I need an advice regarding links to local Wikipedia images in Wikidata. If the image is free then we put it on Commons and add a link to this file through P18 property. But if the file is protected by copyright (f.e. because of NoFoP in the country), we can upload it only on local Wikipedia site. But anyway I need to create a link to the image in the Wikidata item. How can I do it? --Voll (talk) 21:02, 28 April 2016 (UTC)
Any official, semi-official or unofficial workaroud for such want (non Commons image)? Like "No value" and "Unknown value" for datavalues. --Voll (talk) 08:52, 29 April 2016 (UTC)
Links to non-free pictures on random wmf-wikis has been voted down in our talks here. They are useless for our purpose. But url-links to repositories with non-free images is maybe possible. We already have many links to databases with such images. -- Innocent bystander (talk) 09:02, 29 April 2016 (UTC)
When improving the statements about an item we might introduce redundancies that might hinder future refinements. For example, consider an item that I recently edited:
When I found this art gallery it had a country (P17) statement with the value United States of America (Q30), which is certainly true. Indeed, a great many nascent entities in this project have helpful country (P17) tags like this.
Once we are done adding P131 to all items that could have it, I suppose it is. In the meantime, it allows to do simple checks. --- Jura06:07, 30 April 2016 (UTC)
A bit of redundancy is no problem so please don't remove country statements as redundant. It makes it easier to find things. Multichill (talk) 13:50, 30 April 2016 (UTC)
I think I wouldn't exaggerate if I claimed that most localities on Earth belonged to more than one country during the time of their existance. Borders are forever changing, countries come and go (as do settlements and corporations). Library of Alexandria (Q435) was always part of Alexandria (Q87) but never really part of Egypt (Q79) (an claim currently actually added to the item, luckily I possibly found a perfect trial case). It was part of Ptolemaic Kingdom (Q2320005) and later Roman Empire (Q2277) but never part of modern Eqypt, so country (P17) is certainly useful here if you want to be exact. DGtal (talk) 09:48, 1 May 2016 (UTC)
I think we sometimes have statements like "Library of Alexandria" located in: "Modern Egypt", but such statements should not be among the preferred statements. -- Innocent bystander (talk) 10:40, 1 May 2016 (UTC)
greek government items : duplicates ?
I'm not an expert in greek governement, but Q23836182 and Q23836195 seems to be duplicate items - vice ministers of the same(?) minister. Anybody can confirm ? I could have missed something. One minister : Q23836195. Only one other example of ministers, so it might be a change in the minister itself ... author TomT0m / talk page11:35, 1 May 2016 (UTC)
I've just completed transcluding and proofreading A Naval Biographical Dictionary (NBD) and it has been suggested that some or all of the name entries should be added to WikiData.
I have no idea how to do this, what I do have is a spreadsheet with around 5000 entries of this kind:
NBD start page: 1354
NBD end page: 1354
name: Charles Bampfield Yule
NBD article name: Yule, Charles Bampfield
Wikipedia article: Charles Bampfield Yule (or x if no article)
See this example article, they vary in length from a few lines two a page or two.
@GreyHead: a sample with basic steps can be found at WikiProject_DNB#Examples. Essentially, you will get an item for every article page and link these to/from the item for the person. If the person doesn't exist yet at Wikipedia (or at Wikidata), you can create an new item. It should be fairly straight forward. I will try to update the list of tools on the project page. --- Jura13:22, 1 May 2016 (UTC)
@Jura1: Thanks, that appears to be for the DNB - Dictionary of National Biography - a different source? Are those tools still useful . . . and is there any way to use a spreadsheet or CSV file as a source? GreyHead (talk) 15:31, 1 May 2016 (UTC)
standard structure of the item
Hello. I have a very detailed and not very urgent question for you.
When I explain to newbies the items on data I kinda point out that there are four main sections: the description table at the beginning, the "statements", the "identifiers" and than the "wikilink or sitelinks section". It looks quite a natural "first step". I am talking here about "Q" items, or in general itms with a clear connection to the main namespace of a wiki platform.
Is it just me or the page Help:Items lacks of a clear overview of this concept? You kinda see in the "see also" at the end and a little in the text, but shouldn't we insert some schematic picture for it? Like a general coordinate on how to read the whole page th first time when you know nothing.
Also, can someone please explain why the "sitelinks section" does not have a title "sitelinks" and the first one does not have a title like "descriptions and labels"? It is not very important but it would really provide a faster orientation on the basic concepts. Sometimes when I answer to questions I feel a certain lack of homogeneity. For example I can use a direct link to a section, but I have always to describe the first table as "just under the main title" instead of linking it, and for the sitelinks area, I use the wikipedia links for example, but it is bizarre that I cannot use a link with just "sitelinks".
Nothing critical, but I just wanted to understand if there is some specific aspect I am missing. You know what they say, the devil is in the details--Alexmar983 (talk) 14:27, 1 May 2016 (UTC)
It might have been written before there were "identifiers" or even "statements". Thus the (IMHO over-weighted) part on the use of descriptions. As it's translated into 100 languages, it was probably easier to do a separate page on statements. If you have experience in explaining Wikidata to newbies, you might be a good candidate to re-write it. --- Jura15:16, 1 May 2016 (UTC)
ok I 'll think about it. I will make some suggestion for a better introductory part. Not urgent, BTW. Please if anyone had any issue with newbies explaining some concept, let me know...--Alexmar983 (talk) 10:47, 2 May 2016 (UTC)
Wikidata weekly summary #207
Here's your quick overview of what has been happening around Wikidata over the last week.
Wikidata's coverage as of April 2016Difference of Wikidata's coverage between April 2016 and 6 month before
Substantially reduced server load for item and property displaying (phab:T132662)
There are currently some issues with the order of latitude/longitude inn coordinates in the query service map visualization. It will be fixed tonight.
Removed unsupported sort and dir parameters from the wikibase.api.RepoApi JavaScript module. This may break user JavaScript calling getEntitiesByPage (phab:T119856).
Worked on new flyers for institutions that want to cooperate with Wikidata and developers wanting to use our data (will be published on Commons once they're done)
Moved forward with internationalization of the query service interface (not on translatewiki.net yet but being worked on)
Worked on making it possible to extend SPARQL queries in simplified natural language version. It will also no longer add query prefixes when editing the query. Those are not live yet.
Fixed a bug where admins got a blank page when trying to view deleted revisions (phabricator:T132645)
Investigating issues with bad suggestions for properties when adding new statements (phabricator:T132839)
Many Wikipedias have articles for both. There's been a lot of debate over merges, but so long as at least one Wikipedia has a separate article for each, we need to maintain separate items. --Yair rand (talk) 20:46, 1 May 2016 (UTC)
I'm using Dynamic Wikidata Lists to create a kind of multilingual Wikiproject work list for Biosphere Reserves. Its working really well except I'm unable to add the official URL of the site (Property:P2520) into the table for some reason, it just says CLAIMS rather than displaying the url or providing a link. Does anyone know if I'm doing something wrong or if it simply doesn't allow this for some reason?
Qualifiers for official urls and other urls where more than one language version exists
Hi all
I'm looking for some guidance on adding multiple urls to the same property (Property:P2520 UNESCO Biosphere Reserve url) on items. Some of the reserves have descriptions in more than one language (English and French) and the descriptions are hosted in different places on the website with no clear way of switching between them. My assumption is that I should include both URLs with a qualifier for each for the different languages, is this correct? If so which qualifier should I use?
equivalent class (P1709) is an URI that points to a similar class in ontologies. Is this an identifier for the class ? No emergency to answer but this comes to my mind because of
Update: I now see what was going on -- I had added those values previously, and QuickStatements was refusing to add them again, even though they were now qualified by a source. -- The Anome (talk) 22:16, 4 May 2016 (UTC)
@Edgars2007: Thank you! I will change that. First, I'll go back over my previous changes and remove them, ready for the new, improved statements. -- The Anome (talk) 22:26, 4 May 2016 (UTC)
@Edgars2007: Ah. I've just realized that (a) QuickStatements can't remove statements, only add them, and (b) I don't have rollbacker powers here. Is there some easy way to make these 100-or-so edits go away so I can add them again correctly? Or just to add the correct references, since the data itself is I believe valid? -- The Anome (talk) 22:39, 4 May 2016 (UTC)
For WOEID you can probably just remove wrong statements with Autolist2 instead of using rollback. You can use http://regexr.com/ to extract items from your user contributions page if you did not save your QuickStatements tabular data. --Lockal (talk) 08:22, 5 May 2016 (UTC)
Today is African World Heritage Day and I have created a small project to help improve knowledge of African World Heritage Sites on Wikipedia in all languages. I'm using Wikidata as the central index for the project using Listeria, its a really amazing tool. Please get involved and also share the link (especially in languages other than English) to encourage more people to take part, there is a social media sharing section at the bottom of the page.
Research about user participation in Wikidata - call for participation
Dear Wikidata users,
We are a group of researchers of the Web and Internet Science group of the University of Southampton.
We are currently conducting a research aiming to discover how newcomers become full participants into the Wikidata community. We are interested in understanding how the usage of tools, the relationships with the community, and the knowledge and application of policy norms change from users' first approach to Wikidata to their full integration as fully active participants.
This study will take place as an interview, either by videotelephony, e.g. Skype, phone, or e-mail, according to the preference of the interviewees. The time required to answer all the questions will likely be about an hour.
Any data collected will be treated in the strictest confidentiality, no personal information will be processed for the purpose of the research. The study, which has submission number 20117, has received ethical approval following the University of Southampton guidelines.
Users interested in taking part or wishing to receive further information can contact us by writing to the e-mail address ap1a14+wikidata_user_study@ecs.soton.ac.uk
We aim at gathering about 20 participants. The recruitment will take place until Sunday, May 15.
Thank you very much, your help will be much appreciated!
Hi. Wikidata is a database and some values, once inserted are fixed and no longer require changes. For example:
date of birth and death of a person
title of a TV series
historical population at a specific date of an administrative division
length of a film
identifiers
and others, ....
Now it is not technically possible, but if we consider it useful, we can ask to the developers to implement a mechanism how described below. Now administrators can only block the entire page, with different levels of protection. But if it would be possible to block individual statements, we can lock the values that will not change over time (as the examples above). This would bring a big advantage for the patrolling and the reliability of data.
Then, in the guideline, we can decide which properties can be blocked, or just statements with references, and so on. I found no other discussion on the topic; what do you think to activate this mechanism? --β16 - (talk) 08:02, 5 May 2016 (UTC)
Above facts may not change, but statements about them may change or become more detailed. --- Jura08:09, 5 May 2016 (UTC)
It wouldn't be helpful to block adding new references or qualifiers. Occasionally it is also useful to directly modify the value, either to increase precision, or to fix a flawed import. However, monitoring such changes is very important for preventing vandalism. I proposed some months back that such edits be tagged by AbuseFilter: Wikidata_talk:Abuse_filter#Tag_changes_to_sourced_statements. --Yair rand (talk) 08:24, 5 May 2016 (UTC)
That a claim has been protected does not necessary mean that such protection never can be revoked or that such claims never can be edited. It is up to our policies to decide that. I, on a daily basis, remove false claims, even those who have (bad) sources. The bot-import we made of GeoNames ID (P1566) from the database itself (a wiki just like WP) was a really bad idea. -- Innocent bystander (talk) 10:23, 5 May 2016 (UTC)
Support I suggested that in the past. In cases of data imports, the fact that the database did had the value is a fact that won't become anytime wrong. The initial wikidata plan was to deprecate such claims if the values are considered wrong later. It might need a little common sense use not to bloat the database however. author TomT0m / talk page07:01, 6 May 2016 (UTC)
These facts are often plain wrong. I often changed dates of birth by making them more precise. Funny at that is that as a result references are no longer realistic because we differ from the sources. In the past people did not like me removing such references even though they became factually incorrect. Thanks, GerardM (talk) 09:43, 5 May 2016 (UTC)
If a statement can be blocked, not means that must be blocked. If a date is not precise (only year of birth, for example) or disputed, it should remain free modificable. But, for example, the identifier to an external database, with a precise reference, this I do not know how it can be modified.
Moreover, being a new feature, it can be constructed so as to avoid that you change the existent data, not to prevent the insertion of new information (qualifiers or references); but personally this solution of partial block does not really like; it's only a hypothesis. --β16 - (talk) 09:40, 6 May 2016 (UTC)
2 questions on saints
Churches with names of saints: is there a property to connect both? Some saints have 10 - 20 churches carrying their name.
Secondly we know very little of the early Celtic saints, very often we are limited to the century they flourished, not the exact year; yet I'm unable to note their dob/dod or fl as a wider span such as fl. = 560s or dob = c. 466 or fl. = 6th century. Llywelyn2000 (talk) 15:32, 5 May 2016 (UTC)
We have named after (P138) to connect a church with the saint it is dedicated to, but I don't know of a way to do the inverse.
But please use sources. Here in Sweden there used to be a "Municipality of St Peter and St Paul". It was actually not mentioned after the saints, but after two parishes. If the parishes are mentioned after the saints or of the churces of the parishes is beyond my knowledge. -- 78.73.94.16516:27, 5 May 2016 (UTC) (Innocent bystander)
I've now created my WOEID (P1281) dataset, starting with a very cautious subset of ~20,000 matched items. I've used QuickStatements to add some of these, and spot-checked some of the entries so far, all of which seem to check out. (See Waarschoot (Q911983) for an example.)
I'm now looking to add the other 20,000-odd, and after reading the data donation page, it looks like pywikibot is the right way to do it. If this works, I potentially have another ~100,000 more to add later once I've improved my WOEID matching code. Can anyone tell me how to apply for permission to run a bot account to make these additions? -- The Anome (talk) 16:32, 5 May 2016 (UTC)
Give the best precision. E.g., if source says "c. 1456", then write "1450s", if the source says "end of 20th century", then write "20th century". Or I didn't answer your question? BTW, question for everybody: can we have something a little bit more precise than "20th century", if we want to put "end of 20th century"? At least, something like "second half of 20th century"... --Edgars2007 (talk) 09:28, 6 May 2016 (UTC)
I am wondering how should the country of an embassy be handled? Not sure if P17 ("sovereign state of this item") should be the represented country or the host country? or list both with a qualifier? (which qualifiers?) or do we have other appropriate properties to use instead? P1001 ("applies to jurisdiction")?
