website, when no relevant "authority control" property exists (for instance, because the website is too messy)"


As user:JayWacker points out on the former's talk page, artificial light (Q138502) and lighting (Q210064) look very similar. However, links to different articles on the German, Spanish, Italian and other Wikipedias. Perhaps they need better English labels and descriptions? Can someone who speaks one or more of those languages please oblige? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:58, 19 February 2017 (UTC)

  • From German perspective: lighting (Q210064) seems to have a proper label (could also be illumination, but this is just another term for lighting according to the enwiki article and is already set as an alias); artificial light (Q138502) would probably be better described by artificial light. The dewiki article describes it as contrary to daylight (Q16491). —MisterSynergy (talk) 10:17, 19 February 2017 (UTC)
    • I've followed the lead of MisterSynergy, renamed one of the items to 'artificial light', and tried to add appropriate properties to distinguish them further.--Pharos (talk) 19:56, 19 February 2017 (UTC)
  • In what sense is artifcial light the opposite of daylight? Isn't the opposite of light darkness? ChristianKl (talk) 08:46, 20 February 2017 (UTC)
  • Now I see your point—I did not watch these items. The concept of “opposites” does not have a clear definition which allows to contest a given opposite relation for correctness. The distinction between artificial light (Q138502) and daylight (Q16491) has basically a technical motivation, not a physical one; we can thus neglect that there might be other, technically not relevant parts of light in an another physical context (e.g. extraterrestrial radiation). But here both types complement to the total light (Q9128), while being in fact totally disjunct types of light. Yes, an opposite of (P461) relation seems appropriate to my opinion. —MisterSynergy (talk) 22:52, 20 February 2017 (UTC)

British History Online VCH ID[edit]

As can be seen at Wikidata:Property proposal/British History Online VCH ID, Andrew Gray, Charles Matthews, James Heald and I were recently involved in a cordial discussion about how best to represent topics mentioned in Victoria County History (Q7926668), for which I had proposed a property.

To my surprise, before we reached a conclusion, this was marked as ready and, shortly afterwards, British History Online VCH ID (P3628) was created.

Two points arise:

  1. Please can property creators and others be more mindful of ongoing discussions?
  2. What's the best solution regarding VCH? Please discuss at Property talk:P3628#Way forward

@YULdigitalpreservation, ArthurPSmith, ChristianKl: also. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:16, 19 February 2017 (UTC)

I'm sorry. The discussion looked to me like it reached it's conclusion. User:Jheald answered the open issues and nobody opposed the property in it's current form. ChristianKl (talk) 17:14, 19 February 2017 (UTC)
Note I marked it as ready after reading JHeald's "Latest view" comment, it seemed the discussion had concluded. ArthurPSmith (talk) 16:28, 20 February 2017 (UTC)

Commons data file property[edit]

It seems to me that there is no property to link to the data files in Commons, the .tab JSON files and the .map GeoJSON files? If that is the case, I could propose a property for at least the map file. – Susanna Ånäs (Susannaanas) (talk) 15:14, 19 February 2017 (UTC)

The corresponding datatype to link to the data namespace on Commons is not yet available. --Pasleim (talk) 14:35, 20 February 2017 (UTC)
@Pasleim: Is that datatype expected to be coming soon? Or would a property with datatype:string and a URL formatter be useful in the meantime? Jheald (talk) 21:16, 20 February 2017 (UTC)

Items for school year & academic term were merged[edit]

Q1757373, which did have the English label "school year", has been redirected to academic term (Q915466), which has the English label "academic term". Given that there are (usually) three terms in a year, should they be de-merged, or were the English labels wrong?

@Infovarius: FYI. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:55, 20 February 2017 (UTC)

I think it should be de-merged. In German "Schuljahr" wouldn't be used for an university. ChristianKl (talk) 15:47, 20 February 2017 (UTC)

Wikidata weekly summary #248[edit]

executive body (P208): "government" or "cabinet"?[edit]

The documentation for executive body (P208) (Property_talk:P208) gives as its example Israel (Q801)Cabinet of Israel (Q2578249). However, it also has an "Allowed values" constraint of executive branch (Q35798), and Cabinet of Israel (Q2578249) (as well as most other countries where the Cabinet, rather than the Government is set as the value) fails this check (Wikidata:Database_reports/Constraint_violations/P208#Value_type_Q35798), as the item is usually a subclass of cabinet (Q640506) (with individual Cabinets being instances of that), rather than a direct instance of cabinet (Q640506).

From what I can see, many of these countries don't have a separate item for the national government, distinct from the Cabinet, but some countries do: e.g. the UK has Government of the United Kingdom (Q6063) of which Cabinet of the United Kingdom (Q112014) is a part, and United States Cabinet (Q639738) is a part of Federal Government of the United States (Q48525).

