As you maybe know Lsjbot (Q17430942) is today massproducing articles on svwiki and cebwiki with GeoNames (Q830106) as one of the primary sources. I do not care what you like these articles. I do not want that discussion here and now. I want to know what we can do to keep up with the all the items without claims that produces for us here. I have tried for some time to add statements based on the categories on svwiki. Unfortunatly, the category-tree do not give us much information and many categories are very small.
Another way to do it, is to go to the the source itself. These are all the "types" available in Geonames. Lsj do not produce articles about them all, only those who are considered notable on svwiki and are well (enough) maintained at Geonames. I guess we can make P31-statements based on these "types". But which item best fits which "type".
Mon compte a été bloqué le 29 décembre 2015 à 19H37 par Grind24 . raison : spamming . blocage pour une durée infini .
J'aimerais savoir comment j'ai fait pour faire du spam , et comment ne plus en faire .
Le blocage pour une durée infini , je trouve sa un peu excessif . J'en suis navré de ce que j'ai fait ( mais je demande des explications )
et je vous demande de débloquer mon compte ( pas forcément dans l'immédiat [ c'est normal le spam est interdit ] )
@Snipre: C'est fait! Peut-être qu'il fait allusion au blocage du compte Skyrock98 sur mw. Comme l'indique la raison du blocage, l'utilisateur avait créer la page Mont Hikami incluant des liens externes et même un adresse mail pour le joindre. Ce qui est fortement considérée comme non respect de notre politique (des règles de mw). Voici une copie de ce qui était à la fin de la page <<SI VOUS AVEZ D'AUTRES INFORMATIONS , ENVOYEZ LES MOI VIA E-MAIL -> sonmail@exemple ( envoyez en sujet : Mt Hikami ) merci d'avoir pris le temps de lire tout ce qui a été écrit , et de m'avoir envoyer des e-mails>>. J'ai retiré l'adresse e-mail car je crois que wikidata ne l'autorise pas ni les autres wikis d'ailleurs. @Skyrock98: votre compte à été débloquée, mais la prochaine fois lisez bien les règles (scope) du projet avant de créer ou faire tout autre modifications qui peux être directement considérée comme du spam ou même du vandalisme. Je peux aussi vous envoyez le contenue de la page par e-mail au cas ou vous en aurez besoin. Merci pour la compréhension.-Grind24 (talk) 16:27, 2 January 2016 (UTC)
FEATURE REQUEST: Adding .json to a url should give the json format
I’m currently working on a Lua module (to replace wikitext, at least partly, in hu:Sablon:Cikktörténet), and it would be great if I could query the connected item (e.g. for featured articles in other languages, or to check if the badges passed as parameter match the badges stored in Wikidata). There is a function in the mw.wikibase module (client/includes/DataAccess/Scribunto/mw.wikibase.lua, around line 65) which gets the item ID based on mw.title.getCurrentTitle().prefixedText, this means for me that we could change it to content page name (since talk pages won’t have own connected items, so it won’t cause conflict). --Tacsipacsi (talk) 23:20, 1 January 2016 (UTC)
It is an old story where the software should update when new dumps are available. This is where the pipeline breaks. Thanks, GerardM (talk) 06:35, 2 January 2016 (UTC)
When adding statements manually, "property suggestor" suggests what property could be added next.
Since last year, it starts out with P31/P279.
There are a few other improvement that had been suggested. I wonder which you think are most needed. Maybe our feedback could be used to help the developers choose among the various ones requested. --- Jura13:25, 2 January 2016 (UTC)
Teaching machines to make your life easier – quality work on Wikidata
Hey, New blog post and some analysis is out. It turns out if we use ORES properly it can reduce the workload of reviewing human edits by 99%. If you help and label edits in WD:Edit labels. We can both efficiently revert vandalism and route good-faith newcomers to support and training. :) Amir (talk) 23:46, 2 January 2016 (UTC)
The Chinese articles seem to reflect the English ones : one about the cleaning product/disinfectant Dettol, the other about the chemical compound Chloroxylenol. Why do you think they need to be merged ? Koxinga (talk) 18:30, 4 January 2016 (UTC)
I understand the confusion now. The Chinese wikipedia used manual interwiki links and the links where wrong. I will correct them Koxinga (talk) 18:33, 4 January 2016 (UTC)
The 15.6 million items according to Special:Statistics corresponds to the number of "Good Articles" in grafana. There it's written good articles are items which are not redirects and not empty. What counts as empty is however not mentioned. It can't be the number of items without a statement (1.3 million [1]) and it can't be the number of items without a sitelink (2.6 million [2]) but probably it's the union of those two. --Pasleim (talk) 10:35, 4 January 2016 (UTC)
Wikidata:Data donation
Hi all
I'm working on the page Wikidata:Data donation which is meant for organisations interested in adding their data to Wikidata, I've started a new version from scratch here, I would appreciate comments and information, I am also thiking about if there could be a better name for the page e.g Data partnerships and importing data. Pinging @LydiaPintscher: (largest contributor to the page) and @Magnus Manske: who's text I have copied from here.
I think the main things I'm missing are:
Helping the organisations to understand what data they have that may be relevant
Deciding on if the amount of data that is relevant constitutes a copyrightable chunk i.e if they need to make the whole database available under CC0
How to make the data available under CC0 e.g a OTRS system for Wikidata
An example: English Wikipedia gives the name of the actor behind the German comedy character en:Atze Schröder.
German Wikipedia does not, since the actor threatened to sue Wikimedia Deutschland for giving his name.
Previously, he had won lawsuits against several German newspapers that gave his real name: he insists on remaining anonymous, and courts have sided with him.
What’s the situation for Wikidata?
As a layman, I’m afraid that it might be the same situation as German Wikipedia, since AFAIK there seems to be an association between Wikidata and Wikimedia Deutschland.
What country is Wikidata based in, legally?
—Galaktos (talk) 22:25, 22 December 2015 (UTC)
Why would the German Wikipedia not include the name? It's not the "Wikipedia for Germany" it's the "Wikipedia for speakers of German", whereever they may be. It is hosted in the USA; and WMDE is not responsible for it. Also, is there a reliable source for his threat; as I'd like to cite it in en.WP. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits11:32, 23 December 2015 (UTC)
Thanks for bringing this issue up Galaktos. There are different thoughts and arguments when it comes to legal questions. The server location argument has been brought up in numerous discussions on German Wikipedia in the past. However even German Wikipedia's own Wikipedia:Bildrechte states that copyright on German Wikipedia is evaluated based on German, Austrian and Swiss law, applying the most restrictive of the three. That's a perspective the German Wikipedia community has established and it remains to be seen if it can stand the test of time outside German Wikipedia. In July 2009 the district court of Hamburg rendered a verdict (LG Hamburg, Urteil v. 02.07.2009, Az. 325 O 321/08) on a Wikipedia related case including the question of jurisdiction. The verdict states "Das angerufene Gericht ist international und örtlich zuständig. Die streitgegenständliche Veröffentlichung ist bestimmungsgemäß auch im Inland und auch im Bezirk des Landgerichts Hamburg über das Internet abrufbar." Roughly "The court's jurisdiction applies on an international and local level. The publication belonging to the case is intended to be accessible through the internet, domestically and in the district of the district court of Hamburg." Other courts in other countries or Germany might decide differently about applicable jurisdiction and authority. That said, answering the question in which country Wikidata is legally based in might not be very fruitful. To my knowledge there are no cases or lawsuits in Germany that relate to Wikidata specifically. As for Andy's question: The German discussion page of the Atze Schröder article has a lengthy archive with a plethora of discussions about why and if the actual name of Atze Schröder should or should not be used in the article. I have neither time nor inclination to read them, but the information is there if you chose to read up about the issue. As far as the reliable source goes, here's what you are looking for. Regards, Christoph Braun (talk) 21:11, 24 December 2015 (UTC)
Comments: there is always the chance that a person will not sue "Wikipedia", Wikimedia Deutschland, or the Wikimedia Foundation (since in some countries it is already verified in court that it is not liable for the content), but sue the user who put the information on wiki, either in Wikipedia or in Wikidata. That does not mean that the information would be deleted even if the user will be found "guilty". Other users (presumably anonymous) could put the information back to the item/article.
But, there is a risk for reusers of Wikidata content. It is most probable that someone who use Wikidata content for some application, will not be aware of the issue. If he is based in Germany, he can easily be sued for something that it is displayed in his own website/application. I feel that this is something that should be discussed in the future: Data that does not put Wikidata and Wikimedia at risk, but can put reusers at risk: military information, confidential medical records...
FTM, the source of the real name is the user curated IMDb?