Here is a query that lists all items with instanceof = embassy, with country and administrative territory: [1]
If I remember correctly diplomatic missions are legally considered an Extraterritoriality, so the embassy of A in B is an exclave/enclave part of country A. However, in practice this will basically make the item useless, so I recommend we register it as if it is part of the host country. DGtal (talk) 08:37, 1 May 2016 (UTC)
From what I've read (e.g. here), the land is normally still sovereign territory of the host country anyway. - Nikki (talk) 18:55, 1 May 2016 (UTC)
My understanding is that the embassy is part of the operating country, but the land on which it stands remains the sovereign territory of the host country. The land and building may be owned by either country or a third party (e.g. the American embassy in London is built on land leased from the Duke of Westminster [2]). Thryduulf (talk: local | en.wp | en.wikt) 22:40, 1 May 2016 (UTC)
For ownership, we have owned by (P127). Either way, I think setting P131/P17 to where the land is located makes the most sense, because that seems like the most consistent way of using it (e.g. if I wanted a list of castles in the UK, I'd most likely use P31 and P17, if I change P31 from castle to embassy, I'd expect a list of embassies located in the UK, not a list of UK embassies in other countries). - Nikki (talk) 16:34, 2 May 2016 (UTC)
Even if you have solved the problem of deciding which property to use for the host country and the operating country, you still have problems with Embassies not located on the territory of the country the ambassador is sent to. Case in point: Ambassadors to the Vatican City, almost all of which have embassies in Rome. In this case you need to have 3 properties: one for the Vatican City, another for Rome or Italy, and another for the operating country. —seav (talk) 02:39, 2 May 2016 (UTC)
Using country (P17) in that way doesn't seem like a very good solution to me because it's difficult to get that meaning from the property (we currently say that country (P17) is a subproperty of located in the administrative territorial entity (P131), which should mean that the country (P17) statement is the country it's located in). I don't think we should be afraid to add new properties if it helps us store data in a clear way that makes it easier for everyone to add and use consistently and correctly, and given that this question has come up a few times without any real conclusion, this does seem like a case where dedicated properties would be useful. - Nikki (talk) 16:34, 2 May 2016 (UTC)
Things can be a little more tricky than they first look. From what I know, the Swedish ambassador (or similar) responsible for Andorra is located in Barcelona, Spain. I have no source for that, it was something I heard on the radio in the 1990's. And I also know there is a Swedish ambassador responsible for the Holy See, but I have doubts that she has her office in the Vatican. It is most likely located in Roma, Italy. -- Innocent bystander (talk) 16:53, 2 May 2016 (UTC)
I'm thinking country (P17) isn't so good for representing this information, other than maybe just as a location property. I am not sure of a good solution, though. Maybe owned by (P127) can work for the represented country or maybe a new property for this. Aude (talk) 22:33, 4 May 2016 (UTC)
Multiple labels for different items is no problem. It is in the statements that disambiguation becomes obvious. Thanks, GerardM (talk) 11:12, 8 May 2016 (UTC)
I don't think so :/. E.g. In that cases it is very inconvenient to add statements manually when there is no description. --ԱշոտՏՆՂ (talk) 11:19, 8 May 2016 (UTC)
Writing a description could even be more inconvenient? Descriptions aren't even structured data. If there are statements, bots could add descriptions.
Obviously, it all depends on the use you try to make. I found that Duplicity generates quite useful autodescriptions and, with the presence of dates, usually allows to identify already existing items for the same person. --- Jura11:49, 8 May 2016 (UTC)
Commons Category P373 versus Sitelinks/other links that links to a commons cat
Is there any guideline where to put in the commons category info? For example Q1376960 has a statement with a property:p373 (commons cat) while Q7521354 has only a sitelink under "other side" and a link to a commons category. Are these two methods for the commons category consider equivalent? If not, in which field should be the commons category info put in?--Wdwd (talk) 15:45, 8 May 2016 (UTC)
Sitelinks to Commons are broken by design. Since Commons works very different from other projects, we would need to have multiple sitelinks to it. (Due to the fact that we have galleries on Commons, but usually it makes much more sense to link to the category.) This is not supported. Therefore we use Commons category (P373) and Commons gallery (P935) to link to Commons. --Srittau (talk) 17:01, 8 May 2016 (UTC)
Translation and templates
As you know, I'm working on a template integrated into {{Item documentation}}. This is almost done, an example in :
Talk:Q11163999.
Although I'm running to problems for automatic translation:
Although i'm using translated templates like {{Doc disjoint union}} (see Module:Requests/union_of) their expansion in the module does not appear to be translated. Is this a bug ?
unrelated, but still : I wrote {{Capitalize}} which work when used with a raw string, for example {{Capitalize|plop}} => Plop but fails when applied to a template, for example {{Capitalize|{{Int tools}}}} => Tools (uncapitalized for me).
About second. Hmm... Maybe it's because first character of (expanded) {{Int tools}} is "<"? Tried with template, which content was "test". Everything was working, so it's probably not a template transcluding issue. --Edgars2007 (talk) 18:49, 8 May 2016 (UTC)
Mm I don't think so. First it's a "noinclude" tag so having it in the result of the expansion would be weird. Second I reversed the stuffs and it still does not work. Could you put your example here to check if it's language dependant ? author TomT0m / talk page18:54, 8 May 2016 (UTC)
Wikidata query to include 'is part of' or 'has part'
Hi
So I have worked with others to add all the UNESCO Biosphere Reserves to Wikidata and created a list on the Listeria tool to provide a structure for people interested in writing about them. Within Biopshere Reserves there are very often geological features eg mountains, lakes, rivers, towns etc. I know I can link to them by using the properties 'part of' and 'has part', my question is once I have mapped these and added them to Wikidata how can I include them within the Wikidata query on my Listeria list? Currently my query is claim[463:14018439]
In Germany, as well as in several other countries, there is a tradition to designate one species of bird (and also for several other animals and plant groups) as the special Bird of the Year (Q896971). What would be the best way to add this to WikiData? I thought about using award received (P166), however that property is supposed to be for people, organizations and creative works. Would it be better to bend the definition of that property a bit to include animal and plant species (there's even Rock of the Year (Q1519491)), or would a specific property be the better choice? instance of (P31) would also another possibility, but I don't think that'd be a good choice. Ahoerstemeier (talk) 14:26, 9 May 2016 (UTC)
Another problem I ran into doing the job exposed in the previous thread (I do this here because we never talk enough of those tools :p)
@Magnus Manske: : I tried to use petscan to compute the difference beetween a category present on certain wikidata items talk pages (Category:Item_with_class_reports) with a query to see which items have union of (P2737) or disjoint union of (P2738). Problem : I had no result with petscan for the category. Is it possible to get the items whose talk page are categorised with petscan ? author TomT0m / talk page18:48, 8 May 2016 (UTC)
You can get the items with the category on the talk page like this, output it into TSV or PagePile, then use "Manual list"/PagePile to subset with the Property. Or, in one go via template inclusion (see the "Templates&Links" tab). If you do this regularly, add it to the bug tracker ("Issues") so I won't forget it. --Magnus Manske (talk) 11:57, 10 May 2016 (UTC)
Wikidata weekly summary #208
Here's your quick overview of what has been happening around Wikidata over the last week.
Events/Blogs/Press
Past: OpenTechSummitslides for SPARQL intro at OpenTechSummit
Our Main Page now has a section for popular items to show you what is trending on Wikidata based on several people editing an item over the last few days
No longer showing a restore link for inaccessible revdel'ed entity revisions – you are welcome though I think you could disclose this or this one whereas the latter had immediate effect and the former will even influence clients. Unless you are making a release tomorrow, the former can be added to the next issue. Matěj Suchánek (talk) 19:17, 9 May 2016 (UTC)
The « Jumiekan Patwa Wikipidia » is now out of the incubator and open as a new independant Wikipedia. But apparently, it's not yet possible to link them on Wikidata. Could someone do what need to be done? Plus, can a bot import all the old fashion interwikilinks from this Wikipedia ?
I'm having real trouble finding the latest best practices for linking to Wikidata in Commons. I would expect the info to be at Commons:Wikidata but it's not current - it says "More flexibility in accessing information from Wikidata is hoped to be possible in Spring 2015."
I've been pretty much away from Wikidata for the last year. I know random access has been made available, but how should it be used? And which bots to create templates from Wikidata are actually running? Help? - PKM (talk) 19:31, 10 May 2016 (UTC)
At the moment I'm a bit hesitant to add new identifiers for items on DMOZ (Q41226) because the regex that checks for well-formed links is failing for pretty much every identifier in a language that isn't English (and even for some English labels). I'm not sure that the '\w' is capturing every non-whitespace, URL-appropriate Unicode character; is there any good way to fix these false positives? Mahir256 (talk) 06:18, 11 May 2016 (UTC)
(Judging from how IDs on both sites are constructed, they don't look the same at all.)–My concern is that Unicode characters outside of regular ASCII aren't being detected as Unicode characters properly when checking for constraint violations. I am somewhat uncertain how to rectify this issue as (upon looking at it) the regex ought to capture most non-Latin characters, yet somehow it seems to fail. Whether they fail because of a lack of ending backslash is irrelevant; identifiers with characters in the same extended Latin ranges, for example, still show up as ill-formed. Mahir256 (talk) 06:53, 11 May 2016 (UTC)
i contacted them via email, their answer was:"we have given the Gene Wiki project the permission to import UniProt data into wikidata and as far as I remember they import more or less the parts you mention below (for a currently limited set of species). I would therefore recommend that you contact them to see whether you can coordinate with them to avoid that wikidata will contain multiple copies of the same data." So it's already set? --Ghilt (talk) 21:53, 9 May 2016 (UTC)
Don't think it's misplaced. Name, ICAO code, runway length and surface do all match. Only the country is different. Maybe the author of the article does know more @Darafsh: --Pasleim (talk) 20:06, 11 May 2016 (UTC)
ArticlePlaceholder is now live on first 4 Wikipedias
Hi everyone :)
Last year Lucie started working on the ArticlePlaceholder in order to fullfill Wikidata's promise of supporting especially the smaller Wikipedias. Today we have rolled it out on the first 4 Wikipedias: Esperanto, Haitian Creole, Neapolitan and Odia. When someone searches for a topic where no local article exists but Wikidata has data we will show an ArticlePlaceholder with this information and encourage the reader to create an article. I hope this will help these Wikipedias by offering their readers more content and by turning more of them into active editors.
In order for the feature to work well we need labels for items and properties in these languages on Wikidata. A lot exist already but if you want to help out you can find items and properties that need labels in these languages at https://tools.wmflabs.org/wikidata-terminator and d:Special:ListProperties.
What we rolled out today as usual is a first version. Based on the feedback from those 4 Wikipedias we will expand and improve it. One of the next things we will do is add the option to translate an article from another language if it exists using the Content Translation tool and fix known bugs. Things we know are still broken or need work:
language fallbacks in the properties are not working so you will see a lot of P1234 and so on until a label is added on Wikidata in that language
long identifiers break out of the identifier box on the right side and don't look good
right now you only get an ArticlePlaceholder in the search results when Wikidata has at least 3 links to other Wikimedia projects and 3 statements. We might need to tweak this number still based on the first Wikipedia's feedback. We limit this in order to not encourage readers to create an article that will be deleted right after they created it because it isn't notable.
With many events, e.g. earthquakes, wildfires, bombings, explosions, etc. there are often reports about how many buildings are destroyed and how many are damaged. At 2016 Fort McMurray Wildfire (Q24004490) I've tried to represent that as:
but I'm not convinced there isn't (or shouldn't) be a better way to do this, particularly is it's not always going to be clear with incremental updates which point in time qualifier relates to which quantity qualifier. Is there a need for one or two new properties here or am I missing something? Thryduulf (talk: local | en.wp | en.wikt) 22:07, 5 May 2016 (UTC)
Indeed. First the "cause of" statement links an event to a class of event, which is imho clumsy.
But "is it's not always going to be clear with incremental updates which point in time qualifier relates to which quantity qualifier" you just need to create several claims in similar issues. You're trying to put too much information in one claim.
I'd suggest a property for damages however. Properties in
Hi,
I noticed that in the entry on Dante there is a mistake. The entry lists among authors that influenced Dante, also Andrew Davidson. How can delete this wrong info?
I am new here. I would really appreciate your help.
Thanks,
Mbegali (talk) 21:03, 8 May 2016 (UTC)
I've added some new features to User:Yair rand/DiffLists.js. The script now also works on History and Contributions pages, and has some more advanced filtering options. It is now possible to filter out identifiers, or only show links to projects in particular languages, or projects regardless of language.
Watchlist demo, with filters set to hide non-English label changes and identifiers.
I'm trying to find some Wikiprojects with simple sets of relevant properties and items, and use the script to make some Wikiproject-specific recent changes systems using {{Wikidata list}} and Special:RelatedChanges. Suggestions would be most welcome. --Yair rand (talk) 08:30, 12 May 2016 (UTC)
Mergers limitation
In recent time i am fixing lot of incorrect mergers. If incorrect merger is not detected in time, they can be very hard to fix, as new data related to both concepts are added.
These incorrect mergers are usually performed by inexperienced user. I do not call for for granting merge rights to admins only again, but i think that we can consider some limitation for merge rights, for example to grant them at least to autoconfirmed users. --Jklamo (talk) 11:28, 7 May 2016 (UTC)
Well, would not the result of that be that they move sitelink by sitelink instead? I see a lot of very disruptive behaviour from very well-experienced WP-users from time to time. They see WD as a repository for interwiki and has no insight in the purpose of this project. -- Innocent bystander (talk) 07:29, 8 May 2016 (UTC)
Moving sitelink by sitelink may casue less, harm as at least incorrect aliases/labels/descriptions will not be moved. I agree that problem is merging based on sitelinks (while ignoring properties/aliases/labels/descriptions). --Jklamo (talk) 16:30, 8 May 2016 (UTC)
Defining the age of a group of people within a qualifier
Hi all
I'm trying to work out how to add data that includes an age range of a group of people. E.g one of the values will be for Out-of-school children of primary school age in pre-primary education in Antigua in 1998
I plan to use the property 'Out of School Children' and the qualifiers for Antigua and 1998 are clear, however could someone explain how I could add a qualifier for 'of primary school age'?
I would recommend using the generic "of" qualifier with the item "primary school". Is that sufficient for your needs? --Izno (talk) 19:24, 9 May 2016 (UTC)
I'd like to add a page with the details of our business Action Victoria.
We are a business coaching organisation in Melbourne Victoria Australia.
I'd like to add the basic business details like name, logo, contact details, website and social media listings for our business.
Can anyone suggest the data that I should be uploading?