Surveying how this is used generally isn't a lot of help — most countries don't have executive body (P208) set yet at all, and of those that do, there's a slight preference for the "cabinet" rather than the "government", but the numbers are small enough for it to be inconclusive:

SELECT DISTINCT ?country ?countryLabel ?body ?bodyLabel ?parentLabel
  ?country wdt:P31 wd:Q6256 .     # find instances of country
    ?country p:P208 ?statement .  # with a P208 (government) statement
    ?statement ps:P208 ?body .    # which we assign as ?body
    FILTER NOT EXISTS { ?statement pq:P582 ?end_date }  # with no P582 (end date) qualifier
    OPTIONAL { ?body wdt:P279 ?parent } # and set ?parent based on either subclass
    OPTIONAL { ?body wdt:P31 ?parent }  #                          or instance of
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
ORDER BY ?countryLabel

Try it!

I'm assuming that it's generally a bad idea if the example given for a Property fails one of its primary constraints. Is the solution here to (a) change the constraint to allow a Cabinet (either exclusively, or in addition to a Government); (b) change how cabinets are modelled, so that they can be set here without this constraint failing; (c) make sure this is set to the Government item, not the Cabinet item in each country (creating the relevant items where needed); (d) something else entirely?

Or from another angle, which should we be saying?:

< United States of America (Q30) View with Reasonator See with SQID > executive body (P208) See with SQID < United States Cabinet (Q639738) View with Reasonator See with SQID >


< United States of America (Q30) View with Reasonator See with SQID > executive body (P208) See with SQID < Federal Government of the United States (Q48525) View with Reasonator See with SQID >

--Oravrattas (talk) 16:46, 20 February 2017 (UTC)

well it sounds like at least the subclass/instance modeling here is wrong. But I'm not sure what else I would recommend. ArthurPSmith (talk) 17:10, 20 February 2017 (UTC)
Pictogram voting comment.svg Comment Oh, it looks like we have a mixture of terms in place, the synonym here needs to be removed. The body here is meant to be government and it is meant to exclude branch. Looking at what is around, a nice definition is
the executive (the Australian Government) is responsible for enacting and upholding the laws established by the legislature. Certain members of the legislature (called ministers) are also members of the executive, with special responsibilities for certain areas of the law.
and it separately mentions the legislature and the courts as the other two arms.  — billinghurst sDrewth 22:42, 20 February 2017 (UTC)
as a slight aside, that definition is quite specific to the British/Australian-type system. Much of the world maintains a much stronger separation of powers, where a minister cannot also be a member of the legislature. --Oravrattas (talk) 09:55, 21 February 2017 (UTC)

Henry Elsynge confusion[edit]

How should we proceed when most of the authority control records seem to be wrong? In the case of the British parliamentary historian Henry Elsynge (Q18730405) (father) and Henry Elsynge (Q15072637) (son), authorities seem to be following the 1889 Dictionary of National Biography in giving the son's dates as 1598-1654 and attributing a series of works on Parliament to him. The current online DNB gives dates of 1606-1656 for the son and explicitly says "The various parts of the treatise entitled The Manner of Holding Parliaments in England, or, Modus tenendi parliamentum apud Anglos, sometimes attributed to Elsynge, were in fact the work of his father and namesake." Even the 1768 edition of this work in Commons points out that attribution to the son is incorrect and it should be the father, and I have some recent scholarly papers that say the same.

I am happy to deprecate the old dates and the authorship with the more recent info, but do we need to do anything about the authority controls which seem to be associated with the wrong person?

For now I've written an article for EN:WP on the father and tagged the article on the son as disputed - if there's no disagreement I'll make the changes there, but I'd prefer to get WD sorted first. - PKM (talk)

@Andrew Gray, Charles Matthews, Jheald: FYI. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:22, 21 February 2017 (UTC)

Instance of "type of cheese"[edit]

blue cheese (Q746471) is a subclass of (P279) cheese (Q10943).

But it is also an instance of (P31) type of cheese (Q3546121). (English article: Types of cheese, French article is about types of cheese pastes, Italian article is about cheese classification)

Should I remove this second statement? It seems redundant and type of cheese (Q3546121) does not seem like a solid item to build upon. Thanks! Syced (talk) 03:27, 21 February 2017 (UTC)

No, there's no reason to remove valid data. Wikidata's philosophy is that it's okay if different people categories an item slightly differently and seeks to present both perspectives. ChristianKl (talk) 09:16, 21 February 2017 (UTC)

Displayed name of Noongar language[edit]

Nys language label in Wikidata

If I expand "In more languages" for Wedge-tailed Eagle (Q752465), the Noongar language label is itself labelled as "nys", rather than in full (see screenshot, above). Where and how can this be fixed? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:18, 21 February 2017 (UTC)