Another thing to consider aside from what German courts say: AFAIK his real name was never widely published in any reliable source. It was only "revealed" by some people on small, private websites and forums on the internet by digging through legal documents and filings (that don't outright say H.A. = A.S., just make it plausible). Even the English Wikipedia only uses IMDb as a source for it, which is not considered a realiable source for biographical information on Wikipedia as it's just user-generated content.So I'd say remove it until someone can provide a reliable source to back up the statement. --95.89.238.2401:57, 5 January 2016 (UTC)
Adding "imported from" much later
Hoi, a bot is adding "imported from" to items and statements. When this is done while importing, it makes sense. When it is done at a later date it is a fraud. Wikipedia is not a reference so it is wrong on first principles when it is suggested that a Wikipedia is a source. A source and the source the info came from are two very distinct things. Thanks, GerardM (talk) 06:23, 4 January 2016 (UTC)
I do not run such a bot, I reverted a few changes and it is suggested I seek community input. IMHO all these changes done in this way should be changed to for instance "stated in".. Thanks, GerardM (talk) 06:57, 4 January 2016 (UTC)
For me it's not logic to add a imported from Wikipedia, when not importing the data at the same time. How can you know it's imported from a specific Wikipedia? I don't think the user/bot should continue to do that. Mbch331 (talk) 09:51, 4 January 2016 (UTC)
Since we do differentiate between "legit" sources and those only purporting possibly unsourced or unstable information (both applies to data extracted from the individual Wikipedias), IMHO the consensus is not to use stated in (P248) in this case. So imported from Wikimedia project (P143) could either be renamed to something in the lines of "recorded in" or "reported by" or a sub-property with that meaning could be split of. -- Gymel (talk) 10:10, 4 January 2016 (UTC)
Gymel This is not a solution: if the article changes, and the data is deleted or modified, your reference system is dead. Wikipedia is not a source. So even if we accepted in the past to use such system it is time to accept to change and to follow the rules. All data imported from WPs without reference are already ignored by german WP and this will the same by the French WP. So using imported from Wikimedia project (P143) is useless work. Snipre (talk) 12:58, 4 January 2016 (UTC)
Therefore the distinction. But unless we remove here all unsourced statements and do not allow adding them again we really need some kind of "provenance" information for our dealings here. -- Gymel (talk) 13:04, 4 January 2016 (UTC)
We have to stop it because like mentioned by Mbch331 there is no guarantee about the exactness of this addition and then because the minimal to do is to add "imported from" with the revision number of the article containing the information. Snipre (talk) 12:50, 4 January 2016 (UTC)
So you suggest the bot should remove the statements first and then immediately re-add them in order to be allowed to annotate them with some approximation to a proper reference? -- Gymel (talk) 13:06, 4 January 2016 (UTC)
I see, he has been clogging my watchlist, too. My impression is, that User:Gbeckmann exploited the Personendaten template on German Wikipedia and added statements to wikidata. And in case the statement was already present, he added the reference as if just added. IMHO this improves the situation on Wikidata. -- Gymel (talk) 14:49, 4 January 2016 (UTC)
At Help:Sources#Different types of sources it is clearly stated, that "pages on Wikipedia ... are not appropriate as sources for Wikidata statements". So adding "imported from xxwiki" later (not after import, but just by comparsion of data in item and data in article) is not just misleading but also useless. --Jklamo (talk) 18:08, 4 January 2016 (UTC)
Not sure if it's clear to everyone, but if the same statement is imported from several sources, Wikidata includes this statement just once, but consolidates "imported from" in the source/reference section of statements. Thus it may appear that "imported from" was added just later, but actually it's an import from a different source that is consolidated. Obviously, we do encourage sourcing from multiple sources. --- Jura07:58, 5 January 2016 (UTC)
Bacterial information update
Hello,
I am currently working on a personal interest project were I try to link phenotypic information such as biosafety_level, oxygen_requirement, etc to genetic information.
I have looked through several sources (DSMZ, GOLD, EBI) but GOLD is practically one of the better ones (although there are errors and many empty fields in there).
Now I am curious if people are working on bots or integrating sources into wikidata. I am not sure if it is allowed to copy from GOLD and merge this into wikidata but if so this would be a great start.
I also encountered wikipedia and wikispecies (not sure why wikispecies exists) but these remain difficult to access when you are coping with 100s or 1000s of strains ( or am I doing something wrong?).
Are other people working on this or has this already been solved or coped by other resources which I somehow missed?
Maybe good to mention why I am looking into wikidata is because the integration of phenotypic information and genetic information is currently on a semantic (RDF) level and this would greatly interlink with wikidata.
@Jjkoehorst: You might be interested in our genewiki project. The aim of this project is to add content on genes, proteins, diseases and drugs to Wikidata from authoritative resources. We have recently presented a our work at the SWAT4LS conference. The ProteinBoxBot, as our bot is called, consists of a python framework, which is available on bitbucket. Whether or not it is allowed to add content from GOLD (or any source) depends mainly on the license which is attached to the resource. Best is to contact the database owners and ask whether or not it is allowed to redistribute their content under a CC0 license.
To expand on Andrawaag's comments, the GeneWIki MicrobeBot is currently adding gene and protein information to Wikidata, specifically for bacteria, from NCBI and UniProt. As noted on our project task page https://www.wikidata.org/wiki/User:MicrobeBot we are in the process of loading the 120 NCBI bacterial reference genomes. It would be great to add phenotypic data to the network of knowledge. Putmantime (talk) 17:01, 5 January 2016 (UTC)
Thank you all for the comments, regarding : @ProteinBoxBot: @MicrobeBot: both look very interesting and promising but maybe out of big data curiosity if you want to add protein information for all species or at least the fully sequenced once (bacteria currently have ±4.500 strains) it will consist of a lot of proteins. Would this not blow up wikidata or am I thinking to small or are you intentions to stick to 'reference genomes'? I will look into the projects!
@Andrawaag: thank you for mentioning the workshop as I am from the Netherlands it is not that far but might interfere with teaching responsibilities...
Also I will contact DSMZ and GOLD to see if they are willing to have the data into wikidata as it is currently stuck on my private RDF store which I find a pity as it needs to get out there.
Just had contact with DSMZ and we are allowed to upload their content into wikidata! Of course we have to explicitly mention where the sources come from and where possible add the articles from them. PMID: 24214959 and PMID: 26424852. Shall we start a DSMZ discussion group?
Hi. At en.wiki I searched for the item "en:D:Ream" (for technical reasons is at en:D Ream). I was redirected here to an article titled "Ream". I tried in other wikis and the same happened. This is because "D:" is an internal code. This has happened with other articles, the only one I can remeber is Lad: A Dog (a 1900s book), which redirects you to lad.wiki. I tried to create a soft redirect here at "Ream" but there is no option available. What can be done for this? Tbhotch (talk) 23:00, 5 January 2016 (UTC)
Hoi, when a scientific program is run on a University .. How do I indicate who works on that project and who is the lead scientist? Thanks, GerardM (talk) 06:58, 4 January 2016 (UTC)
UNESCO are interested in importing data from UNESCO Institute for Statistics into Wikidata, I think the education data would be a good place to start, I have some questions:
How do we decide which data withing the Education section is suitable for importing?
How do we decide if the amount and type of data that is relevant and could be imported constitutes a copyrightable chunk i.e if they need to make the whole database available under CC0?