Hello @Shelleykknox: the first step would be to check our notability policy to ensure that the item is something we want to include here on Wikidata. If it meets the policy, then you are free to create the item and add all of that information. Please let me know if you have any questions about whether or not your business is notable. Ajraddatz (talk) 02:52, 11 May 2016 (UTC)
I generally agree. However, if they can meet the burden of evidence required to create a notable item here, then really that's a benefit to the project. Ajraddatz (talk) 08:56, 13 May 2016 (UTC)
University professor
Hoi, as far as I am concerned, this is a title and not a job. It is not that you work or teach at a university that makes you a professor. So I prpose to remove all occupation professor and make it similar to a member of parliament (they are politicians). Thanks, GerardM (talk) 07:05, 12 May 2016 (UTC)
We need more than one item for professor anyway. It has several different meanings only in Swedish applied to Swedish circumstances. -- Innocent bystander (talk) 08:35, 12 May 2016 (UTC)
The distinction between a "job" and a "title" is 0. People are paid to perform work with a certain title. Your argument could be extended to "teachers" as well, which is ludicrous. --Izno (talk) 12:46, 12 May 2016 (UTC)
Well, the "title of professor" in Sweden does not always require that you have an employment or are paid a single cent. We who are a little older are used to see professors as the managers of a department at the University. That is no longer a requirement. They today do not have to be managers of anything. It today tells more about your academic merits and your income than anything else. -- Innocent bystander (talk) 13:24, 12 May 2016 (UTC)
The difference between the job and the title is massive. People teach at a university and for all kinds of reason are not called a professor. Being employed at a university and doing research or teaching does not make you a professor. It is a title only pure and simple. Thanks, GerardM (talk) 15:25, 12 May 2016 (UTC)
Some of the problems here are language-specific. The label of university teacher (Q1622272) in Swedish says nothing about "professor". It says "university teacher". It looks like the same in German (de). A person with this title in Swedish, normally translates it to "University Professor" in English and nobody has any objections to that. But it also depends on the rank of the person. I have myself, acted as teacher at a University. But since I have no post-graduate degree, I could not call myself "university teacher". I acted as an engineer (Q81096) even when I gave lessons. -- Innocent bystander (talk) 08:28, 13 May 2016 (UTC)
The English-language version could be changed to "university instructor", to denote someone who (as the whole or part of their duties) teaches classes in a university. Professor is a title separate of this responsibility. We won't be able to fit all cases, but from my research this could fit the most. Ajraddatz (talk) 08:47, 13 May 2016 (UTC)
Weapons - Their instances and subclasses
It seems to me like a Q2295322 is a subclass of Q170382. It's however listed as a instance_of. Unfortunately it's not the only concepts that is handeled this way. Was there some decision to tag weapons like Q170382 this way? --ChristianKl (talk) 14:19, 12 May 2016 (UTC)
The next object where it appears is Smith & Wesson Model 1 1/2. There are also lot's of other Smith & Wesson Model's for which it's the case. Is there a way to automatically recategorize them all?
Cross wiki notifications will be released by default on May 12 at 23:00 UTC.
During the beta phase, the cross-wiki notifications feature was enabled by over 18,000 accounts across more than 360 wikis. We receive great feedback from a lot of very happy users. After that 3-months long beta period during which we made adjustments and that feature is now ready for a release by default.
Users who don't want to receive cross-wiki notifications will be able to turn them off on their preferences on each wiki. If you haven't activated Cross-wiki Notifications during the Beta phase, you may receive old unread notifications from other wikis.
For this to be useful, we need a system which allow us to set the preferred language for every project, not once for each project. Getting notifications in Banjar was not useful at all.
We need an optional common user_talk page in the same way we can have a common user page. Getting a welcome-template in almost every project we visit is not helpful at all.
I would love a way to set preferred languages which would apply to all wikis by default (overridable locally in the same way we can override global user pages). I must have changed the language setting on hundreds of wikis by now and I still keep finding more. - Nikki (talk) 08:05, 13 May 2016 (UTC)
I guess we just have to learn to ignore all those identical notifications in dozens of wikis. --- Jura08:31, 13 May 2016 (UTC)
As said in the original post, you can opt out of this feature (and all other notifications) through your preferences. Ajraddatz (talk) 08:45, 13 May 2016 (UTC)
Aah yes, I understand what you mean. Hopefully global preferences will come around sometime - that would fix the default language issue as well. There might even be a script that can do this; I'll ask around and report back here on that. Ajraddatz (talk) 08:54, 13 May 2016 (UTC)
Yes, there used to be a script on meta, but you have to give your password and username to a Steward to be able to do it. I used that script on my last username but it does not affect this account and it does not affect new projects. -- Innocent bystander (talk) 08:59, 13 May 2016 (UTC)
Hmm, do you remember who it was that ran the script? It might be possible to duplicate that functionality with OAuth, and without the need for a steward to do it. I'm heading off to bed now, but I'll look into that more tomorrow. Ajraddatz (talk) 09:03, 13 May 2016 (UTC)
For a period, it used to be two Stewards who maintained the bot which run over all projects to modify user-pages and script-pages. One of them had a script (s)he used to modify user preferences. Since Global js/css and global user pages came, I seldom see them, and I have forgotten their usernames. -- Innocent bystander (talk) 09:31, 13 May 2016 (UTC)
Number and list of all labels and properties created in any given language
Hi, can anyone please share the number and list of all Wikidata labels and properties created in any given language? --Psubhashish (talk) 11:43, 13 May 2016 (UTC)
When running a test adding coordinate locations using QuickStatements, I noticed that the precision for each statement is set to ±0.000001° by default, and there appears to be no way to specify the precision you want. See the example coordinate location (P625) statement on Wikidata Sandbox (Q4115189).
Does anyone know of another way (apart from bot edits) to add a large number of coordinate locations with precision? NavinoEvans (talk) 17:51, 13 May 2016 (UTC)
What's the relationship between Q902296 and Q2715121
I looked through various terms around the French concept of phytosanitaire. The English translation that the WHO used for their treaty seems to be phytosanitary.
At the moment it's not clear to me whether Q902296 and Q2715121 are synonyms or whether one of them is a hypernym of the other. Can someone with more French than
me clear up the confusion?
single-species genus and species for extinct animals
I am not 100% proud of what I did in Talk:Q732917. Well, I was asked by a wikidata-newbie to help him fix the problem of connecting two items and I face again a problem I wrote a thread on the local project "life beings" of itwiki many months (years?) ago.
So it seems to me noone really discussed it, and if they did please let me know what we decided.
The problem is easy: extinct species, mainly reptiles in my experience, are "taxonomically isolated" and most of the time the name of the genus in "assigned" whilst naming the species. Now some languages such as itwiki go for the genus concept, other ones for the species but in the end is the same stuff, there is no point in not cross-linking them, you are describing the only species of the genus, it is just the way you write the incipit and if you add a line in the lateral infobox
This time I am not ignoring, I go public: what should we do? I was bold in the moment, I confess. To me the most logical think is to stop at the genus level, be tolerant with the link of "species " articles and accept the species name as a secondary description. If we want to have them separated, one item for the species and one for the genus (and we can't have an item for the species without one for the genus, it is a necessary link), it will take some awareness to gear because languages may continue to connect them "randomly".
Anyone else? I can link from local projects later, if there is no previous discussion. I can start a second discussion where you think is more appropriate, but we should take a final decision. Don't rollback anything, I can fix it if you don't agree, just tell me.--Alexmar983 (talk) 11:01, 2 May 2016 (UTC)
(f.c.) update: when cited or archived I should use the correct expression of "Monotypic taxon", just learned it--Alexmar983 (talk) 09:13, 3 May 2016 (UTC)
You will not be misunderstood if you use "monotypic taxon" but you should be aware that it is not correct. - Brya (talk) 11:15, 3 May 2016 (UTC)
A genus is not a species, so there should be an item for the genus and an item for the species. Then put all the sitelinks for genus and species in one item (unless this is not possible or plausible). This item will be the genus for fossil taxa, and the species for extant taxa. Routine. - Brya (talk) 11:21, 2 May 2016 (UTC)
that was not very clear. It is written in a very confident way, but sounds confusing I hope just to me. One sentence say "there should be an item for the genus and an item for the species" and the other one "put all the sitelinks for genus and species in one item". First of all, which item of the two? If you do not fix a standard I can have one group of wikilink for an extinct species/genus under the species item and another one on the genus item. Secondly, if make two items available soon or later someone may link to the other one if fits. Or if they find some group of links under "genus" and some under "species" they will start to standardize them in a direction or another. that's a "ballet" that will never stop. Maybe that's why it's routine (I agree that's a routine I keep finding), but shouldn't we have a clear guideline to link in this case? I can link your answer if asked but it won't do it, IMHO.--Alexmar983 (talk) 12:32, 2 May 2016 (UTC)
If missing interwiki is the only problem, see the members of sv:Kategori:Sidor som hämtar kompletterande interwiki med hjälp av Modul:Interwiki at svwiki for a possible solution. They all summon interwiki from at least two items. Therefor sv:Försäter has interwiki to nl:Försäter even though they do not share item. The Swedish article is about a group of villages, while the Dutch is only obout one of them. Since they do not share statistics, they should not share item.
In the same way, interwiki between a Genus and Species-artiles is possible. But this template on svwiki has mainly become popular among users who edits disambigs like SAN (disambig) and San (disambig). -- Innocent bystander (talk) 13:38, 2 May 2016 (UTC)
It is a constantly recurring topic. Like I said, there should be, as a rule, two separate items, one for the genus, and one for the species. Given the structure of Wikidata, it is impossible to have an item for a species without also having an item for a genus. Sometimes there is only an item for the genus (indeed for dinosaurs, and the like).
Sitelinks (iw-links) go:
to both items if there is duplication (that is there is one or more Wikipedia that has a page each for each topic), in which case the links for the species pages go into the item for the species, while the links for the genus pages go into the item for the genus.
to one item: namely the species level (this is the default, and always the case for still-living, extant taxa)
Where has this been codified? previous discussion would be useful, in general it should be written on a help page and accessible through a clear link, especially if it is constantly recurring topic. BTW I link from some local projects so we are sure local users kinda know that. --Alexmar983 (talk) 07:33, 3 May 2016 (UTC)
Alexmar983: I encountered the same problem for projects treating a same species under different names, each of these names having a Wikidata item associated. See this discussion, where Brya did exactly the same "dirty" thing: linking "name1" interwiki on "name2" item page rather than on the "name1" one... Totodu74 (talk) 07:50, 3 May 2016 (UTC)
thank you. So I should have asked at Wikidata talk:WikiProject Taxonomy . My apologies, maybe when can move the discussion there when finished. BTW I don't have any strong idea... I just hope that if "we" all agree on a method for this type of situation, maybe we can jus write it down on Wikidata:WikiProject Taxonomy main page, providing some automatic maintenance list similar to those shown by Innocent bystander but at a meta level. Something like "list of genuses with only one species linked". We kinda now we have to check them regularly. Do we know at least how many of these items we have? We all say we've faced the problem sometime, so they are not few aren't they? Just humble suggestions, of course, I am not the expert here. But please write this down. It would much less dispersive to a "newbie" than giving a link to one or two previous discussions.--Alexmar983 (talk) 08:09, 3 May 2016 (UTC)
No, if a genus is monotypic (has a single species), there should not be separate articles for the genus as species; they both refer to the exact same animal. And no, this problem does not only apply to extinct animals. A living example is the kakapo[3], which is a single species within a genus (which is the only genus within its family). The difference with modern animals is that they have common names for their article titles, whereas the extinct ones only have the scientific name. Having different Wikidata entries for each taxonomic name that just refers to a single animal is ridiculous, overcomplicated,and doens't help any readers. The example, Daemonosaurus chauliodus, is the exact same thing as Daemonosaurus. Not a single word could be written about the former that wouldn't aply to the latter. FunkMonk (talk) 08:47, 3 May 2016 (UTC)
As an aside on the Kakapo, yes, it's a single species in its genus, but not in its family, there's two genera with three extant and one extinct species in its family Strigopidae ;-) MPF (talk) 09:03, 3 May 2016 (UTC)
I think not so much out of date, as failing to follow the agreed international IOC classification; some other authors do split Strigopidae into two families, but since en-wp's WP:BIRDS adopted IOC as the best peer-reviewed taxonomic authority, they should stick to it. - MPF (talk) 11:09, 3 May 2016 (UTC)
Different Wikipedia's could in theory follow different classifications. So it seems Wikidata would need entries for both, even though they are theoretically synonymous... FunkMonk (talk) 13:33, 4 May 2016 (UTC)
Did someone say it was "only" for extinct animals? I just started from that because that's the situation I've always faced. In any case even if we assume that separate articles are not necessary, which I agree, and even if all the projects agree locally to do that (which they do, more or less as far as I know), still I kinda agree we may need in an archive like wikidata two separate items. From the practical point of view in the beginning I did not, but I can accept there is a bigger picture... Question: what do external databases do in this case? I am worried about confusion with Identifiers too.--Alexmar983 (talk) 08:50, 3 May 2016 (UTC)
@FunkMonk: I fully agree with you, though it can't be discussed here on Wikidata if a Wikipedia project decides to have 2 distinct articles for each monotypic genus and its species, but anyway that's not the point here: we are discussing the case of "Russian dolls" articles with the genus as title on WP1 and the species as title on WP2. On Wikidata, genus and species can't be treated on the same page, but we need to find a way to provide interwiki links between WP with different title choices. For example take a look at Q13507550 (Cycloseris wellsi) and Q3948967 (Coscinaraea wellsi). Totodu74 (talk) 08:57, 3 May 2016 (UTC)
That one seems to be a different issue, which is how to treat scientific synonyms. Such should simply be merged here. As for genus and species, if we can have them separate here, but make them both link to for example the same English Wikipedia article, but two different Italian Wikipedia articles, then that would solve it... FunkMonk (talk) 09:13, 3 May 2016 (UTC)
I'd say best to have separate Q numbers for a monotypic genus and its species; any currently monotypic genus will sooner or later become polytypic as fossils of its extinct ancestors are discovered (or if genetic evidence shows the living taxon is actually two or more cryptic taxa). - MPF (talk) 11:09, 3 May 2016 (UTC) Addenum: and of course, Wikispecies adopts separate pages for each rank, one for the genus, and a separate page for the species - MPF (talk) 11:12, 3 May 2016 (UTC)
As I was saying here in French my doubt is hat even if we are adamantly clear that we use for example the species item and that we must have two items, the local user may not know that. If (s)he creates the article with the name of genus (according to the local policy) and (s)he connects it by clicking on the left bar (s)he has two option, the genus item and the species item. How can (s)he know that the interlinks are all hosted in the one with the species title? We have thousands of potential article to do on dozen of relatively active languages, every time this is done we have a potential source of confusion. if we don't write it clearly, and explain at least here, how can we avoid useless duplicates of work? Plus, a monitoring list or something could be really useful.--Alexmar983 (talk) 10:20, 3 May 2016 (UTC)
The default result is already useless duplication, it seems... The matter is how to solve the useless duplication, or how to prevent it from happening in the first place. I had this exact problem with an article I created recently. I linked this English genus page[5] to the Italian species page[6], but afterwards I discovered there is also an Ukrainian genus page[7], which can now not be linked to the former. FunkMonk (talk) 10:33, 3 May 2016 (UTC)
Codification runs a bit behind. The Wikidata datamodel is dynamic and evolves constantly. But codification when it happens can be found here.
But the whole question of how many species there are in a particular genus is a question of taxonomy. And this, all too, often is dynamic. Adding references to any taxonomic opinion is of paramount importance. An unreferenced statement that "this genus has only one species" is just stating an opinion, quite possibly one among several.
Actually, if we get to the bottom of it, this is just another example of how irrelevant and impractical bijective links are. (n:n links would be much welcome.) —Tinm (d) 02:33, 4 May 2016 (UTC)
There are always going to be some monotypic taxa, and different wikis will always have different ways of dealing with them. The standard at the English Wikipedia is explained briefly at here. Purely bijective links are simply wrong. Wikidata must be able to handle links of every kind of arity. Peter coxhead (talk) 08:28, 4 May 2016 (UTC)
@Tinm, Peter coxhead: An alternative is to develop WD:XLINK. But @Lea Lacroix (WMDE): this would need a lua API to get the statements with an item of the page which have the item of the current page as value. It's a main usecase of this : if some wikipedia has an item about a genus and that another as an item about the unique species, we have to go from the genus to the species to generate automatically a sitelink. Although it comes to my mind that (@Tinm, Peter coxhead:) {{Union of}} would totally fit in this case for a substitute of an inverse property (sourced of course) that could be used on (say the taxobox template) to generate a sitelink. author TomT0m / talk page09:33, 4 May 2016 (UTC)
A good test case is the taxonomic hierarchy: order Amborellales – family Amborellaceae – genus Amborella – species Amborella trichopoda. The order, family and genus are monotypic. The English Wikipedia has one article at the genus; the German Wikipedia has one article at the species; the French Wikipedia has articles at the order, family and species; the Spanish Wikipedia has one article at the family; and so on. At present these aren't properly connected. There needs to be a proper solution. Over to you guys, as I don't work here. 86.113.153.21513:24, 4 May 2016 (UTC)
the sentence Over to you guys, as I don't work here. together with the fact you are posting here is kind of self-contradictory :p author TomT0m / talk page16:43, 4 May 2016 (UTC)
@TomT0m: what I meant by "work" is that I have no idea whether the right kind of links can be set up in Wikidata and if so how; I just know that that strictly 1:1 doesn't work. Peter coxhead (talk) 10:14, 6 May 2016 (UTC)
This depends on what one calls "properly connected". It actually looks pretty good, especially compared to pre-Wikidata days when the weirdest things were linked together. There won't be a perfect match until and unless each Wikipedia has a page at each level, and this is unlikely to happen. - Brya (talk) 16:30, 4 May 2016 (UTC)
It's unlikely to happen (at least at the English Wikipedia) because it's there's a strong consensus that it would be the wrong thing to do. Wikidata's inter-wiki links have to be able to cope with what wikis do; Wikidata can't determine this. Peter coxhead (talk) 10:14, 6 May 2016 (UTC)
Taxa aren't the only area where there's a problem with only 1:1 links. The English Wikipedia decided to split one article on "berry" into two: w:Berry for the general sense, and w:Berry (botany) for the strictly botanical meaning. Neither are currently correctly linked to other wikis in Wikidata. If the other language wiki has two articles, then this can be fixed, but many don't. I tried to sort this when the split happened, but then realized that I couldn't do it, so gave up. Please find a way to fix this, otherwise it will continue to reflect badly on this aspect of Wikidata. Peter coxhead (talk) 10:14, 6 May 2016 (UTC)
Yes, everything dealing with a fruit tends to be messy. I went through the sitelinks for both berries. These Wikipedia pages prove to come in four categories (not counting two pages that do not belong here at all):
pages strictly from the botanical perspective,
pages from the botanical perspective that also point out that the culinary perspective is different,
pages from the culinary perspective that also acknowledge there is something like botany,
pages from the culinary perspective that ignore that there is something like botany.
The problem is not with Wikidata (and I doubt it can be fully solved here), but at the Wikipedia's. Wikidata makes it possible to at least handle such cases in the best possible way, although clearly short from the ideal. - Brya (talk) 16:33, 7 May 2016 (UTC)
Links to redirects are just a dirty & intractable way to allow 1:n links. If n:n (or 1:n, but that wouldn't be consistent) is to be allowed, it has to be done cleanly or it's going to be a real mess. I'm against links to redirects. —Tinm (d) 16:39, 14 May 2016 (UTC)
Help with a query
Hi. I've done a query to list the given names of all people born in Palma (Q8826). Can somebody help me to do the same with people born in the island Mallorca (Q8828). I've tried but I don't know how to make the query take into account all the geographical items contained in the island, including Palma. Could somebody help me? Paucabot (talk) 16:01, 15 May 2016 (UTC)
We have two separate entities for the plant and the food. However some of the WP entries, including the examples, cover both concepts. Do we then need a third item for combined coverage? All the best: RichFarmbrough, 22:43, 13 May 2016 (UTC).
Certainly a problem exists. Wikidata was never properly planned from the start. It now has several conflicting objectives. Possibly the centralised handling of interwikis should be completely split from the "database of all things". The two don't fit together. Danrok (talk) 23:13, 15 May 2016 (UTC)
perhaps it would be best to create properties for "next higher rank" and "next lower rank" (perhaps "ranks immediately higher than"/"ranks immediately lower than" would be better labels?), which would be useful in any field that has ranks (military, police, fire service, etc) and potentially also any other sets with ranked order (e.g. longest bridges, highest mountains, etc). Thryduulf (talk: local | en.wp | en.wikt) 13:28, 14 May 2016 (UTC)
I think this is a much larger problem and that it deserve a much larger discussion about modelling orders in general. Would deserve a section in Help:Modelling. The point of a order is that an order deserve an item by itself, like "taxonomic rank order", "french military ranks order", "french honorific titles" (Knight of the Legion of Honour (Q10855271)), episodes of TV series in their series and/or season ... So I think what we actually need is a property to indicate in which order the precedence occurs. I think a generic property "ranked in" (which might exists for sport https://www.wikidata.org/wiki/Property:P2291 ) is enough. The statement can then be qualified by the usual properties. author TomT0m / talk page07:42, 15 May 2016 (UTC)
Please be aware that military ranks differ from country to country and new ones have been introduced over time.. So do use qualifiers at all times. Thanks, GerardM (talk) 14:12, 14 May 2016 (UTC)
Defining a Wikibase extension of SPARQL property paths
Hi people, a little project I'm working on (with perspectives) is Module:PropertyPath.
The short term goal is to define a function that can take a sequence of properties like P279/P31 and to compute the statements that can be found starting from an item. For example if we start from "example item", that
The idea come from the "PropertyPath" functionality of SPARQL. In the future I'd like to include some more propertyPath functionality like :
Alternative : (P279|P31)/P31 where (P279|P31) means that we can follow either P279 or P31 claims, and the result must be followed by a P31 because of the /P31
Repetition : (noted with "*") (P279|P31)*/P31 means that we can follow any number or path like "P279/P31/P31/P31" ... (optionally 0)
The point is that property path is limited because there is no handling of qualifiers path for example. So to be more usable for us a few extra functionalities would be useful like :
conditional following : P279[Pxx]</code : the path matches if P279 is qualified by Pxx
qualifier paths : P279:Pxx : P279 is qualified by Pxx and we wand to follow not the main value of the claim but the value of the qualifier.
I can imagine two functionalities for this : compiling this into a SPARQL (partial) query and as such create a Category:Partial query to use this with the query service for us to write query, and in lua for us to use in infobox code for example. Comments on (un)usefulness/ ideas / support ? author TomT0m / talk page09:48, 15 May 2016 (UTC)
It might be indeed, maybe later. At the moment I plan only to use on-wiki available features and experiment a little bit. If that become popular why not adding a service like the language one to the query service. author TomT0m / talk page14:22, 16 May 2016 (UTC)
This is the thing I am waiting for so long. I made some experiments myself (here examples) but the method is very inefficient, especially for taxonomy due to very long path from the bottom to the top. Paweł Ziemian (talk) 18:40, 16 May 2016 (UTC)
Wikidata weekly summary #209
Here's your quick overview of what has been happening around Wikidata over the last week.
In a nutshell, later this week this will provide a one-time "Welcome" message in the wikitext editor which explains that anyone can edit, and every improvement helps. The user can then start editing in the wikitext editor right away, or switch to the visual editor. (This is the equivalent of an already existing welcome message for visual editor users, which suggests the option to switch to the wikitext editor. If you have already seen this dialog in the visual editor, you will not see the new one in the wikitext editor.)
I want to make sure that, although users will see this dialog only once, they can read it in their language as much as possible. Please read the instructions if you can help with that.
I also want to underline that the dialog does not change in any way current site-wide and personal configurations of the visual editor. Nothing changes permanently for users who chose to hide the visual editor in their Preferences or for those who don't use it anyway, or for wikis where it's still a Beta Feature, or for wikis where certain groups of users don't get the visual editor tab, etc.
There is a slight chance that you see a few more questions than usual about the visual editor. Please refer people to the documentation or to the feedback page, and feel free to ping me if you have questions too!
Finally, I want to acknowledge that, while not everyone will see that dialog, many of you will; if you're reading this you are likely not the intended recipients of that one-time dialog, so you may be confused or annoyed by it—and if this is the case, I'm truly sorry about that. This message also avoids that you have to explain the same thing over and over again—just point to this section. Please feel free to cross-post this message at other venues on this wiki if you think it will help avoid that users feel caught by surprise by this change.
New to this, unable to add spouse can someone explain what I'm doing wrong
I'm new to wikidata so forgive my ignorance. I'm attempting to add a spouse to Q6162079, but any time I try to add her name, the save link is grayed out. I was able to add the references, just not her name. Can anyone explain to me why it is greyed out and what I'm doing wrong? Offnfopt (talk) 07:43, 22 May 2016 (UTC)
As it's set to "no value" now, you'd need to change it to "custom value" (see GUI on the right).
Just to make sure I understand this correctly. So in order to add a spouse, the spouse also has to be notable and if they aren't then the information for the spouse can't be added? If that is the case then I guess I should just delete the spouse/references. Offnfopt (talk) 08:08, 22 May 2016 (UTC)
If you think that it is important to have the spouse you enter it. It is then notable on the back of the spouse. This is Wikidata remember! Thanks, GerardM (talk) 10:59, 22 May 2016 (UTC)
I have added water course since the categories in the Lsjbot-articles includes all kinds of flowing bodys of water. In many cases, these statements can be replaced by river, that is more specific, but it does not fit everywhere. We do not have one single river in Sweden (Q34) for example. -- Innocent bystander (talk) 16:34, 15 May 2016 (UTC)
There is maybe no precise and scientific way to handle this, since language may differ. So, yes, there is a large risk of "unscience" here. In Swedish it even differs between different parts of Sweden. When we, most likely never, can agree about the quantifiers of these entities, a general item like Water course is to prefer? An example, Jordan River in Palestine (Q23792) is named River, but has the size of a small river (Q3529419).
I mean, every running water that isn't made by man, is probably a subclass of Water course. But a very low percentage is a (subclass of) River, at least now after the Lsjbot-project has added so many articles about small water courses. -- Innocent bystander (talk) 17:42, 15 May 2016 (UTC)
I agree, it is difficult to quantify this subjects. But calling everything "River" look very strange. As I said, in Swedish vocabulary, we do not have a single river in Sweden. Volga, Rhen, Lena, Nile etc are rivers, but Viejejåkhå above does not look like a river to me. -- Innocent bystander (talk) 18:12, 15 May 2016 (UTC)
Maybe in a perfectly logical world, if different languages have different classifications of types of watercourses, wikidata would have items corresponding to each of the specific classification definitions independent of language. For example, if there's "river as defined in Russian (and some other languages?)" which applies to a greater number of actual watercourses than "river as defined in English", we would have some items listed as instances of "river as defined in Russian" but not "river as defined in English". However, that's probably a needless level of complexity. An alternative would be to create class-level items or property statements that specify more precisely the useful information - instance of "navigable watercourse" rather than "river", perhaps, or "width" "1 m" rather than instance of "stream"? ArthurPSmith (talk) 14:43, 16 May 2016 (UTC)
Did we really need a perfect world ? I think that we should just rely on the sources. If there is a source telling that the item is about a *whatever*, then we indicate the data accordingly. If there is no source, the general watercourse (Q355304) is the best option (because all river (Q4022) are watercourse (Q355304) but the contrary is not true). And for the specific characteristics, there is specific properties; we shouldn't use instance of (P31) for that (like width (P2049)). Cdlt, VIGNERON (talk) 15:39, 16 May 2016 (UTC)
Apparently there is some discrepancy with attempted murder (Q81672) (Sammyday reported it on the french bistro but no-one really answered). Right now, in most language the label is « attentat » (or a close word ; in fr, de, da, cs, et, eu, hr, etc.) and the description is more or less about an attack (with or without killing, toward someone or something, in fr, de, it) but in English the label is « assassination » and the description « deliberate killing of a prominent person or political figure » (same thing for es). Can someone check it and confirm that it should be split in two different item ? There is murder (Q132821) that has maybe to be check too.
The Swedish article well fits with the English description you wrote above. ("deliberate killing of a prominent person or political figure") -- Innocent bystander (talk) 14:01, 17 May 2016 (UTC)
I mean that the content of the Swedish article in this page is about "deliberate killing (or a try to kill) a prominent person or political figure". The word used here is also used to describe the Charlie Hebdo shooting (Q18718876) The article is illustrated by the murder attempt at Ronald Reagan. The beginning of the article tells that the word is derived from a latin words that means "attempt to kill". This word is not used in a legal context in Swedish. To do an "attentat" is of course illegal, but the term is not used when you come to the court. You are then accused of "Mord" or "Dråp" or "Vållande till annans död", depending on the level of intention and planning. -- Innocent bystander (talk) 16:12, 17 May 2016 (UTC)
Creating a list of missing Wikipedia articles with source URLs using a Wikidata query
Hi all
Yesterday UNESCO made all its Biosphere Descriptions available as CC-BY-SA meaning that they can be used as the missing 400+ English language Wikipedia articles for the sites :) I'm currently trying to create a Wikidata query to make this job easier for people, I have a query but I'm struggling to add a column for Property:P2520 which will link through to the description and also the image for the Wikidata item. Could someone help me out pretty please?