the first thing to do is to link Wikidata to UNESCO databases by including identifiers. In that way you negate a lot of issues because people can go there and look at data at the source. Thanks, GerardM (talk) 11:35, 5 January 2016 (UTC)
Thanks @GerardM:, can you elloborate a little more? What I'm ideally looking for is a clear list of steps to get from an organisation being interested in importing data to the point where the data has been imported. John Cummings (talk) 15:29, 5 January 2016 (UTC)
Have a look at Wikidata:Property_proposal/Term#Datasets. But in the case of UNESCO data I have some doubts about the possibility to integrate this kind of data. This is high density data and just by selecting some data you can transform an item in a heavy object which can't be opened due to its size: just take 20 different topics times 10 or 20 years, including for statements the reference and a qualifier (the year) and you can have 200-400 heavy statements. Time data series with a high frequency over a long period are a challenge because Wikidata is not curently able to store that information in a very efficient way
We need a new datatype for tables or datasets in order to reduce the number of statements and to keep the items as light as possible for data loading. Snipre (talk) 23:42, 5 January 2016 (UTC)
Thanks Snipre, I understand what you mean about yearly updates of data, perhaps a workaround would be to import only the most recent data and then work backwards once a solution has been found? That way all the legwork has been done with adding new properties etc. John Cummings (talk) 13:41, 6 January 2016 (UTC)
Dear all,
the deadline for the call for proposals for Wikimania 2016 has been moved on 17th January 2016, so you have 10 days to submit you proposal(s). To submit a presentation, please refer to the Submissions page on the Wikimania 2016 website. --Yiyi.... (talk!) 09:36, 7 January 2016 (UTC)
ISNI
Hoi, I found that ISNI records are often mall formed. For me they are in the sidebar and I do not know how to toggle this. Also.. Is there a bot that can fix this? Is it fixable by bot? Thanks, GerardM (talk) 09:20, 31 December 2015 (UTC)
They should do when shown, not when stored. When You click on a source, the website should open correctly. Thanks, GerardM (talk) 16:14, 1 January 2016 (UTC)
a bot is changing the ISNI number and consequently it no longer works in Reasonator.. This is NOT a good idea. Thanks, GerardM (talk) 14:22, 7 January 2016 (UTC)
the bot preserving a standard format. There are 270,000 statements containing spaces. Everyday around 10 ISNI records get added with double spaces and another 10 records without spaces. If no one is correcting it, it will end up sooner or later in a mess. --Pasleim (talk) 14:43, 7 January 2016 (UTC)
It is irrelevant that 270.000 statements are wrong? There should be no spaces at all. With spaces linking to ISNI fails. That is what is essential. THanks, GerardM (talk) 22:08, 7 January 2016 (UTC)
So 270.000 statements never worked in Reasonator...? Looks like Reasonator should adapt to the correct format. --Succu (talk) 22:37, 7 January 2016 (UTC)
Showing the "correct" format is just a formatting trick. Easily done if you want to do that. Disparaging Reasonator does not make for a fine argument. It actually is an additional reason to correct the data into a workable format. Thanks, GerardM (talk) 06:02, 8 January 2016 (UTC)
It may be old but it is definitely not "well" established. You suggest that it other data consumers have a usage... Please show me one. It is wrong, you know it. Thanks, GerardM (talk) 20:52, 8 January 2016 (UTC)
Thanks Vlsergey. Citing Pasleim from above „There are 270,000 statements containing spaces.“ But you are ignoring this fact, GerardM with vapour like „Disparaging Reasonator“ or „It is wrong, you know it.“ - Well done. --Succu (talk) 21:25, 8 January 2016 (UTC)
First item of series
If we have a series of items all linked with part of the series (P179) to a series, is there an easy way to determine the first item of the series? For TV series this would be the pilot or episode 1. Items may have qualifiers for previous/next or just a qualifier for the series ordinal or properties with dates. --- Jura10:14, 3 January 2016 (UTC)
It's a bit complicated to query, but it works. Only a few item have this currently set (about 20-30 series). Maybe we should seek to define a more direct way. --- Jura06:33, 4 January 2016 (UTC)
It could be, but these are mostly TV seasons (1295), many I just defined. Besides, the first item could be either the pilot or episode "1" or "S1E1". There are currently more episodes using "novalue" than "1". --- Jura07:23, 4 January 2016 (UTC)
Actually series ordinal (P1545) is string datatype so unfortunately the order in sequence value might actually be "S1E1", which is imho a very bad idea since the all point of having an order is that it should be naturally orderable. This should be a real number so that we can use "order by" in sparql naturally. For "in beetween" numbers like 1b, we should use fractional values, for example 1.5 author TomT0m / talk page20:10, 7 January 2016 (UTC)
Yes, that is exactly what I have used in many cases. Properties like inception (P571) is often hard to use. An entity may change it's nature many times and the articles cover many phases in their life. There can here also be interwiki-conflicts between different projects. Svwiki has different articles about Swedish Municipalities before 1971, when there existed many more types of communes that it does today. There is therefor one Swedish article about Stockholm municipality 1971- and one for 1862-1970. I am not aware of any other project having any such article. But Stockholm was founded long before that! -- Innocent bystander (talk) 09:27, 6 January 2016 (UTC)
I think it would also make sense to mark the non-historic relations as preferred values so that by default queries will not find historic relations. (For example, a query looking for all Dutch municipalities would currently find Bussum (Q9909) if there is no special-casing for end time (P582).) --Srittau (talk) 19:22, 7 January 2016 (UTC)
Queries can be quite difficult. I was thinking about a query help desk to have a central place to get help with queries. Do you think this would be useful? Would you be willing to help out? Multichill (talk) 18:28, 7 January 2016 (UTC)
Class A war criminal (Q5369948) include Wikipedia article in Chinese, Korean and Japanese language. Things are fine for Chinese and Korean Wikipedia, however interwiki links aren't displayed for Japanese Wikipedia. I don't know the reason for that so I try to delete and readd the ja wiki link in wikidata, but while it show me the Q5369948 entry have three item when I read the page, once I enter the edit mode, it show me there are four items, and consequently after I delete the ja sitelink from the item, I can't save it. How to fix that?C933103 (talk) 18:59, 8 January 2016 (UTC)
When I got to jawiki now I see the interwiki links. Next time they don't show up, the only thing you need to do is add ?action=purge after the URL on jawiki. (So in this case it would be https://ja.wikipedia.org/wiki/A級戦犯?action=purge.) That should be enough to restore the interwiki links. If everything is fine on Wikidata and the interwiki links work on other linked wiki's, usually it's a cache problem on the server. Mbch331 (talk) 21:09, 8 January 2016 (UTC)
Get qualifier on item on wiki that supports access to Wikidata?
I have been working a bit with creating infobox templates that automatically collects data from Wikidata, like movie properties for instance. However, is it possible to fetch the qualifiers for a certain statement as well? Looking at the movie example, a movie can have different release dates in different countries. I guess you would specify different release date statements and add a country qualifier to which country it applies. How can this information be displayed in a template on a Wikipedia version that supports access to Wikidata? Are there any examples out there? //Mippzon (talk) 15:07, 5 January 2016 (UTC)
(redkonf) @Mippzon: Ja, det går. Det kräver dock att man använder lite klurigare metoder än att bara man använder sådant som {{#Property:PX}}. Vi kan tex bygga ut vår Module:Wikidata (på svwiki) så den stöder dina mallar. Det går också att bygga specialdesignade moduler för just dina mallars behov. -- Innocent bystander (talk) 15:21, 5 January 2016 (UTC)
Thanks for both the general answer, and the local one! I'll try to reach out to @Innocent bystander: on sv-wiki for some guidance.
At Wikidata:WikiProject Movies/Infobox, there are few infoboxes for films, but most only display basic information. If you test (in preview) on the articles of Q15624215 in cs/hu/ruwiki with the empty infobox, you will see that they merely output the full list of dates.
Thanks! Those are some really good examples, at least I got some idea of how this works. I'll continue to test around a bit. //Mippzon (talk) 16:26, 5 January 2016 (UTC)
On the French Wikipedia there is a user who is actualy a race cyclist, and he has programmed a lot of infoboxes with WikiData and Lua. I tried to find back his username, but failed so far, but he created some nice examples last year, I guess that is exactly what you was looking for. Edoderoo (talk) 20:53, 5 January 2016 (UTC)
Thanks @John Cummings for the ping. I'm afraid I don't have an answer at the moment, but it sounds very similar to a problem I've encountered with trying to access the 'first light' qualification for the SPT from Q1513315 - the core of the problem being that qualifiers may not be as standardised as properties, and you can't check for property numbers since they may be repeated. It may be possible with en:Module:Wikidata, but I haven't figured it out yet. Also pinging @RexxS, who may be able to help. Thanks. Mike Peel (talk) 19:25, 9 January 2016 (UTC)
opening this thread to document and continue our discussion from yesterday with @almereyda and @toka
Context:
Projects such as TransforMap or inventaire.io could make use of an opened linked, & contributive database with a different scope than Wikidata but re-using its properties and entities, and its interface
The obvious solution would be starting new instances of Wikibase but it currently miss federation features, namely:
allow to make use of both local properties/entities and remote/wikidata ones
allow to customize the naming system, replacing the incremental P and Q ids by uuid-like ids to make properties/entities imports among instances easier
The Wikidata/Wikibase team has added federation to the roadmap but made clear that it would take time to get there. We couldn't find the dedicated issue or thread though.