Tracking more information about Wikimedia movement affiliates
Greetings. I am looking into the idea of tracking relevant data about Wikimedia movement affiliates on Wikidata. In addition to providing helpful data about affiliates on Wikidata, this could also be used to maintain content on other wikis, such as Meta-Wiki and possibly WMF-Wiki in the future. Nearly all affiliates already have items as they also have pages on Meta-Wiki. However, I think we can utilize and format these items a bit more. I am hoping to get some feedback on this idea and the addition of a couple properties to help implement it.
I have mapped out what information I think would be needed, and how they relate to existing properties and other elements:
instance of (P31) (current can be linked to main items for type; proposed and former would need to be created, but may also have usage as Meta-Wiki pages)
Focus area flag (used in short listings to help depict affiliate - generally a flag representing affiliate's region/country or focus. Examples on Press page on WMF-Wiki)
I think I'd prefer not to use a "focus", and certainly not as a text field. Try it out with "main subject" and see which orgs and where we bump into issues, as I'm skeptical that we won't already have items for most focuses, and where we are lacking we can probably just create a new item. Regarding the code, see no issue, so you should probably propose that at WD:PP. --Izno (talk) 13:16, 19 May 2016 (UTC)
Thank you for your feedback. I think we could make that work. Some tricky ones would be things like "LGBT+" (which is LGBT in Wikidata), "Women, Spanish language" (which is more limited "Women in Spain" on Wikidata), and "United States, North Carolina Triangle" which does not appear to have a Wikidata entry. Any ideas on how to address those variables? Would new items using those as sort of "aliases" or new groupings be okay? --GVarnum-WMF (talk) 17:40, 19 May 2016 (UTC)
I've added aliases to LGBT; the English article at least treats the extended phrase "LGBT/Q/I" in the article called LGBT.
I think in this case, "women" and separately "spanish language" could be claimed and the point would be made that the organization of interest is caring about both rather than one or the other (set as preferred rank both?).
@Srittau: How would you treat a "proposed" vice "active" using. Inception can be used with "active" but what about pre-active. "inception: no value"? Do we actually want to track items which are only proposed and not active or disbanded? @GVarnum-WMF: For the claims you want to make using P31, are there any other values? --Izno (talk) 18:15, 19 May 2016 (UTC)
If you want us to annotate your list, it might be better to place it on a separate page. --- Jura18:17, 19 May 2016 (UTC)
National museum
Hello, we have national museum (Q17431399) which I found via a specific museum. This Wikidata item is linked only to one article in a language version. Besides, I find the definition rather limited, in spite of the many ways a museum can be "national" (about a nation, the most important museum of a country, government-owned).
Well they could be merged as there don't seem to be conflicting wiki-links; they are at least linked together via the property is a list of (P360) so they could also be kept as is. If merged the name should be "national museum" and it should not have the instance of "Wikimedia list article" statement. See this RFC. ArthurPSmith (talk) 14:37, 19 May 2016 (UTC) (note - I edited this comment a little later ArthurPSmith (talk) 14:41, 19 May 2016 (UTC))
What are the notability criteria when it comes to scientific papers and their authors?
Is every paper used as a source within Wikidata and it's authors automatically notable, or are there cases where the policy is that it wouldn't be notable? ChristianKl (talk) 16:15, 19 May 2016 (UTC)
What Jura said. In general, the notability criteria in Wikidata are quite a bit lower than on other projects due to the need for "structural" items. --Srittau (talk) 16:47, 19 May 2016 (UTC)
Is there a possibility to activate two languages? What I mean is that if an item doesn't have a label in language 1 Wikidata checks language 2 first before merely showing the Wikidata id. --Jobu0101 (talk) 13:47, 18 May 2016 (UTC)
The software also looks into the preferences in your browser and the geolocation of your ip. If I remove the Babel from my user_page I therefor get Finnish and some other strange languages into my FallbackChain. That is of no help for me, I do not read those languages. But there is some logic behind that, since a large minority speaks Finnish or Meänkieli where the geolocation thinks I live. For a period I was also given South Sami, but it is gone now. -- Innocent bystander (talk) 06:29, 19 May 2016 (UTC)
Can I somehow force the system to remove the cache and render the page again? And how long does it take when I don't do anything? --Jobu0101 (talk) 10:06, 19 May 2016 (UTC)
When you switch language, you download a page that you never have been at before. You then do not have any cache-problems. So that was expected behaviour. I have seen the old version of a statement for several days in some cases. There is maybe a way to remove the cache, but I do not remember how to do that here at Wikidata. At enwp I see a help page that tells me to add "?action=purge" at the end of the url. Try that! -- Innocent bystander (talk) 04:20, 20 May 2016 (UTC)
@Innocent bystander: I tried it and it doesn't help. I don't think that this is a cache problem. Neither local cache on my computer nor global cache on Wikidata. I think so because I excluded the local cache case by opening the page on several systems which haven't been at the page before. Globally I guess that there are many Wikidata users which have my current language settings for months. So Wikidata would use the cache generated for them. Cabinet Wowereit II (Q873338) hasn't been changed for a long time. So even when the cache is quite old it should show me the labels in German (because there don't exist English labels). --Jobu0101 (talk) 07:38, 20 May 2016 (UTC)
I tried myself to change to de-1 instead of de-0 in my #Babel (not true at all, but anything for science) and I can confirm that I do not see any label in the claim. And I do not see anything but the Qid in the header of my browser either. I normally see the English label there, when there is one. It is not until I change the whole GUI to de, that I see a label. Lydia? -- Innocent bystander (talk) 08:28, 20 May 2016 (UTC)
Fallback like the one you want unfortunately doesn't work (yet). We can not do a fallback that depends on the user's language there because of caching. All fallback we do is the regular MediaWiki fallback there. So if you look at the page in German and there is no German label but an English one then it works because one of the MediaWiki fallbacks is German -> English. --Lydia Pintscher (WMDE) (talk) 08:41, 20 May 2016 (UTC)
As someone doing deletions on a regular basis I check "what links here" a lot. I do not object having new interface, and it looks nice, but I would very much like to have all to information be on one screen in most of the cases. Right now I need to scroll down even if there are zero or one backlink. This can be done either by arranging the options horizontally, or moving them below the results of the search.--Ymblanter (talk) 20:57, 19 May 2016 (UTC)
I've never used a screenreader but that seems unlikely to me. I can't see how blind people would need styled forms, the standard form dropdown replaced by a reproduction in divs and tons of whitespace. To me it looks more like an interface aimed at mobile devices that sucks on a desktop. It seems to have been reverted though, I'm seeing the old one again now, thankfully. - Nikki (talk) 07:03, 20 May 2016 (UTC)
How to document data about impact of oxytocin on trust
I'm new and want to know how to add the results of a specific study. Male U.S. citizens were asked to identify themselves as republican, democrat, or independent, then they were given either 4ml oxytocin or a placebo, then asked to rate their trust in the government. The democrats who received oxytocin rated their trust in the government significantly higher than those who received the placebo, but this pattern was not seen in subjects who identified as republican or independent.
In Wikidata I see an oxytocin item with lots of properties, but I don't see how to add the results of this study. If I try to boil the discovery down to statements, the subject is a kind of male U.S. citizen and the properties are 1. likely to identify as democrat, and 2. trust in government is moderated by oxytocin. At this time we do not know whether the effect could be replicated in non-citizens or females, and we have not pinned down exactly what the type of person happens to be (e.g. is it people with a certain gene? could it include Koko the gorilla?). There is real knowledge here, but can Wikidata handle it? Is the existing Wikidata entry about oxytocin problematic because it describes impacts, but does not specify the species/subspecies in which the impacts manifest?
Finally, for my reference, can I add the sample size, power and magnitude of the statistic as metadata?
Welcome again to Wikidata!⸻First and foremost, in the interest of not seeming suspicious to others, you should create an account and sign all of your posts with four tildes (~~~~).⸻Second, if this study is published and has a valid DOI (P356) or similar external identifier, then you have a better chance of an item on this study not being deleted on the spot.⸻Third, you don't need to add your results to oxytocin (Q169960); you may wish to correspond with User:James Hare (NIOSH) who is as I'm writing this message adding items about papers describing scientific studies.⸻Fourth, as your IP indicates affiliation with the government of Wisconsin, you should declare this prior to adding anything of note. Mahir256 (talk) 03:08, 20 May 2016 (UTC)
I would recommend adding a Wikidata item about the paper. If you have the DOI, or a PubMed ID, or a PubMed Central ID, you can use the Source Metadata tool and it does all the work for you. I highly recommend it, but you need an account for it. As for adding the findings to the item on oxytocin, I don't particularly recommend it, since it's just the findings of one study. James Hare (NIOSH) (talk) 03:29, 20 May 2016 (UTC)
Any property for linking a reservoir with the river feeds it?
Hi all, do you know whether there is any property linking a reservoir with the river(s) which feed the reservoir? If no, does it make sense to create it? --Discasto (talk) 09:30, 20 May 2016 (UTC)
Using Wikidata queries to organise a open license text import into English Wikipedia
Hi all
I'm very happy to say that UNESCO has made the official descriptions of all Biosphere Reserve sites available under a Wikipedia compatible license. Currently around 440 (out of 670) of the Biosphere Reserves do not have an English language Wikipedia article.
These descriptions can be used as the missing Wikipedia articles with very little adaption. I have created a Wikidata query and a set of instructions to help people create the missing articles. I hope that this is useful for other people interested in using Wikidata to organise writing projects etc.
Thanks very much to Navino Evans who did all the data importing into Wikidata (a herculean effort) and Andy Mabbett for helping me with the instructions.
Is there an item "points" that can be used as a unit? I found point (Q1550236) "ice hockey statistic summing a player's goals" and point (Q2353718) "basketball unit of scoring accumulated by making field goals and free throws", but no general such unit of measurement. If it does not exist, where can I propose one? --Bensin (talk) 15:02, 14 May 2016 (UTC)
If there isn't one (and I haven't found one either) then you can just create a new item without needing to formally propose it anywhere. Thryduulf (talk: local | en.wp | en.wikt) 15:33, 14 May 2016 (UTC)
@Bensin: Precis, men det kan vara en bra idé att på något sätt försöka hitta ett sätt att objektet får åtminstone en normal länk riktad till sig. Annars riskerar den att bli raderad som "irrelelvant" med jämna mellanrum. Vi använder nämligen ett verktyg som räknar inlänkar (Vad som länkar hit) för att avgöra om ett objekt är relevant eller inte. Och jag tror inte länkar till enheter ger utslag i denna länkräknare. Men ett sådant objekt är tveklöst relevant, det är bara våra verktyg som är lite klumpiga! -- Innocent bystander (talk) 15:38, 14 May 2016 (UTC)
In frwiki we have an article that was names "point (game and sports)" at some point. Ontologically a point is scored when a player create an instance of a class of event, defined by the rule of the game, whether it's a card game or a sport game does not matter much, I think. author TomT0m / talk page07:32, 15 May 2016 (UTC)
No, they are subclasses of the same concept which have specific meanings in the different sports. I've added that to the former (it was already present on the latter) and to point (Q7207922) that I've just found. Thryduulf (talk: local | en.wp | en.wikt) 23:41, 15 May 2016 (UTC)
As TomT0m pointed out above, ontologically it doesn't matter if a point is scored in a card game or sports game. The game rules are then defining if it's called point or goal. Claims like in Q5354210#P1346 are pleonasms and are opposing the idea of providing structured data. --Pasleim (talk) 09:32, 16 May 2016 (UTC)
Maybe so, but there is also the problem with our templates in Wikipedia. How do they know if we use points or goals? In Ice hockey, there is both points and goals in the same game. And the sum of points a teams players makes in one single hockey game normally exceeds the numbers of goals the team makes. -- Innocent bystander (talk) 10:36, 16 May 2016 (UTC)
because 1 goal = 1 point :) This is more complex in games where there is several ways to score and each way has its own point numbers. author TomT0m / talk page10:57, 16 May 2016 (UTC)
I don't think a general point concept should ever be used as a unit. In nearly every case it should be possible to pick a more specific unit. It similar to how degrees isn't a valid unit for temperature. At best a unit that specifies the rule set that's used for the unit. In Go for example games sometimes have a different result when counted with Chinese or Japanese rules. ChristianKl (talk) 08:37, 21 May 2016 (UTC)
Conflict of interest
Hoi, When the work of a scientist is tainted because of a conflict of interest, how do you mark this in Wikidata? Thanks, GerardM (talk) 13:31, 18 May 2016 (UTC)
@GerardM: As Wikidata is not saying the truth, I don't think it is our task to judge anything about the work of that scientist. The only thing is perhaps to look for the use of articles and other documents written by this scientist as reference in WD and to mark them using a specific property in order to link them with the document produced by Greenpeace. Again thi is not the task of WD to provide any kind of way to say that a guy was acting in a bad or good way. The only thing which is interesting is to offer the possibility to reader to have the reference to a document which dispute the truthfulness of another one. Snipre (talk) 18:23, 18 May 2016 (UTC)
Don't list the conflict of interest directly as conflicts of interests on the item of a person. If a person has a conflict of interest because they are employed by organisation X, the fact that they are employed by organisation X is information worth entering into Wikidata.ChristianKl (talk) 08:48, 21 May 2016 (UTC)
UEFA Euro 2016
Is there a plan to represent the full 2016 UEFA European Championship in Wikidata? Meaning that every single game result is stored here? Are there showcase items to see how to do that? --Jobu0101 (talk) 12:52, 20 May 2016 (UTC)
The only thing I can find is league level above (P2499) and league level below (P2500) which are for linking items about adjacent leagues, not what you are after. The discussion when they were proposed them did suggest a need for "promoted to" and "relegated to" properties for items about teams but nothing came of that. In your case I think you'd want a third set of properties "competitor or team promoted" and "competitor or team relegated" (which could use "promoted to"/"relegated to" as qualifiers if necessary). If nobody has any better suggestions, I'm happy to propose them for you if you aren't comfortable doing so yourself. Thryduulf (talk: local | en.wp | en.wikt) 22:18, 21 May 2016 (UTC)
Hi, I can't find any information about this artist (Q24174825). However we can find quite a number of his/her works on the Net ([10]). Regards, Yann (talk) 23:11, 21 May 2016 (UTC)
P1244 (P1244) was created to specify phone numbers in the url format. This would be Wikimedia Foundation (Q180) => tel:+1-415-839-6885 . The format needs to be enabled for Wikidata. On creation of the property, users were advised that this would eventually happen and this is now possible.
phone number (P1329) should include the same as a string, but has no built in format validation, so people need to clean up entries after users first entered them. Number of uses are currently fairly limited, but if we try to include data from Wikivoyage, this may increase eventually.
If we agree to activate P1244 (P1244) going forward, we can delete phone number (P1329). Even if I personally don't think this is key data for Wikidata, I think we need to provide a consistent way to include it. --- Jura06:26, 17 May 2016 (UTC)
I honestly don't care about activation, but we should have exactly 1 properties for this. Suggestion: Keep "phone number", delete "phone number(URL)", and add a formatter URL to "phone number", since the root data is the phone number and not the URI. --Izno (talk) 13:34, 17 May 2016 (UTC)
I don't care either way, but we certainly should have only one property. --Srittau (talk) 15:40, 17 May 2016 (UTC) (Okay, a slight preference for the non-URL version if it is kept in a normalized format. --Srittau (talk) 15:41, 17 May 2016 (UTC))
+1 to that idea. I also prefer adding a formatter URL for the same reason, which I'm pretty sure I said somewhere else before. - Nikki (talk) 13:25, 19 May 2016 (UTC)
The property P1329 we created on a temporary basis has normalization issues, so P1244 would be a simple way to have normalized data. The property for email addresses works in a similar way. --- Jura07:23, 21 May 2016 (UTC)
After a failure to save the Q127840 damaged, and no error displayed on my screen. I understood what happened there, when I saw the edits of user Edoderoo. Please, someone to check the code. --Francois-Pier (talk) 03:01, 22 May 2016 (UTC)
This happens to some items every day, almost all links get removed. A different user every time. Often I have the feeling that a user tried to merge two items. Edoderoo (talk) 06:35, 22 May 2016 (UTC)
I'm starting to see some Wikidata actions filtering into Wiki Commons which have very disturbing, and seriously unethical implications. For example, at Category:Sitta frontalis I see "English: Velvet-fronted nuthatch (linked to the en:wp article) .... Canadian English: Velvet-fronted Nuthatch ... British English: Velvet-fronted Nuthatch" (neither linked to any wikipedia article). Why is British English not considered synonymous with English (i.e., English as spoken in England)? And if all versions of English are to be considered different, why is there no entry for American [American English, en-us], nor for Indian English (which is the language used where this bird actually occurs)? What I find disturbing and unethical is the worry that the USA has as an act of cultural imperialism taken it on itself to hijack standard English [en] as being synonymous with American [en-us], when it should, if anything be synonymous with standard English English (i.e., English as spoken in England, where the language originated, "en-en"), or if not with that, then with no version of English taking supremacy over any other version of English (as is policy, though frequently not practice, at English wikipedia). - MPF (talk) 10:11, 22 May 2016 (UTC)
Hoi, English means "it does not matter what flavour it is". Effectively it is en-us. It is a given do not be a Don Quichote, it just does not work that way. Having alternate flavours will make it only more embarassing. Thanks, GerardM (talk) 10:57, 22 May 2016 (UTC)
Hi. I think what is happening here is that "English" means "whatever the enwiki article is called", and the rest are taken from Wikidata labels, so no particular variant of English is being artificially preferred in this process. There really is no "standard English", and hasn't been for a long time, as the majority of English speakers, both first-language and second-language, live outside the UK. -- The Anome (talk) 10:59, 22 May 2016 (UTC)
In which case, with [en] being for "generalised English", there also needs to be a separate entry for American English, as well as British English, Canadian English, etc., to prove even-handedness. - MPF (talk) 11:11, 22 May 2016 (UTC)
The more sensible option would be to deprecate en-GB and en-CA; the number of circumstances where this is actually useful is trivially small, and the potential for confusion is too high to make having them useful. Andrew Gray (talk) 11:15, 22 May 2016 (UTC)
That may be enough for English, but this cannot work for all the languages variants in the world. For example, we currently have a "Chinese" label, which is half traditional and half simplified Chinese, and often very different from Hong-Kong, Mainland or Taiwan labels. What I would like to see is a good fallback mechanism if only one of the different possibilities is filled, then it is enough to create a default EN-US locale, and maybe one day completely remove the default "English" or "Chinese" label. Koxinga (talk) 14:13, 22 May 2016 (UTC)
I guess the main problem is that we do not have enough users here to fill all these labels with local variants of a language. The differences between sv-se and sv-fi demands that we have to separate them sometimes. But I am not aware of one single user here who know enough about sv-fi (Swedish as it is spoken in Finland) to fill all our items with labels in that version. -- Innocent bystander (talk) 07:07, 23 May 2016 (UTC) (Who studied Northumbrian English in school)
The problem here (at Wikidata) is the interface, which makes it much easier to not maintain multiple variants. The sensible thing to do, in my opinion, would be to improve the interface so that maintaining n variants of a language is no longer n times as much work. For example, for variants which are mostly the same, it could collapse them by default and any changes would apply to all of the variants. When it is a term which varies regionally, there could be a way to expand it and edit individual variants (which would then stay expanded). That would allow regional variants while vastly reducing the amount of effort needed to keep the rest of the labels and descriptions in sync. - Nikki (talk) 09:37, 23 May 2016 (UTC)
That's mostly commenting on editing labels/descriptions here. For Commons, it would probably make sense to only show regional names if they're actually different. - Nikki (talk) 09:42, 23 May 2016 (UTC)
can confirm that this happens to me too, I start at the sitelink section, but then jump somewhere to the ID section. Using Chrome on Mac. --Denny (talk) 18:25, 20 May 2016 (UTC)
@Lockal: Thanks for the code. I'm have a large set of wikidata IDs (corresponding to organisms, as it happens), and would like to forward users on to the corresponding site link of that creature, where it exists for their language. But where an equivalent language sitelink doesn't exist, I'd like to direct them to the wikidata page, so they can look at the data present on WD, and check out the sitelinks for themselves, in case they speak another language. The simplest way to do that would be to direct them to the appropriate section of the standard WD page. But your code is an alternative, although then they won't see that there is a load of other data there too. HYanWong (talk) 00:15, 24 May 2016 (UTC)
New type of entity ? entity whose existence is not proven
Papremi (Q3894902) seems to be an antique capital of Egypt that has not been actually discovered but is mentioned into some old greek texts. I created an item for this kind of entities, Papremi (Q3894902) whose definition is precisely that. Please comment if it's a weak definition or if this is a duplicate. author TomT0m / talk page10:10, 22 May 2016 (UTC)
When I read the article about "entity", it does not look like something have to be proven for it to be an entity i.e. "assumed entity" == "entity". Only one man has said that Eric VIII of Sweden ever has existed, and I am not sure he was very well informed. It is possible that his record only was propaganda. -- Innocent bystander (talk) 07:19, 23 May 2016 (UTC)
@Innocent bystander: Then I named it wrong. I meant an entity that is mentioned on text that have or might not have existed according to recent historical datas or historian point of views. author TomT0m / talk page16:23, 23 May 2016 (UTC)
@TomT0m: I do not think it is important to use a scientifically exact definition of the word "entity" here. The important thing is to describe the nature of a claim. We have mythical entities like the norse god Odin, who is a historical person according to some records. This very Odin is then also #2 in one version of all lists of Swedish kings and an alien in a science-fiction universe. He is also a character in computer games etc etc. There is a continuum of myths/fictions/legends/ideas and sometimes they are entangled in each other. -- Innocent bystander (talk) 07:05, 24 May 2016 (UTC)
It would be really useful to be able to give geographical features an indicative radius. This would not be a precise value, but instead a rough-order-of-magnitude diameter of a very crude bounding circle (or even, say, 90% coverage circle) around their indicated centre point. For example, a reasonable figure for the "radius" of London (Q84) might be 30km, but it wouldn't really matter if a figure of 20km or 50km was equally representative -- the point is that it's not 3m, 30m, 300m, 3km, or 300km, or 3000km. (City of London (Q23311), by comparison, would have an approximate radius of about 2.5 km: I've added this experimentally to the item by hand, just to give an example.) This would be very useful for doing things like automatically scaling maps, and also as an extra data source to be used during cross-correlation of other geodata databases to quickly find possible matches and eliminate obvious mismatches.
We have several possible sources for this sort of data, which should allow large numbers of these radii to be assigned quite rapidly.
For example:
The CC0 Flickr shapefile database contains a large number of bounding boxes which could easily be used to generate approximate radii -- I am currently working on assignments of WOEIDs to Wikidata items which will enable these values to be estimated very simply from their database
Both Wikipedia and Wikidata have point data and nesting hierarchy data for places and administrative areas, which could easily be used to assign approximate radii to administrative areas such as parishes, counties, regions, and even countries.
I think we can do this with existing properties, if we use the current radius (P2120)or diameter (P2386) (see below) property, qualified with sourcing circumstances (P1480) of circa (Q5727902) to indicate that it is very approximate, with no indicator of precision.with a suitably large tolerance, (see below)
I think you're right about the use of a very large tolerance value instead of the "circa" qualifier. You might also be right about "diameter", as "radius" implies a well-defined centre, which is not always the case. I'd be happy with the use of either radius or diameter to define the approximating circle. I've amended my proposal accordingly. -- The Anome (talk) 11:18, 21 May 2016 (UTC)
Not quite. Consider, for example, near-linear features like Lac-Mégantic (Q142020), which may have much larger bounding radii than their area would suggest if they were round-ish. This is not to disparage "area", though -- we should have both! -- The Anome (talk) 10:44, 21 May 2016 (UTC)
Revised proposal: assignment of "diameter" to geographical features
Here's my revised proposal. We should be able to add the diameter (P2386) property to geographical objects, together with a substantial error bound. This should be interpreted as defining the diameter of an approximate bounding circle that would contain that object. If other editors agree with this, I should be able to provide this data for large numbers of geographical objects, and can apply for bot permission to add it. -- The Anome (talk) 12:06, 21 May 2016 (UTC)
Even if all we get is an order of magnitude level of precision then I can see the value in this. My only question is whether we want to use diameter (P2386) or a specific property ("bounding box on map") as geographical objects that happen to be circular or cylindrical, or contain circular or cylindrical elements, could end up with two diamater statements with very different values. Thryduulf (talk: local | en.wp | en.wikt) 12:46, 21 May 2016 (UTC)
There are pros and cons in both approaches, and no single number is really sufficient to summarize a shape. Really, the only proper way to record the shape of a thing is to have a detailed shapefile. I'd prefer something simple that sort-of-works right now, to something like latitude/longitude bounding boxes that look superficially more sophisticated but are actually less mathematically elegant (potential for misleading over-precision -- consider horizontal vs. diagonal features, which have the same bounding box but a difference of 40% in size, changes of scale on the ground, singularities at the poles and wrap-round at 180 longitude), and aren't much better in practice on typical wiggly-shaped geographic features. (The five parameters of a bounding ellipsoid would actually be better, but is way too complex a solution for not nearly enough improvement over circles or boxes) We can have both, if desired.-- The Anome (talk) 13:12, 21 May 2016 (UTC)
the shape of Stockholm
Shapes can look really awful. Where can we find a standard for how such a diameter should be measured? Will such a statement be given anything but guesses in the majority of cases? Most of the shape-files available out there are in most cases rough approximations. -- Innocent bystander (talk) 14:13, 21 May 2016 (UTC)
I absolutely agree. But rough approximations are better than nothing, and I can add them easily from a number of sources. -- The Anome (talk)
By the way, regarding algorithms, I can think of several. The simplest and crudest is to estimate some sort of maximum distance from the notional centre point we already have on Wikidata. If we start with a bounding box, for example, we can calculate distance to the furthest corner and nearest edge, double these, and take these as the limits of the interval for the diameter. Alternatively, we can do this with the centre of the bounding box, or the centroid of a shape. Finally, if we have a true shapefile, we can solve the smallest-circle problem, which gives a unique answer -- and would work fine for, for example, Stockholm. -- The Anome (talk) 21:18, 21 May 2016 (UTC)
Yes, probably Northernmost + Southernmost Latitude and Easternmost + Westernmost Longitude is more appropriate also for me. --ValterVB (talk) 15:48, 21 May 2016 (UTC)
"coordinate of northernmost point" etc. are not lat/long values defining a bounding box, they're extremal points. Extremal points are not always available, and they can't be derived from lat/long bounding boxes, which often are. But I can add these too, where known. My central concern is that at the moment, we have no concept of the size of geographical features, and any data, even very imprecise data (providing that lack of precision is clearly represented in the data) is better than none. I'm happy to generate anything the community desires, if it can be implemented easily. But extremal points are not viable for the sort of low-quality data we have freely available to us in the public domain. -- The Anome (talk) 21:10, 21 May 2016 (UTC)
OK, here's a straw man proposal. How about two properties: diameter of enclosing circle (a distance), and centre of enclosing circle (a point), that is not necessarily the same as the point given from other data, for example data which gives capital cities as representative points for countries... -- The Anome (talk) 21:30, 21 May 2016 (UTC)
A dedicated property which defines what the diameter means would be better than using the existing property, but I still don't understand why it has to be a circle and how that is better than storing the bounding box. Neither shape is perfect but boxes are more flexible than circles and the map things I'm familiar with (OSM and Leaflet) also use bounding boxes not circles with diameters. The fact that some of the data you have is already bounding boxes seems like even more of a reason to store bounding boxes rather than storing data derived from it, because then we have a citable source for the statements. - Nikki (talk) 09:08, 22 May 2016 (UTC)
I'm pretty sure some of what is above has been considered before (outside the context of the current sister project proposal). I'd have to dig up the property proposal though. --Izno (talk) 10:11, 22 May 2016 (UTC)
We certainly need something to serve the purpose of encoding the rough size of irregularly shaped geographic objects, or people are just going to encode "zoom level" which is the worst of all worlds. Diameter and lat/long bounding box seem like the best two options, at the moment: diameter is my favorite because it's just a single number, but if people like bounding boxes, I'd happy to use that, too. Without creating a new data type, we could encode the bounding box by two geographic points, one at the northeast corner and the other at the southwest corner. -- The Anome (talk) 11:18, 22 May 2016 (UTC)
Based on this discussion, I'm increasingly of the opinion that a bounding box is the best option. This allows the mapping program to optimise the display for a given shape in the current window size and shape - a portrait window and landscape window need very different zoom levels to optimally display Chile (Q298) for example, which is not possible to convey using only a diameter. Thryduulf (talk: local | en.wp | en.wikt) 19:15, 22 May 2016 (UTC)
@Nikki: Yes, that sounds excellent. If someone wants to make that proposal, I'd be happy to populate those properties for as many items as I have data for. I'm up to 28000 WOEID assignments already, so I should be able to populate at least those many straight away -- and I have up to a couple of hundred thousand more that I could potentially populate once I've resolved the WOEID assignments. -- The Anome (talk) 12:49, 25 May 2016 (UTC)
The precision on a geocoordinate itself was meant to capture this a little. A church would have a higher precision than a city, a city a higher one than a country, etc. Why does that not work? (This is an honest question) --Denny (talk) 15:42, 23 May 2016 (UTC)
Because the size of a thing is not the same either as the precision, or the accuracy, of its coordinates. In the spirit of my example above we might have an object 30m across whose position we only know to the nearest 1km, or somewhere like London, which is 60km across, but has a conventional notional centre point (at Charing Cross) with a position we know to within +/- 1m. Conflating the two kinds of interval might be convenient, but we're not doing ourselves any favours by doing so.-- The Anome (talk) 18:08, 23 May 2016 (UTC)
Several reasons: 1, if we want to reference a source, we can only say what the source says. It might be more or less precise than we think it should be for the size of the object, but that doesn't change the fact that that source says those are the coordinates. 2, coordinate precision is not a very intuitive thing. I expect that people entering coordinates based on maps will typically just copy whatever the coordinates say at the point they choose and use whatever precision is chosen by default when they add the statement, which is likely to be overprecise. 3, IIRC, some other tools which can add coordinates always select the same precision. 4, it doesn't give enough information to help people usefully fit the object onto a map (see the Chile example above). The first point means that we can't any have useful correlation between precision and size, the second and third mean that even if we did, it would be very hard to get useful data from it and the fourth means that even if the first three weren't an issue, it's not informative enough actually solve one of the problems we want to solve. - Nikki (talk) 19:35, 23 May 2016 (UTC)
From what I know, our Coord-templates in Wikipedia uses "type:city(1000000)" or something like that to help identifying how large zoom you have to make into a map. Of course, that is very rough, but help us separate Tokyo for random village. -- Innocent bystander (talk) 13:03, 25 May 2016 (UTC)
TREE[182624][][131] AND (CLAIM[31:532] OR CLAIM[31:659103] OR CLAIM[31:16858213] OR CLAIM[31:640364])
The problem is this list mixes settlements from different administrative levels no matter how I sort it. I want to have either:
the villages from a commune under the respective commune - no idea how to ask for that
only the villages and cities in alphabetical order - I expected to get that by removing CLAIM[31:659103], but that just makes autolist return an empty list, presumably because tree returns something else.