For users of MediaWiki, it should be possible to define "internal" properties and items (in separate namespace) that combine external data (from Wikidata) with internal data (on their Mediawiki instance). If I recall correctly, there may already be a phabricator task for it. A possible use for this at Wikimedia could be Commons. --- Jura12:32, 9 January 2016 (UTC)
What you could do fairly easily, I think, is to create a new data type "Wikidata entityid" that clones the current "wikibase-entityid" one but instead of linking to the the local Wikibase entities it would link to Wikidata ones. Then, in order to query both dataset, you would just have to merge both your local instance RDF dump and Wikidata RDF dump into a unique SPARQL endpoint (after making the RDF export tool support your new "wikidata-entityid" datatype that would output Wikidata URIs). Tpt (talk) 17:16, 9 January 2016 (UTC)
It seems that in RDF technologies every database could have its own ids that could be mapped with a "owl:sameAs" or similar property. No need to mess with the ids ... author TomT0m / talk page19:54, 9 January 2016 (UTC)
Ok will do, let me do a little digging to make sure it doesn't already exist with another name on some other wiki. Reguyla (talk) 21:18, 8 January 2016 (UTC)
I've created a new item for the eswiki page, and added some claims to him: Aaron Anderson (Q22006480). I couldn't find any other items on wikidata for him. If you find any articles about him on other wikis, please add the sitelinks to the item. Silverfish (talk) 18:11, 10 January 2016 (UTC)
Automatic pattern-based label translation
Hello, many items have standardized labels, for example 'en:1999 in art', 'de:Kunstjahr 1999'. Its can be used for translation pattern generation, for example 'en:# in art' ↔ 'de:Kunstjahr #'. The pattern can be used for adding missing labels, for example label 'de:Kunstjahr 1924' can be generated for 1924 in art (Q372520). ~30000 ru labels were added using this approach today. Quality of generated labels is looked good. I have plans to apply this approach to another languages. But my language knowledge is limited. Could somebody verify quality of this approach for another languages: de, fr, it, es, en. The approach can be applied for another languages too. Current algorithm uses Arabic numbers as variable part, but it can use country names, Roman numbers and etc. also. — Ivan A. Krestinin (talk) 23:25, 9 January 2016 (UTC)
I would suggest to check out the date.py script in pywikipediabot - link. It contains date data in several languages. Back in the day, when interwiki links where updated on the wikipedias themselves, this script was used by interwiki bots.--Snaevar (talk) 02:55, 10 January 2016 (UTC)
For italian version: some problem with upper/lower case, so need add some rules: if translated label start with number is ok, if translated label start with Categoria: is OK, if translated label start with 2 or more upper letter is OK, is translated letter start with letter dot letter (ex. U.S) is OK, other rule but more complex, if item is a album (Q482994) or subclass is ok, if is human (Q5) is OK, if is astronomical object (Q6999) and subclass is OK, just for start :) I have a lot of other rules. If you can follow this rule you can update the Italian label, then if is possible to have the update list, I can check again. --ValterVB (talk) 09:33, 10 January 2016 (UTC)
Ok, I try to add some checks. Bot uses existing labels for pattern generation. Bot settings are very strong now: all found labels must have the same pattern and number of existing labels must be 20 or more. If some labels must be lower cased, the best way is fixing existing labels. Bot does not replace existing labels in any case. — Ivan A. Krestinin (talk) 11:36, 10 January 2016 (UTC)
The article Dassel in the german wikipedia (de-wiki) recently lost its link to wikidata and thus doesn't link to other languages any more. Can someone please find out what's the problem with wikidata in this case? Thanks!--Wibby02 (talk) 10:48, 10 January 2016 (UTC)
which would solve most cases. It does not seems to have received negative comments on the principle of doing that, yet it seems it needs more support.
Yet I am still in dispute with Fomafix and Andreasmperu that want to restore the intuitive but not really correct claims. This situation has to move, can we do something at least here ?
Also, I agree that in that case, has part(s) (P527) is not correct to define the class. I like the proposal for the disjoint union class, which is equivalent to owl:DisjointUnion and is therefore clearly useful and sensible. --Srittau (talk) 17:34, 10 January 2016 (UTC)
Then please both support the proposal if you did not yet :) I guess it does not need a lot of support to be created yet. author TomT0m / talk page18:19, 10 January 2016 (UTC)
Continuation of discussion if Regierungsbezirk (Q22721) should be subclassed for every state of Germany
@91.9.123.135:/@178.4.4.253:, @GerardM:, @Mbch331:, @Edoderoo: I would really like to hear the opinion of any of you regarding the discussion we had here before christmas.
After Gbeckmann and I wrote our opinions (pro-delete, at least for some of the disputed items), the discussion turned silent and the items we discussed are all still in existence. I didn't consider the discussion finished, so I waited for an answer from anyone who expressed, that these newly created items should be kept. As I noticed today, that the discussion has been archived, I thought I'll repost here to continue the discussion. I'd at least like to hear any comments about Gbeckmanns, Srittaus and my points of view, now that we have explained them to you.
I don't really get, what you mean by that. Is it confirmed, that the IP is Tobias Conradi?
Do you mean, that the edits by the IP should be reverted? I did that and definitely won't do it again, because admins said, that I'll receive a block if I continue. That's the whole point of this discussion, I want to discuss this matter, because the IP user raised concerns about me going ahead and deleting items created by him/her.
Do you mean, I should ping User:TobiasConradi instead of the IP? Or that I should mention this on the RFC page although it is closed?
Could you please elaborate a bit, because I'm not really getting it. And I don't really want to sift through the whole RFC, as it's quite long. --Floscher (talk) 14:36, 2 January 2016 (UTC)
Yes, I know. So I shouldn't have pinged him/her? I wasn't aware that I shouldn't ping blocked users, just wanted to enable him/her to react (even if it's unlikely that an IP receives notifications after 2 weeks of inactivity). And I understood the mention by Pasleim rather as a hint that it could be Tobias Conradi, rather than a hard fact that the IPs are him. Anyway, thanks for bringing that issue to attention again. --Floscher (talk) 15:50, 2 January 2016 (UTC)
If I remember well, some ppl wanted to delete the items, because there was no wikipedia-article related to the item. But that is definitely not a good reason to delete, many items exist just to relate to other Wikidata items, like mayor of Woerden (Q20130623), used to connect all these items. The same will happen to these states of Germany, that might have politicians, towns, etc, related to them, even when the state nowadays does not exist anymore. Edoderoo (talk) 13:54, 2 January 2016 (UTC)
My main argument was that these are in my opinion not notable, with the exception of those with Wikipedia articles.
I think there is no question whether to keep the items for the states themselves. However, items were created for the government subdivisions of each state. To take an example, we have articles for :
Exactly, my point was not about notability. My argument was that the Regierungsbezirke (etc.) are conceptually the same in the different states. Therefore the distinction between the instances should be made via properties. I fear that otherwise we go down the Commons route of having intersection classes for everything. (That make sense in Commons categories due to the lack of proper querying tools, which we have here.) In other words, for a particular Regierungsbezirk the question "what is it?" can be answered by instance of (P31)Regierungsbezirk (Q22721), the question "where is it?" can be answered by located in the administrative territorial entity (P131). --Srittau (talk) 14:59, 2 January 2016 (UTC)
Thanks for commenting here again, I'm able to completely subscribe to your argumentation. It's just that I chose the different argumentation route to rely mainly on notability, because I felt that notability would give me a more solid foundation to argue against keeping those items. --Floscher (talk) 15:50, 2 January 2016 (UTC)
Ok, I still don't see major arguments against deletion of those items. So I posted now on the discussion pages of the government regions, that I consider requesting deletion again in about one week. --Floscher (talk) 13:48, 4 January 2016 (UTC)
Hi. I've marked some statements about Yugoslavia (Q36704) and Czechoslovakia (Q33946) being neihbors with Romania (Q218) as deprecated, but the change was reverted with the explanation "it is past data, not wrong". However, Help:Ranking#Deprecated_rank says: "The deprecated rank is used for statements that are known to include errors or that represent outdated knowledge." So in my opinion, those past neighbors should be marked as deprecated, but I would like to find out what the practice is around here.--Strainu (talk) 23:04, 10 January 2016 (UTC)
I understand the term "outdated knowledge" to mean knowledge that has been proven to be wrong, not "knowledge about outdated situations". From what I understand: "preferred" = current value, "normal" = other values, "deprecated" = values that are (and have always been) wrong. --Srittau (talk) 23:07, 10 January 2016 (UTC)
I would say: Add "Romania" as "preffered" neighbour to "Yugoslavia", but "Yugoslavia" as "normal" neighbour to "Romania". This since Yugoslavia had Romania as neighbour until the end of its existence. For Romania, Yugoslavia used to be a neighbour, but isn't any longer, and therefor should have a lower rank than the statement that tells that "Serbia" today is its neighbour.
The regular API has a parameter to get a redirected item or not. To my humble opinion, only ill programmed external tools will fail on redirects. Edoderoo (talk) 17:04, 11 January 2016 (UTC)
Hi, there is a edit war between user:HolgerFroese and I so I try to find consensus before act.
See item diaspora* (Q251649): it has both software version identifier (P348) with publication date (P577) as qualifier for each version, with the latest version marked as preferred, and publication date (P577) as single statement for the latest version release date. I remove the single statement for the latest version release date as it is a duplicate of the one in software version identifier (P348). However, HolgerFroese keeps on inserting it because he says that removing this statement will break «ALL Wikipedia Diaspora articles in any language, except Italian», as Italian Wikipedia uses Template:WikidataQ (Q18808659), a template based on a module to fetch information from qualifier, while (English?) Wikipedia doesn't.
I think that this is a problem for English Wikipedia, which should work to create such a template in order to fetch data from qualifiers. We must not have duplicate data on each item. So I'm looking for consensus to remove the statement. Do you agree with me? --★ → Airon 9008:59, 11 January 2016 (UTC)
While I agree that templates should be instructed to fetch data from/based on qualifiers, there is no issue in adding first publication information (also) in P577 and it seems advisable to me to have both. Nemo10:45, 11 January 2016 (UTC)
Date of first publication can be a normal claim, all other versions should only be qualifiers. As the software isn't published for the first time, but just updated to a newer version. So the date of first publication is only for a specific version and that's what qualifiers are invented for. To give extra information for a specific claim. If publication date (P577) would be the date of the latest version, then you are saying the other versions were published before the software was published and have a negative age (as if someone has a birthday before he was born). Mbch331 (talk) 11:20, 11 January 2016 (UTC)
Using publication date (P577) separately for the first publication (which usually conincidences with publication of the first version) can be useful, even if it is redundant. It allows to query for all works with certain publication dates, without having to special case software products, especially since many software products can not be described using versions, e.g. websites). --Srittau (talk) 11:42, 11 January 2016 (UTC)
My intention is to keep all Diaspora* articles up to date in a central place. It seems to me the Italian solution is actually the best one but doesn't work on most (all other?) Wikipedia languages at the moment. Keeping the publication date as Property, even it is redundant, will keep all other Diaspora articles up to date. If there is an other solution please help. HolgerFroese (talk) 16:24, 11 January 2016 (UTC)
Ok, I think that there is consensus in using publication date (P577) for the date of the first release. Moreover, it is difficult to automatically understand which is the first version, as not always is composed by number and not always it is used the alphabetical order.
@HolgerFroese:: I'm sorry but I have to repeat myself. It's not our problem if other wikis uncorrectly implement their templates. Ask your wiki's developers to build a module like Module:Wikidata (Q12069631) to enhance basic {{#property:}} tag. Maybe in the future the #property tag will have all the feature used by itwiki's Module:Wikidata but we can't permit to have useless data or useless items. Best regards, --★ → Airon 9018:21, 11 January 2016 (UTC)
@Airon90:: There is a 'date of release' for publication date (P577) available also. There is an simple question: How does this affect the Italian article or is this issue for having duplicate data only? I don't know if you are a diaspora* user but if - there are more user in other languages too. Keeping all articles up to date should on your behalf also. It is of course a 'useless item' for the Italian wiki but required for all other languages at the moment. I think this is more a debate on principles than a question of advantage of Wikidata. HolgerFroese (talk) 18:47, 11 January 2016 (UTC)
A consistent data set is very much an advantage for Wikidata and its reusers, including Wikipedias. If the templates are wrong, they need to fixed, not the data changed to accomodate the broken templates. --Srittau (talk) 22:26, 11 January 2016 (UTC)
Removing publication date (P577) or changing it to the date of the first release don't affect it.wiki because the creator of the module worked hard to create a good tool
I may be a D* user or I may be not. The point is: I don't work on Wikipedia and Wikidata to make propaganda ;)
The item you created is useless as per Notability policy, not because it is not useful for Italian Wikipedia
I do care about up to date articles on Wikipedias. That's why I work on Wikidata and that's why I suggest you - and I keep on repeating it - to ask your wiki's developers to create a module to fetch data correctly
Consensus in using publication date (P577) for the date of the first release has been found. Please, don't rollback my edits again.
In the FRBR model spirit, there should be one item for the projet, plus one item for its releases. It would be consistent with the "work/edition" scheme in the books field. author TomT0m / talk page09:25, 12 January 2016 (UTC)
I am not sure that is necessary, especially for minor releases. Software releases are much more frequent than editions and not really comparable. In a more specialized ontology than Wikidata this could make sense. (Also note that looking at it from a data modelling perspective, we already have anonymous items for versions, introduced by using qualifiers on them.) --Srittau (talk) 10:57, 12 January 2016 (UTC)
There is no real "anonymous items" as this item represents a statement. So to use one release in another statement we need to use the exact same statements. I'd prefer to have a real Wikidata item to make things more regular, as in the end there Will be (and is) items for some releases. author TomT0m / talk page11:26, 12 January 2016 (UTC)
Distinct element per translation (movie or book) is a good approach -- because each translation has it's own date, voice actors, publisher, place, etc. Sometimes -- own codes and even small changes in duration (in movies). I.e. a lot of values. Distinct element per different publications of book is good as well -- because they have a lot of differences. Publication (part of), editor, publisher, year (date), place, ISBN, total pages, etc. Thus such distinct element is good (it's a "pro"), because a lot of data are NOT duplicated and it would be hard to squeeze those different data into single element. Despite the fact that having single element is very nice in some cases -- no arbitrary access, much simple to access data, simpler interwikis. (i.e. "contra") Having those arguments in mind, it's okay to have multiple elements per book, but community is not ready to have multiple elements for movies -- because usefulness of such elements is not good enough (comparing to single element). Having multiple elements per single software is completely nonsense. We have only 2 distinct values per edition -- date and version number (sometimes -- platform). This can be very easy squeezed into single element. Moreover, since absence of reverse links in Wikidata (unable to get "linked to" elements via LUA API) we are still required to hold all links to all versions from original element. Having > 50-100 versions will make this construction unusable from Wikipedia due to LUA limits. -- VlSergey(трёп)12:29, 12 January 2016 (UTC)
There is many many thing we could say about a release : added feature, removed feature or complete list of features. Right now the feature model does not imho work really well. Plus there is a few proposal that are tight to a release : the name of the archive file, url in the official repo, dependancies ... Software distributions with a specific version also have a specific version of the sofware, very often. author TomT0m / talk page12:36, 12 January 2016 (UTC)
So... who will actually use all of this? Is this information present in some Wikipedia already? I suppose no. Later we can create new element per each version of software, but so far is good for almost everyone, who actually uses or may use Wikidata. -- VlSergey(трёп)12:54, 12 January 2016 (UTC)
Hi, everyone. I was wondering if/how I could insert a time value with precision higher than 'day'. For instance, it is verifiable that 1977 Vrancea earthquake (Q751139) occured at 21:22:22, but all I can input for point in time (P585) is the date, not the time. I went around looking for other high precision times, such as the launch time of a space vehicle, but all I could find was data with precision to the day, not hours, minutes or seconds. Is this a known limitation? Andrei Stroe (talk) 12:55, 12 January 2016 (UTC)
Great ! That and StrepHit suggest a bright future for references in Wikidata. Do we have ideas on how well the web has microdatas atm ? Big website or database ? author TomT0m / talk page19:11, 5 January 2016 (UTC)
what I don't understand is how you intend to use this. As it only can add source to already existing statements, you should use it multiple times on the same item.
Also, could it be used to add claims. I saw on one film distribution that it added sources on some names, but other names existed in the same source, that were not added. Do you think it could be done ? or too risky ? --Hsarrazin (talk) 20:17, 5 January 2016 (UTC)
@TomT0m: No idea about how much of the web is covered by microdata (I am sure there are some numbers somewhere). Althoguh not all microdata is useful for this case.
@Hsarrazin: I don't quite understand the first part of your question "you should use it multiple times on the same item".
In theory the same process could be used to add more statements but I don't feel that something like this should work in quite such an automated way.
@Addshore: sorry, english not my own language : I meant, as it only works on existing claims, you would have to use it several times on an item, after new claims have been added, or not ?
in fact, I do not understand how you intend to use it : systematically, run with a bot, or on a one item at a time basis ? and how can other people use it ?
and my question about importing statements was more on a semi-automated basis. Automatic adding of statements, even sourced, could result in a mess... I agree with you on that :) --Hsarrazin (talk) 21:51, 7 January 2016 (UTC)
Yes if more statements get added to an item the script would have to be run again in order to reference them! It is however easy enough to write a sparql query for example which can identify all statements of a given property on a instance of a given type that do not contain references, thus making identifying items that need to be rechecked easy. Right now the script is still being developed but I imagine once more complete I would simply do a one time run (and then top up runs after that). Although in theory the script will be there for anyone to run! ·addshore·talk to me!17:34, 8 January 2016 (UTC)
This looks very promising - thanks for working on it. :) In terms of what would be useful for reuse on Wikipedia, it would be good to have some way to whitelist sites that can be considered as reliable. For example, in the diffs there are lots of IMDB links, but IMDB isn't accepted as a source on the English Wikipedia because their data is crowd-sourced. I'm guessing this would need a whitelist, but maybe there's a better way of doing it? — Mr. Stradivarius♪ talk ♪06:05, 15 January 2016 (UTC)
One date with multiple places of publication would seem correct to me. Although date type properties seem to be poorly handled at present, because at any given point in time, the date and time differs depending on where you are in the world. Danrok (talk) 20:31, 12 January 2016 (UTC)
Maybe if it really is dubbed. Here only films with an audience marketed toward children are dubbed, and even such films are often displayed in two version, one dubbed and one with the original voices. -- Inn
@Innocent bystander: I agree. In that case I support (and also I used) the option mentioned by Danrok. But the new item should be created anyway, didn't it? It's a derivative work of an existing work. --Escudero (talk) 20:02, 13 January 2016 (UTC)
In Germany, both the dubbed and original versions are usually released at the same time, although there are not many cinemas that show the original version. --Srittau (talk) 12:54, 14 January 2016 (UTC)
Promotional use of Wikidata - companies requesting name change in Wikidata
In en:WP:OTRS there is ticket:2016011110025957 in which a public relations staff person for a company writes in asking for a name change in their Wikidata entry.