Working on classification? Important paper for you.
Hey :)
If you are working on classification here on Wikidata I can highly recommend the paper "State of the Union: A Data Consumer's Perspective on Wikidata and Its Properties for the Classification and Resolution of Entities" that was recently published at ICWSM. You can find a pdf version here. It gives a very good overview of the difficulties faced by someone who wants to make use of our classes and their relations. --Lydia Pintscher (WMDE) (talk) 09:43, 18 May 2016 (UTC)
I didn't read the paper in details but the main problem is that wikidata is the result of different classification systems from different wikipedias. Here we have to merge the works of different communities or work teams in WD into one unique classification scheme. Not easy to handle. Snipre (talk) 15:22, 18 May 2016 (UTC)
I did not read the paper in detail either, but my impression was that the authors have problems converting to a format they can use (like RDF) and that they dislike the fact that the contents of Wikidata are not frozen. The latter is fairly weird, as any database worth anything will keep growing. - Brya (talk) 16:51, 18 May 2016 (UTC)
I've been reading it - still will be thinking about it some more, but it has some fairly good points. The paper proposes a few specific properties, which we ought to discuss in detail under property proposals. But the main part of their problem concerns our instance of (P31) and subclass of (P279) class hierarchy. For some areas (human beings) the class hierarchy usage in wikidata is simple; for others (organizations) it is extremely complex. In both cases they point out things that don't make a lot of sense. For some things (events in particular) they propose splitting off separate wikidata items to attach location etc. data to. I think it would be good to start a discussion on these ideas in Wikidata:WikiProject Ontology. ArthurPSmith (talk) 17:12, 18 May 2016 (UTC)
@Lea Lacroix (WMDE): I did not have time and courage to read it entirely either but a major point is still in your hand :) Wikidata is still not full featured, and this is still a blocker, even if HUGE steps are done at this point : queries are one major way, of course to extract information and document preferred way, as they are both a sign that a model has been chosen by community, a specification of how information is intended be entered (a not well entered data won't fit show up in the query), and a way to monitor datas. And they are still weakly integrated at this point. I'd also like to higlight that Wikibase has almost integrated tool to model, which leaves community with wikipage for documentation. This is not an easy thing to discuss modelling issues and we are actually a very small number to do this (and still we achieve to have conflicts :/ ) I think we should not expect miracles if nothing changes on the tool part for communautary part.
Another critic : I think projects like WikiProject Reasoning could easily be expected to solve the dichotomy "specific classes versus specific properties" which is a fallacy if we can define classes in term of properties and can do inferences with the definition. Still a question of wikidata tools ... but community definitely need the help of tools considering the millions of items the project handles and the complexity of the task. And tools can definitely smooth certain conflicts (I'm thinking item splitting because some people are very territorial about their model and try to isolate themselves to try to control their items) I think for example that human (Q5) versus Homo sapiens (Q15978631) whom a statement was moved recently from one to another is a weirdness related to modelling that is both social and technical but that illustrate pretty much how things can be difficult on a project like this one ... author TomT0m / talk page09:39, 21 May 2016 (UTC)
Yeah that is certainly a big part of the issue and one of the reasons Markus and his team created SQID for example. And I think this is pretty good and helpful already. Do you have ideas/suggestions how we could integrate it better? Or any other things we can do that would make it easier? --Lydia Pintscher (WMDE) (talk) 10:33, 23 May 2016 (UTC)
@Lydia Pintscher (WMDE): For better integration of markus tool, integration to templates {{Item documentation}} or {{P'}}, {{Q'}} is definitely useful (this is half done) but they are not that popular in community. Gadgets could help, but they have to be opted-in so i suspect this is kind of power users matter. I could for example replace the reasonator template in my templates.
On thing I already talked about but that would be actually VERY useful is that classes, items,queries and properties become more integrated. For example if we associate an item to a query - say, a class, that if an item is in the result set of any of this query (interpreted in an analog way of class expression in SPARQL), the query appears somewhere on the item page without the need to explicitely state that the item is an instance of the class. Maybe in the form of an "inferred statement" ? I know for example that elastic search has a feature that could help to implement this that allows to recompute stuff on edits that touches an item ( https://www.elastic.co/fr/products/watcher ) and be notified. This could allow to maintain a set of those "inferred statement". author TomT0m / talk page11:56, 24 May 2016 (UTC)
Thanks for the clarification! I think those are good points. I have no idea how we do this well in the UI yet but let's see. As for gadgets: gadgets can also be enabled by default for everyone. (Though we should definitely not make use of this too much.) --Lydia Pintscher (WMDE) (talk) 20:24, 24 May 2016 (UTC)
@Lydia Pintscher (WMDE): After a still incomplete but more careful reading of the paper, I think the authors could better serve in an "on wiki" discussion than in a paper and participate here instead of taking a (very consumeric) approach for data extractions :) they have good points but some of their ideas were already discussed here. I'd have a few answers for them, but mostly what we lack here is workforce to organize and document the organisation of stuffs. I suspect we could very more easily clean the class tree of organisations with just a few edits than implement their idea, and I think some of their ideas would vanish in a few discussions here. author TomT0m / talk page10:54, 25 May 2016 (UTC)
Yes I totally agree. I am currently scheduling a call with them to see if and how we can make the things they found actionable and useful for all of you here. --Lydia Pintscher (WMDE) (talk) 22:03, 25 May 2016 (UTC)
@TomT0m: Before cleaning we should discuss and approve at community scale some classification schemes. We face 2 different problems:
Different classification system between different fields (use of a specific system for chemicals and another one for car models)
I agree with you about the fact that the origin of these different systems is mainly driven by closed team working on their items. We should focus on differences in classes inside a same field first. Trying to create one global classification system for all fields is a lost of time: the system will be too abstract. Snipre (talk) 11:37, 25 May 2016 (UTC)
@Snipre: I don't think so. Just using instance of (P31) absolutely does not mean that we use only one classification system. Second, it's way more flexible to create an item for a classification and mark every classes in this classification as "class in this cassification", per Help:Classification. On the over hand do you imagine asking for a property for every subfield in the universe ? It's WAY more hard and a lost of time to ask for a property for this. In the end it would be way more inefficient. author TomT0m / talk page13:36, 25 May 2016 (UTC)
Take chemistry for example. Existing ontologies actually use generic properties for classification. The problem in wikidata for chemistry is more inconsistent definition for the "chemical element" concept whose nobody actually seem motivated to sort out. I don't see a property to classify metal/non metals, another one to classify radioactive isotopes and so on. author TomT0m / talk page13:40, 25 May 2016 (UTC)
Thanks for the reference to this publication. I am currently trying to tidy up the classification of classification systems or more general knowledge organization system (Q6423319) and it first was a mess (see current graph). One problem is that Wikidata items are inherently fuzzy because they combine meanings from multiple languages that rarely fully overlap. But there is neither a commonly agreed upon classification in any domain. I would call myself an expert in the domain of knowledge organization systems but the number of similar items such as data model, conceptual model & ontology also confuses me. So if you are interested in backing up Wikidata with existing classifications, help to adding instances and types of classifications and taxonomies to Wikidata! -- JakobVoss (talk)
Dear Colleagues we have presented to the Wikimedia Foundation, the IEGrant proposal "ZARZUELA: The Hispanic Musical Theater in Wikipedia": Classification, Digitization and Description of Zarzuela Sound Archive and Iconographic Files in Wikipedia. Zarzuela is the hispanic genuine musical theater (XVII- XX centuries) a mass spectacle -dramatic, lyrical, musical and choreographic combination- for more than three centuries. As a result we find composers, librettists and scripts of La Zarzuela in the Iberian Peninsula, Argentina, Paraguay, Cuba, Mexico... There are also written Zarzuelas in the various languages of the Iberian Peninsula and America (castillian, basque, galician, catalan, majorcan, yopará, etc.) You can find the project linking: https://meta.wikimedia.org/wiki/Grants:IEG/ZARZUELA:_The_Hispanic_Musical_Theater_in_Wikipedia
If you like this project we would be very grateful if you sign the "Endorsements". We are looking for your support. Thanks a lot. --Jacinta Grey (talk) 17:05, 26 May 2016 (UTC)
Queridos compañeros, hemos presentado a los IEGrants un proyecto denominado " ZARZUELA: The Hispanic Musical Theater in Wikipedia" que mediante la clasificación, digitalización, y paso a dominio público pretende poner al alcance de todos el archivo sonoro e iconográfico de la Zarzuela y de paso recuperar muchísimos contenidos sobre el teatro musical hispano para la Wikipedia. Este es un trabajo de investigación y digitalización del Archivo sobre la Zarzuela del coleccionista Arturo Gil Pérez-Andújar, desarrolllado hasta ahora por la Fundación de la Zarzuela y al que nos hemos sumado varios wikipedistas con el asesoramiento de la Asociación Wikimedia España (las primeras gestiones fueron con Santi Navarro y luego nos ha ayudado muchísimo y pacientemente Jesús Tramullas). Podéis ver el proyecto completo en: https://meta.wikimedia.org/wiki/Grants:IEG/ZARZUELA:_The_Hispanic_Musical_Theater_in_Wikipedia
Hemos sido seleccionados y estamos en la última ronda pendientes de la decisión última de la Fundación Wikimedia.
Este es un proyecto global en el ámbito hispano y latinoamericano, pues la Zarzuela fue un fenómeno cultural y musical que mezcló artistas y creadores de las dos orillas. Y que tuvo repercusión importante, pues era muy popular, allende los mares y también en toda la Península Ibérica (llegando a escribirse zarzuelas en catalán, vasco, gallego y en alguna lengua nativa de América): Para conseguir que la Fundación Wikimedia elija este proyecto, entre otros, necesitamos todo el apoyo de los wikipedistas, por lo que si os gusta la propuesta, sería genial contar con vuestro "endorsement" en español o en inglés al final del proyecto: https://meta.wikimedia.org/w/index.php?title=Grants:IEG/ZARZUELA:_The_Hispanic_Musical_Theater_in_Wikipedia&action=edit§ion=17
Una vez que consigamos la IEGrant, empezaremos a organizar todo el trabajo de clasificación y digitalización, buscando los recursos bibliográficos y hemerográficos digitalizados que permitan mejorar, ampliar y crear nuevos contenidos. La idea es organizar un edithtaton sobre este tema durante 2017.
This is not a property, but an item. No discussion is necessary before creating items. If your question has another background, please specify. --Srittau (talk) 02:21, 22 May 2016 (UTC)
I don't think Magnus Manske's tool was meant to be used this way. It's just confusing in a multilingual environment to duplicate an item with labels and descriptions in dozens of languages and then edit the label in one language to say something different. No wonder Nono314 didn't understand the change. --- Jura06:28, 22 May 2016 (UTC)
Thank you Jura for shedding some light on what actually happened. Looks like every new technology will be misused...--Nono314 (talk) 11:20, 22 May 2016 (UTC)
Good question. Partitive relations and subclass relations for abstract concepts are a little hard to distinguish. I would tend to favor the use of part of (P361) in this case as I'm not even sure why we would consider, for example, biophysics (Q7100) to be a class in the first place - what could be the meaning of an "instance" of biophysics? biophysicist (Q14906342) is certainly a class that has instances, but the subject itself? Further discussion on this perhaps ought to take place somewhere under Wikidata:WikiProject Ontology. ArthurPSmith (talk) 15:41, 26 May 2016 (UTC)
@ArthurPSmith: This questions is related to the nature of sciences. If science is a class of theories that describe something, then for example theory of relativity (Q43514) is an instance of physics, that describe some interraction in the universe (and a definition of the universe itself). If I take a dictionary, a science is "a set of consistent set of knowledege related to the observation of some facts". Then this is a class of knowledge. It then make sense to classify "physics" as a specific kind of those knowledges, hence a subclass. author TomT0m / talk page16:31, 26 May 2016 (UTC)
@TomT0m, Dr.üsenfieber: No, I don't agree that science is a "class of theories that describe something". Much of science is not theoretical. And science encompasses the process of developing understanding, not just a particular state of understanding. "Physics" includes the ways in which physicists think about and model and look at the world - that's not a theory, that's essentially a "worldview". I would say theory of relativity (Q43514) is an instance of "physical theory" if we had such a class, or just "theory" as it is now. But I can't ascribe any meaning to the idea of an instance of "physics", or any other scientific field or academic discipline. ArthurPSmith (talk) 17:51, 26 May 2016 (UTC)
@ArthurPSmith: The definition is taken from a real dictionary, hence a reference (I took the first definition in a french dictionary). Any reference for you ?
dictionary.com : 2 . systematic knowledge of the physical or material world gained through observation and experimentation. 1 . a branch of knowledge or study dealing with a body of facts or truths systematically arranged and showing the operation of general laws: the mathematical sciences. => seems consistent with a class of theories, in an extended sense, and as a class of knowledge). 3 . any of the branches of natural or physical science. consistent with sciences as subclasses of each over, considering any discipline might be a subfield of physics if you are a materialist. 4. systematized knowledge in general. => the most generic class regrouping any theory or fact observed, includes maths.