They changed their company name, and they changed the name of their English Wikipedia article, and now they want a change in the Wikidata entry. Their premise is that Wikidata decides the name in the Google Knowledge Graph, and they say if they get help changing their Wikidata name, then Google will present the name as they like it.
I am sharing this here because I expect this is a problem of the near future. I do not want every company in the world scrambling around wondering how to change their Wikidata name, and I do not know what to do for them. This is a little confusing in English Wikipedia and more risky to encourage anyone to control in Wikidata. I am presenting the issue here for consideration and documentation. Blue Rasberry (talk)20:04, 12 January 2016 (UTC)
in general (a) we can't prevent people from editing wikidata unless they really are behaving badly, and (b) if a name has actually changed, surely we do want the correct name in wikidata. There is a record of edits so if there's some sort of problem with a particular person or company we can get an admin to look at it and consider a ban. But generally this sort of edit sounds completely harmless. ArthurPSmith (talk) 20:55, 12 January 2016 (UTC)
They want General Dynamics Mission Systems (Q5531848) renamed to "General Dynamics Mission Systems", which is fine, but company renames are almost always unsourced on English Wikipedia and frequently a problem. This is a division of a larger company, General Dynamics (Q502940), and I am not even sure if it is notable.
I agree with Arthur that the edit itself seems harmless but there is some harm in the use of volunteer time to serve these kinds of requests. It might be worthwhile to note how much time is consumed when companies ask for PR assistance from Wikimedia volunteers on Wikidata. English Wikipedia has its own history with this, and mostly, neither volunteers nor companies are happy with the relationship. Wikidata could by default develop its own culture of permitting a similar tension, or it could do something different in response to corporate queries.
Right now in OTRS there are not templates for guiding people to use Wikidata. I have no idea how companies are even coming to believe that Wikidata influences their SEO. This particular case is not such a problem, but rather, I am sharing this as an example of a trend. Blue Rasberry (talk)15:34, 13 January 2016 (UTC)
Personally, I don't see a problem so far, and lack the imagination to see future problems. If a company requests something, then we either fulfill it if they are right (like in this case), or we don't. After all, we value data quality here and shouldn't refuse requests, just because they come from a source, some of us don't like. If this becomes a problem in the future, we can always look at it then. --Srittau (talk) 16:31, 13 January 2016 (UTC)
If a company merges with some other and changes its names, Wikipedia might re-write an article, but for Wikidata, new distinct items are generally more suitable. --- Jura15:55, 14 January 2016 (UTC)
Are you sure these are actually about the same topic? The latter appears to be about legal rights, while the former doesn't. --Yair rand (talk) 04:51, 13 January 2016 (UTC)
Indeed, the latter seems to be a "sub-topic" of the former. (I am not sure how to express this with the properties we currently have.) --Srittau (talk) 10:53, 13 January 2016 (UTC)
I'm sure. We have in German Wikipedia this Topic for that theme. 5323nhsetdr (talk) 17:49, 13 January 2016 (UTC)
In german language noone knows the meaning of LGBT. 5323nhsetdr (talk) 17:52, 13 January 2016 (UTC)
The German word for G="Gays" is "Schwuler(Single)/Schwule (two and more gays)" and so the acronym LGBT isn't correct in German language and could only tranlated with "LSBT", but that acronym isn't in use in German language. 5323nhsetdr (talk) 17:53, 13 January 2016 (UTC)
I was wrong above, "LGBT rights in Portugal" is not a sub-topic of "Homosexuality in Portugal", because ...
LGBT != Homosexuality (the latter is a sub-topic of the former), but ...
XYZ rights != XYZ (the former is a sub-topic of the latter)
The word LGBT is used in Germany, found it in at least those newspapers: Tagesspiegel, taz, Berliner Morgenpost
These are not the same topics and should not be merged. If those topics are confused on the German Wikipedia, they need to fixed there. --Srittau (talk) 12:53, 14 January 2016 (UTC)
thats not true. The acronym "LGBT" is not used in German language. And the word "Gay" isn't a German word. In German language we say "Schwuler" (Single)/"Schwule" (more than one person). So we can't take "LGBT" as an acronynm, because "G" stands in German language not for "schwule Männer". 92.76.122.1818:24, 14 January 2016 (UTC)
Do your research. Google any of the newspapers I mentioned above together with LGBT and count the number of hits. You might be surprised. --Srittau (talk) 22:36, 14 January 2016 (UTC)
The new options on Watchlist are fine for refining the list, BUT,
the "Show" button to update the list is now lost among numerous textes and cells, and difficult to find. It would be really great to make it a little bigger, and maybe colour-background and/or put it somewhere easier to find in the top of the page...
The new options on your watchlist are coming from Yair rand's script you installed this morning and aren't part of a general update. --Pasleim (talk) 07:57, 14 January 2016 (UTC)
your displaying of the watchlist is great, but the button to update the list is really not visible with the line of options under it... could you please make it a little bigger/or put colours in background ? Thanks a lot --Hsarrazin (talk) 18:51, 14 January 2016 (UTC)
@Hsarrazin: I added a divider between the options added by the script and the regular watchlist checkboxes. Does that help at all? The "Show" button doesn't do anything with regards to the script's filtering options, by the way. Those update the list as soon as they're check/unchecked. --Yair rand (talk) 23:14, 14 January 2016 (UTC)
that filtering is very useful, thanks... the line helps a little... a bigger button would be better, though :)
(Sorry if this is a known topic — I have found nothing in the archives.)
FYI, the babel-based customisation of languages doesn't seem to be compatible with global user pages. (Concretely, I'm using {{#babel:...}} on my “meta” user page but Wikidata shows me the default “four last visited” languages.)
Hi Hsarrazin. I may not have been clear enough. My user page itself works fine, the problem is with the display of items ; the language customisation doesn't (seem to) work when the #babel statement is at meta.wikimedia.org, I have the default behaviour instead. Tinm (talk) 21:46, 14 January 2016 (UTC)
Add references using quickstatements (quantity type)
I can't add references when I use quantity type sentence in Quickstatement. For example, "LAST P2049 +33 S143 Q22026414". Any tip?--KRLS (talk) 20:57, 14 January 2016 (UTC)
Migrating identifier properties to new identifier datatype
Hey folks :)
The new identifier datatype is getting ready. Once we have it we will put identifiers into a separate section on items. This is part of our effort to make items easier to scan in order to find the information you're looking for. Right now identifier properties are of datatype string. Please have a look and see if any properties we should migrate are missing or if any we want to migrate should not be migrated. You can find a list at User:Addshore/Identifiers. This list was made by Adam based on statements on the properties.
(For the technically inclined: The datatype will change for these properties but the value type will stay string.)
I added a lot of missing properties I found from re-running the query using wdt:P31/wdt:P279* rather than just wdt:P31 as was done originally. We might also want to check for all properties that use the word "ID" or "identifier in their name even if they haven't been properly given a P31 statement. ArthurPSmith (talk) 22:40, 11 January 2016 (UTC)
There were 479 id properties in the list when I first looked at it; I added 217 that were missed based on that expanded sparql query, for a total of 696. However, I did NOT add house number (P670) or P969 (P969) - those seem to be new items from User:Jura1. Not sure what the criteria were there, User:Jura1 added a lot of additional items not on my list. ArthurPSmith (talk) 19:59, 13 January 2016 (UTC)
Hmmmmm good question. Is there a particular reason those are URL and not string? We can configure certain statements to go into a special section based on datatype or property id. I'll have to check if we can also mix these two configs. --Lydia Pintscher (WMDE) (talk) 19:47, 12 January 2016 (UTC)
Ok. Thanks. So would it help if we make the replacement for the formatter URL more powerful maybe? I talked to Thiemo about mixing the configs is not possible right now. Will look into more options. --Lydia Pintscher (WMDE) (talk) 13:54, 15 January 2016 (UTC)
Based on the discussion here I'm now not sure what the purpose of doing this via a special datatype is. If the only purpose is for grouping things on wikidata display pages, how about just keying off statements on the property - for instance, any property that is an instance of Wikidata property for authority control (Q18614948) or one of its subclasses would be grouped separately. ArthurPSmith (talk) 19:10, 15 January 2016 (UTC)
why isn't that already standard? I'd forgotten I'd even turned on a gadget to do that. It seems obvious. And I don't see how a new datatype helps. ArthurPSmith (talk) 20:02, 15 January 2016 (UTC)
Grouping isn't the only reason. Another reason is so that we can emit the URI of the entity at the "other" organization in the JSON and RDF representations of the data. I'm not sure the APIs associated with those things should or do have the desire to be able to calculate those URIs. --Izno (talk) 23:20, 15 January 2016 (UTC)
Property "for"
Hi everyone. I've been working on modeling medical monitoring (alliteration!) in Wikidata, and I think having a property for "for" would be useful, if there's not one already. What I'm trying to do is have First aid measures : monitoring and then have things like "blood pressure", "pulse", or "hypertension" as qualifiers. I'm asking here rather than property proposals because I'm not sure if it exists or not. Thank you very much! Best, Emily Temple-Wood (NIOSH) (talk) 20:56, 15 January 2016 (UTC)
Not so easy to explain, it partly has to do how numbers are stored in the Wikidata-database. You can avoid this by entering a value like 123±0 instead of 123. In Windows the ± can be typed with alt+241 (on the numeric keypad). Yeah, it's a pain in the neck, and often does make no sense at all too. Edoderoo (talk) 12:56, 11 January 2016 (UTC)
It seems there is no phabricator request to simply remove the default behavior? This despite that users keep asking for this. --- Jura09:16, 13 January 2016 (UTC)
Can we do another one? The display issue is related, but not the same. It should simply state that manual entries of quantity should be reflected as entered ("12" => "12" not "12 upper bound 13, lower bound 11"). --- Jura08:15, 16 January 2016 (UTC)
I get the issue with Firefox too. It started about 10-12 days ago, I believe, as if the display of statements was triggered before its layout was complete. It also causes lots of trouble to gadget User:Soulkeeper/statementSort.js: quite often, and even more with items bearing lots of statements, the gadget properly sorts properties but the display shows mismatching values - as if their display had taken place before the reordering had completed. Typically, when this happens, there are no hyperlink displayed on any of the values that would normally display such a link. LaddΩchat ;)20:36, 15 January 2016 (UTC)
I confirm the problem in any browser when using User:Soulkeeper/statementSort.js. I often have to re-load the page 3, 4 or 5 times to have active link on IDs.
Yes. That is one of the goals. It'll be linked properly in the backend instead of the gadget. We might need to rely on the gadget for a bit longer after the rollout but this is on the short-term plan. --Lydia Pintscher (WMDE) (talk) 14:57, 16 January 2016 (UTC)
Tested 10 time with Edge, no problem, tested 10 time with Chrome canary (64 bit) always failed, tested 10 time with Chrome 32 bit always failed --ValterVB (talk) 22:22, 15 January 2016 (UTC)
This seems a lot like the problem I reported in phab:T115794. I've been having problems with anything that touches the statements section of the page ever since August and I have to refresh almost every page I load at least once to get it to load properly. - Nikki (talk) 10:46, 16 January 2016 (UTC)
DuplicateReferences
Does anyone know why the DuplicateReferences gadget does not work? I don't see any "copy" links next to references on items anymore. //Mippzon (talk) 16:08, 15 January 2016 (UTC)
A personal record property for sports was recently created (Q21142177). I've immediately hit a stumbling block in that I can't seem to find a relevant unit for times which have multiple parts, for example the marathon world record 2:02:57 (two hours, two minutes, 57 seconds) or the 800 m world record 1:40.91 (1 minute, forty seconds, 91 hundredths of a second). Is there a suitable property which can store these "timer"-style times? This will have ramifications for any timing racing sport. Sillyfolkboy (talk) 15:03, 16 January 2016 (UTC)
@Jura1: That doesn't really look workable (or user-friendly) for the above examples. Should we change the datatype of the property to something more open ended, such as monolingual text? It's quite a significant limitation to the property, otherwise. Sillyfolkboy (talk) 16:56, 16 January 2016 (UTC)
I would store a time like that in seconds (with possible any decimal parts also) and propose to the developers to display times like that better somehow (eg. as hours, minutes, seconds) automatically. But storing as seconds doesn't lose anything for now. There are other open display problems, for example any very large or very small quantity displays badly in wikidata right now. ArthurPSmith (talk) 17:15, 16 January 2016 (UTC)
I'm sure something is at Phab:T56318 to address this, but I can't really find it now. This probably means that no working solution will be available in the next 12 months. If you come up with a clear plan how it could eventually be converted into a time quantity, maybe a string-datatype property could work. --- Jura17:37, 16 January 2016 (UTC)
Can someone delete Epcot International Flower & Garden Festival (Q22118640) for me please? I made it after there already was a version of it (Epcot International Flower & Garden Festival (Q22058796), apparently. Elisfkc (talk) 16:17, 19 January 2016 (UTC)
As described in w:en:Restrictions on geographic data in China, Maps available in China are encoded into a different coordinate reference system known as GCJ-02 instead of the commonly used WGS-84 system, which affect basically all online maps available with mainland China map, including bing/google/here map by making coordinates on those maps drift for tens to hundreds meter, with the exception of OpenStreetMaps or some satellite map. For instance, by clicking the coordinate for Oriental Pearl Tower (Q223207), you can see the map view of basically all map show the coordinate is somewhere inside a river instead of the tower's actual location. Should GCJ-02-based coordinate data be denoted in wikidata for location within mainland China too? There are automatic conversion program source code exist online by reverse engineering. C933103 (talk) 00:09, 17 January 2016 (UTC)
Right now, and for the following 5 hours, I am deploying some database schema changes as requested by Wikidata developers to make sure Wikidata can continue growing as much as we want. This should not impact normal operation of the site (both reads and writes), so you can continue using it and editing it as usual. I want to acknowledge, however, that at certain times, during the maintenance, some things can be a bit slower than usual or some requests may sporadically fail. For example, from 12:11 to 12:15 UTC, recentchanges list freezed for some minutes before returning to its normal operations. This was detected and fixed immediately. I apologize for any inconvenience, hope your understanding for what it is a temporal state that will eventually improve the wiki and if you still detect some issues, please feel free to report them on phabricator.
PS: If there is any German speaker here, it would be great if someone could link/translate this into the German Village Pump so they are also aware, as this could affect that wiki, too.
PS2: Remember to continue editing as usual, as that should not affect the ability to edit or use the site in any way, only provide an explanation for certain errors. --JCrespo (WMF) (talk) 13:21, 20 January 2016 (UTC)
This batch of changes (the most problematic one) has finished. It created more problems than intended while finishing, with some extra errors between 18:14 and 18:19 UTC. But things should be back to normal again. Thank you for your patience. --JCrespo (WMF) (talk) 18:45, 20 January 2016 (UTC)
Proposal and demo script for watchlist and recent changes appearance
Changes to Wikidata items are shown in lists with autogenerated summaries like "(Added [en] label: Some label)", or "(Changed claim: Property: Value)", along with extraneous information like byte change. Not only does this leave out important information regarding the change (Changed the claim from what? How was it changed? Qualifiers?), it wastes space with the beginning part. A shorter and more informative way to list them would be to show a miniaturized version of the diff itself right there in the list: "Property: old valuenew value".
I submitted this as a proposal on Phabricator about a month ago, but it didn't get any response, so I threw together a demo script that reformats RecentChanges and watchlists, to show what I think would be a clearer, more concise way of listing changes: User:Yair rand/DiffLists.js. (Since the script was already grabbing the whole diff from the server, I also threw in a feature to filter changes, if you want to hide, say, changes to labels in languages other than English and Hebrew, or whatever.)