Your definitions and references are fine; it's your interpretations that I disagree with. None of the dictionary definitions uses the word "class", and certainly none of them restricts the meaning to "theories" - specifically they talk about "observation and experimentation" and "systematization" which involve methods not theories or laws. Something "showing the operation of general laws" is not the same as the collection of those general laws, and one of those laws is not logically an "instance" of "showing the operation of general laws". Yes they can be arranged in some form of hierarchy, but hierarchy is not identical to class/subclass trees. ArthurPSmith (talk) 20:49, 26 May 2016 (UTC)
@ArthurPSmith: Then we are in a similar situation similar to religions, who are both a set of practices and a set of belief. Systematization is a polysemic word that can both refer to the action and to the result. You make a good point with this however, science is also a practice. That said, a class, by definition, is a set of object sharing some common characteristics. Here we are clearly in that case : we can recognise a scientific work if the work followed the best practices of the science of this time, and is characterized by the object of the research. This is something common to all scientific work. It's definitely not just a group of practices or knowledge which as parts as any new work can be recognized as science or not whether or not it meets the definition. That would be a theory for example, which could be divided into components (subtheories?), or just a set of observation, which also could be composite and definite (let's take the set of all observation a specific experiment measured). author TomT0m / talk page06:39, 27 May 2016 (UTC)
The first refers to the concept of "Album" as a form of book (a special French meaning?) while the second does indeed refer to picture books. If they should be merged it should be done with the articles on frwiki first. I don't know enough about this to say more, but as long as there are separate frwiki articles these can't be merged. ArthurPSmith (talk) 20:38, 26 May 2016 (UTC)
New mayor in Burjassot
I'm trying to change in Q55688 the alcalde. I've checked with the townhall website that it is Rafa García García. But I cannot put the name in the field. I cannot upload it. My options are:
Create a Q for Mr García García. (Note: García García is one of the most common surname combinations in the Spanish speaking world).
You can create the item, but try to add some more statements to it. Some of Q14135772 might apply to him as well. You could also create an item for "Alcalde of Burjassot" and add that to both with position held (P39). If you feel like it, you could do the entire list of alcaldes, etc. --- Jura09:41, 22 May 2016 (UTC)
A Wikidata concept doesn't point to a name but to a person. The person is named Rafa García García and thus the newly created concept should be named Rafa García García and not Mr García García. If you believe that there are multiple persons named Rafa García García write a description of the concept that specifies which Rafa García García is meant. Furthermore it's good if you add statements to the new Rafa García García concept. ChristianKl (talk) 10:52, 27 May 2016 (UTC)
Done that, more than once; and made suggestions on how to move forward, here and elsewhere. Nothing happened. Since I want people to know about this embarrassing state of affairs, I think you will find your latter statement is false. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits11:14, 27 May 2016 (UTC)
Approaches to keep "articles without items" and "items without statements" low?
Personally, I find that creating items with PetScan (Q23665536) and directly adding a statement can help keeping articles off these lists. Adding at least one statement (e.g. with P31) makes it also more likely that other statements are added.
For PetScan to work well, one would need to identify a series of categories that work well for a given wiki. --- Jura08:25, 27 May 2016 (UTC)
I follow Lsjbot to keep the record low, but he is hard to compete with. I have not become a friend of PetScan yet. I use Autolist, but a large problem is that Autlist is halted by serverproblems a little to often. -- Innocent bystander (talk) 09:54, 27 May 2016 (UTC)
I'd rather not use those as there is/are no statement(s) that could be added to all items that would be created for these. --- Jura17:10, 27 May 2016 (UTC)
New RfC about standards for property proposal discussions
I have begun a new RfC to try and prevent future controversy similar to that regarding the speedy creation of cites work (P2860). It seeks to determine if property proposal discussions should have a minimum duration and/or if property creators should be independent of discussions/events that resulted in the proposal.
However the second one also referenced by ky, tg, uk wikis. I don't know these languages but seems that from these only ru and uk are disambigs all others are not.
What is the best way to resolve this? I would say good step would be to delete Q4339855 and relink ru+uk to Q230259. Just merging Q4339855 into Q230259 is not a good idea as seems ky+tg are not disambigs but specific to philosophy. Fuxx (talk) 20:42, 27 May 2016 (UTC)
I'm currently doing a (slow) import of historic UK MPs. Many of these have unknown birth/death dates, and I'm keen to represent this data rather than just have a missing P569/570 value. Many entries (eg Jean Gery (Q3172296) have "unknown value" for these dates, which is exactly what I want to use, but I can't figure out how to add this - either in the editing interface or in an upload tool like QuickStatements. Am I missing a trick somewhere? Andrew Gray (talk) 22:43, 27 May 2016 (UTC)
Yes, when you go to create the claim after clicking edit, there's two icons on the left of the entry bar. One of those allows you to set a rank. The other allows you to set some special values; no value shows up there, as does unknown value, and some value is the default. --Izno (talk) 02:23, 28 May 2016 (UTC)
Aha, right! Got it. I think I'd only ever tried clicking on the 'rank' button there. Doesn't look like there's an easy way to do it via QuickStatements or similar, though. Hmm. Andrew Gray (talk) 11:01, 28 May 2016 (UTC)
I think the use of each of these properties is probably correct, except possibly speed. Numeric value is necessary because someone may want to know the speed; conversion to SI unit may be necessary because someone may want to jump from c units to another unit of measurement. "speed" seems okay since the domain is any item that moves. --Izno (talk) 12:16, 28 May 2016 (UTC)
Yes, it is probably "light" (or something like that) who should have "speed:299,792,458 m/s", while the speed of light should/could be regarded as a unit. -- Innocent bystander (talk) 16:56, 28 May 2016 (UTC)
Number of students as a property of Universities
Is it possible to get the number of students together with the date as a property of universities?
@Thryduulf: Yup, though we may want to refine it--most schools (in America) report "undergrad students" separately from "grad students", so perhaps there are some subclasses to deal with or another "applies to part" relation to establish. --Izno (talk) 14:07, 27 May 2016 (UTC)
Well reporting undergrand and postgrad students separately is just a matter of there being appropriate items with the relevant relations defined and having two claims for population rather than one. It wouldn't surprise me if the items already existed, but I haven't actually looked. Thryduulf (talk: local | en.wp | en.wikt) 14:12, 27 May 2016 (UTC)
"Student population" of a University can be misinterpreted as the number of people actually living inside the Campus area. Often our University-items are regarded as "instance of:Buildings" (I do not like that, but that is how it looks like) and buildings definitely can have a normal population. Here all post graduate students today are employed by the University, so I do not think they are counted as students at all. -- Innocent bystander (talk) 14:36, 27 May 2016 (UTC)
be_x_old is the old name of that Wikipedia, they changed it to be-tarask some few months(?) ago. Maybe nobody cares about the change? :) --Edgars2007 (talk) 20:55, 29 May 2016 (UTC)
There are multiple codes in use (the language tag for the content, the code used for the subdomain and the internal site ID) which don't always match. People are working on making them more consistent, but renaming wikis is something which has not really been done before (be-x-old to be-tarask is the first, as far as I know) and there are still some unresolved problems with it. I think the display of the sitelinks is probably covered by phab:T112647. - Nikki (talk) 11:49, 30 May 2016 (UTC)
Wikidata weekly summary #211
Here's your quick overview of what has been happening around Wikidata over the last week.
How would I add Out-of-school children of primary school age in pre-primary education to an item?
Hi all
I'm looking for advice of how to add 'Out-of-school children of primary school age in pre-primary education' to an item, the property Out-of-school children exists but I'm unsure how to break down the rest of the statement into qualifiers. Any suggestions would be appreciated. An example of a statement would be Out-of-school children of primary school age in pre-primary education in 1998 and would be added to the item for the area e.g a country or region.
I'm somewhat confused as to what you're asking: are these children in pre-primary education or not in an educational system at all? It seems like 'out-of-school' and 'in pre-primary education' are exclusive. In any case, you could have an item denoting an age range or something and do something like Jhalokati District (Q2093327): number of out-of-school children (P2573)="20015" (of (P642) <boys aged 3-5 years old>). Mahir256 (talk) 03:58, 31 May 2016 (UTC)
Thanks @Mahir256:, so they are in pre-primary education but they should be in primary school, its a measurement of someone being under educated for their age. --John Cummings (talk) 10:04, 31 May 2016 (UTC)
Editing from Wikipedia and co
Hey folks :)
Charlie has been working on concepts for making it possible to edit Wikidata from Wikipedia and other wikis. This was her bachelor thesis. She has now published it.
I am very happy she put a lot of thought and work into figuring out all the complexities of the topic and how to make this understandable for editors. We still have more work to do on the concepts and then actually have to implement it. Comments welcome.
Currently, I am using java to extract etymological relationships, building on DBnary http://kaiko.getalp.org/about-dbnary/ and the output is in (RDF) turtle format, e.g. (English word trance):
Any suggestion/recommendation from Wikidata or bot developers on how to structure the code would be helpful. – The preceding unsigned comment was added byEpantaleo (talk • contribs).
Maybe a separate installation of the Wikidata software on toolserver could be used to store them? Ideally, it could also draw data from Wikidata itself. Is there already such an installation? --- Jura10:36, 31 May 2016 (UTC)
The community should be aware that Property:P1887, for vice counties ("the standard geographical area for county based [biodiversity] recording" in the United Kingdom[11] and Ireland), which was created through the usual process, with consensus, was deleted without consensus. A discussion on the administrators' noticeboard asking for the deletion to be undone has been closed with no resolution, despite objections to that action, and a duplicate property proposal has been started (albeit in the wrong place). There is nothing in the documented policies supporting such actions. The duplicate discussion should be closed, and the property deletion should be reverted. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits11:02, 31 May 2016 (UTC)
The deleting admin obviously had a different view on the matter. Deletion requests (like property proposals) are in the end down to personal judgement. The other admins obviously understand this and did not see any blatant disregard in this case. Please accept that other people have different opinions on some matters and that there is no "right or wrong". --Srittau (talk) 13:15, 31 May 2016 (UTC)
No, deletion requests are not "down to personal judgement", they're down to consensus. Other admins have agreed that there was no consensus for deletion and not one single one has said that there was - but, remarkably, none has yet recreated the property. Perhaps they share your misconception about their flawed personal judgement overriding consensus? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits13:34, 31 May 2016 (UTC)
Nice forum shopping. I've moved the discussion from AN to Wikidata:Property proposal/Organization, as a way of getting past the bureaucratic bullshit of critiquing the form of the PFD close for over a month to the actual merits of whether or not that property should exist. Anyone is welcome to comment there. Ajraddatz (talk) 20:30, 31 May 2016 (UTC)
Alerting the community to an attempt - indeed, repeated attempts - to undermine its consensus. And yes, you "moved" discussion, as you so euphemistically refer to closing an ongoing topic in the face of objections, to the wrong forum: vice county is not a property of organisations. As I have noted in that forum, we have already had three discussions focused on the actual merits of having this property. The first found consensus to create the property; the second and the third - in the later case by your own admission - found none to delete it. Not one single person has said that they did. And you accuse me of forum shopping! Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits20:51, 31 May 2016 (UTC)
ISNI
ISNI (P213) identifier is the only Authority control identifier that shows up in the Statements section instead of Identifiers section. It also is displayed without the link to the isni.org. So it is very hard to verify if it is correct. See for example Q1994456. Can it be fixed? --Jarekt (talk) 13:23, 31 May 2016 (UTC)
I was always using activate "Authority control", but until today ISNI showed up as a text not a link, but it seems fine now. --Jarekt (talk) 15:14, 31 May 2016 (UTC)
I've created a new tool at wikidata-externalid-url to fix these linking problems via a modified formatter URL, if we choose to go that route. Feedback welcome. I may try switching some existing formatter URL's in the next few days (though others are welcome to try it first). ArthurPSmith (talk) 20:29, 31 May 2016 (UTC)
ISOCAT id would be better of as identifier than as statement.
@ChristianKl: you should sign your edits on discussion pages like that - also, there was already a section discussing ISOCAT ID (P2263) on that page, please check again and comment there. There was a reason it hadn't been on the list to convert to external id yet. ArthurPSmith (talk) 20:57, 31 May 2016 (UTC)
Research about user participation in Wikidata - call for participation (update)
Dear Wikidata users,
We are a group of researchers of the Web and Internet Science group of the University of Southampton.
We are currently conducting a research aiming to discover how newcomers become full participants into the Wikidata community. We are interested in understanding how the usage of tools, the relationships with the community, and the knowledge and application of policy norms change from users' first approach to Wikidata to their full integration as fully active participants.
This study will take place as an interview, either by videotelephony, e.g. Skype, phone, or e-mail, according to the preference of the interviewees. The time required to answer all the questions will likely be about an hour. Further information can be found on the Research Project Page: Research:Becoming Wikidatians: evolution of participation in a collaborative structured knowledge base..
Any data collected will be treated in the strictest confidentiality, no personal information will be processed for the purpose of the research. The study, which has submission number 20117, has received ethical approval following the University of Southampton guidelines.
We aim at gathering about 20 participants. Users interested in taking part or wishing to receive further information can contact us by writing to the e-mail address ap1a14+wikidata_user_studyecs.soton.ac.uk
Thank you very much, your help will be much appreciated!
Is it possible, given two job designations, to find the relation between them through Wikidata?
For eg: given CEO and VP- Marketing, is it possible for WikiData to give back something like,
VP - Marketing reports to CEO
or even works under CEO is okay! – The preceding unsigned comment was added by115.113.175.114 (talk • contribs). 11:06, 30 May 2016 (UTC)
Notified participants of WikiProject Names family name (Q101352) 
and after-name (Q4116295) 
Support I suggested that in the past. In cases of data imports, the fact that the database did had the value is a fact that won't become anytime wrong. The initial wikidata plan was to deprecate such claims if the values are considered wrong later. It might need a little common sense use not to bloat the database however. author TomT0m / talk page 07:01, 6 May 2016 (UTC)
⟩ damaged (P3081) 
Done. In case you find another wrong statement, just click "edit", then "remove". Mushroom (talk) 21:24, 8 May 2016 (UTC)
Oppose, such permissions are actually useless, using AbuseFilter to hold up illegal merging is enough. --Liuxinyu970226 (talk) 10:02, 13 May 2016 (UTC)
ecs.soton.ac.uk