I want to emphasize that this script was thrown together rather quickly, and was not built for maintainability, or to be reasonable with regards to number of API requests. It is not bug-free, does not work with IE, and will probably break if the devs change some stuff around diffs at all. Its primary purpose is to push forward a proposal that I'm hoping will be properly built into Wikidata if it is agreed that the style/design is a good one. (If this does not go forward at all, and a year down the line the script breaks, I probably won't fix it.) That said, I hope you'll try it out, and tell me what you think of the design. --Yair rand (talk) 00:04, 11 January 2016 (UTC)
I love that you hacked this together. Thank you. And sorry for not getting to your ticket on pahbricator yet - christmas and traveling for the developer summit... :/ People: Please do test it and give feedback. If this works well for you we can totally look into making this work for everyone. --Lydia Pintscher (WMDE) (talk) 00:49, 11 January 2016 (UTC)
Visual example how "wikidata changes in extended watchlist" gadget works at ruwiki. You can see not only description, but also expand change to actual "what changed"
Ruwiki extended watchlist with wikidata changes
From my point of view, such way of changes display is better for complex changes, when multiple properties were changed in single edit (i.e. not via Wikidata UI). -- VlSergey(трёп)08:37, 11 January 2016 (UTC)
I think the inline display of my script works rather well for complex changes, and it doesn't leave out information or leave in any extraneous stuff like the empty diff columns or " / name" or the redundant autogenerated summary, or similar. I recommend trying it. --Yair rand (talk) 19:34, 11 January 2016 (UTC)
I really love the look of your version of the RC list. It is very readable, and can help catching accidental errors (like same date added on birth and death together because of sliding fingers ;)
Is it compatible with the recent tool that evaluates the probability of spam (coloured background) ?
Those are both interesting general designs for watchlists, but I'm more interested in the amount of content that can be cleanly displayed in the list for Wikidata changes. The default autogenerated edit summaries are both lacking too much important information and including too much unnecessary text. --Yair rand (talk) 23:00, 17 January 2016 (UTC)
Recent changes (https://en.wikipedia.org/wiki/Special:RecentChanges) provides the last 500 changes and changes made in the last 30 days. Is it possible to go back thousands of changes and months or years? Thank you very much in advance!
Yes. If you select the 500 and 30 options, you'll notice that those numbers appear in the URL. Simply change them to however many changes / however many days you want :) Ajraddatz (talk) 21:44, 16 January 2016 (UTC)
Afaik, the limit for recent changes is 5000 changes, as for the search results pages. I personally feel this as a bit limiting, too. Maybe result pages of more than 5000 entries would actually be getting a bit too large, but a pagination which would allow more than 5000 entries in total (e.g. show 500 entries by default, but offer a "Next 500" link, like it is done for e.g. the search results, and allow that for more than 10 times) would be very helpful for patrolling vandalism etc. --YMS (talk) 08:25, 18 January 2016 (UTC)
No, it is limited to only 30 days, number of changes isn't limited. At least, it should be like that (about number of changes). --Edgars2007 (talk) 09:01, 18 January 2016 (UTC)
@Voll: I am a journalist and need as much data as possible to conduct data journalism. @all: Thank you. Who can change the 30-day-limit? What must I do to get more data?
Tool: checkSitelinks was Disambiguator
About this my request, Matěj Suchánek has create a tool called checkSitelinks. Very useful to check if the sitelinks in an item are disambiguation or redirect. I use a lot to check if sitelink in a disambiguation page are all disambiguation page or is necessary split the item. To use it copy and paste the following line into your common.js.
You can acivate it just like any other user script (eg. my new one). Then, when you are editing a client article which has some old interwiki links, you can click on "Remove sitelinks" in the drop-down and the ones on Wikidata get removed. Matěj Suchánek (talk) 20:15, 17 January 2016 (UTC)
I know how to install a script :) What I meant was, is it automatically working ? is there a button ? a link ? where ?
from what you tell, I understand it removes sitelinks on wikipedia pages when those sitelinks exist in the linked element in wikidata, is that right ? is it working in wikidata or in wikipedia ? --Hsarrazin (talk) 21:45, 17 January 2016 (UTC)
You can try on Q204574 On sitelink section you must see [Check sitelinks!] Click on it and near to every sitelink you can see the icon of disambiguation. --ValterVB (talk) 11:56, 17 January 2016 (UTC)
is there also something to clean up labels of an item once a disambiguation page, that has been divided, when the only link is not a disambiguation anymore… I did it manually today on an item with dozens of wrong labels :(( --Hsarrazin (talk) 19:57, 17 January 2016 (UTC)
No, for Wikipedia articles that's manual labor as you need to check which text from 1 article needs to be transfered to the other and where to fit it in, in the destination article. Mbch331 (talk) 10:18, 18 January 2016 (UTC)
Wikidata weekly summary #192
Here's your quick overview of what has been happening around Wikidata over the last week.
Students worked on new datatype to capture mathematical expressions (phabricator:T67397)
Updated property suggester data to give you more up-to-date suggestions when adding new statements
Pages in the module namespace now also get interwiki links (phabricator:T123234)
Reduced number of resource loader modules to improve performance (phabricator:T123233)
Started experimenting with showing an image in the header area (phabricator:T119493)
Fixed a bug where the query text would be moved off the screen on the query service website (phabricator:T120196)
Worked on the remaining blocker for taking the in other projects sidebar out of beta. We need to link to the Commons category and not gallery for articles without frying the servers (phabricator:T94989)
The LCCN website has been down since Friday, does anyone know another way to look up the LCCN numbers? By time it is back up, after the government holiday, I will have forgotten which ones I still need to look up. If I build up too big of a list it will then seem like work to do it, if I wanted to do work, I would do the job I am getting paid for, instead if editing Wikidata, which I do to avoid work. --Richard Arthur Norton (1958- ) (talk) 18:47, 18 January 2016 (UTC)
I want to express that a test bench may interact with either a device under test (Q1206780) or a system under test (Q2376856). In the first case, physically interacts with (P129) is clearly appropriate. But in the latter case, which may include software, the physicality of the interaction is questionable at best. Does anyone know of an appropriate property to express that relationship? Does this argue for the creation of a "communicates with" property? Swpb (talk) 15:33, 15 January 2016 (UTC)
In a literal sense, yes, but software complicates the matter. I think it's a stretch to say that two applications or processes directly "physically interact" with each other when they merely access the same memory, file, or socket, often through multiple layers of abstraction. The interaction appears direct to the participating processes, but this abstracted interaction is not exactly "physical" as one would usually think of the word. By the logic that all communication is physical, the relation between conductor and orchestra, or listserv sender and recipient, is physical; it makes the property so broad as to be meaningless. Swpb (talk) 17:05, 19 January 2016 (UTC)
2016 WMF Strategy consultation
Please help translate to your language
Hello, all.
The Wikimedia Foundation (WMF) has launched a consultation to help create and prioritize WMF strategy beginning July 2016 and for the 12 to 24 months thereafter. This consultation will be open, on Meta, from 18 January to 26 February, after which the Foundation will also use these ideas to help inform its Annual Plan. (More on our timeline can be found on that Meta page.)
Apologies for English, where this is posted on a non-English project. We thought it was more important to get the consultation translated as much as possible, and good headway has been made there in some languages. There is still much to do, however! We created m:2016 Strategy/Translations to try to help coordinate what needs translation and what progress is being made. :)
If you have questions, please reach out to me on my talk page or on the strategy consultation's talk page or by email to mdennis@wikimedia.org.
The transliteration gadget does not work for Japanese Chinese character nor Traditional Chinese character, can that be fixed?
Is it intended for the gadget to transliterate all Chinese characters regardless of what language's Wikipedia the character were from according to mandarin Chinese pronunciation? For instance for Tokyo (Q1490), the Japanese language entry's transliteration is shown as "東jingdu" while the pronunciation is actually Tōkyōto. Of course there are other pronunciation and other transliteration exist for the entry in Japanese but none would yield something like 東jingdu.
Tonight, I noticed something new in the sitelinks section. After each sitelink is a [ref] or [ref sitelinkname] that appears.
When clicked it opens on the right a panel that displays the whole site article, including infobox and templates. see screen capture.
First, I thought it was an effect of a recent tool activation. But when I deactivated the last tools I added (checkSitelinks & DiffLists), it was still there.
not that I know… the Preview gadget looks like this and I still have it displayed when I click on the bubble next to sitelink. These links appeared today…
And the last modification I saw on Preview gadget was last autumn (13 nov.) unless it calls a submodule ? — this is mysterious :) --Hsarrazin (talk) 22:35, 18 January 2016 (UTC)
I thought of a new tool, not of an old one being improved, and sooooo much, by talentuous Magnus. Thanks to him, your working tool improve everyday :))
That's wonderful, because this new panel also solves the problem of seeing more than the article intro (Preview). Now, it would be soooo nice if it also could display wikisource pages (see Q19199343 for instance. @Magnus Manske: --Hsarrazin (talk) 22:07, 19 January 2016 (UTC)
Is there a page compiling all tools and helpers dedicated to referencing and Wikidata ? There is a fair number of efforts right now, it would seem good to keep a trace of them all (and it would be a great demonstration of wikidatas potential ...)
what i can think of
StrepHit
this very cool tool
the webpage metadata extraction tool that was on WWS a few weeks ago
Hoi, I noticed that "Gene Ontology" has taken possession of "congnition". I do not know and doubt that the psychological term cognition is the same. How do I deal with this? Thanks, GerardM (talk) 16:16, 19 January 2016 (UTC)
I fixed the description. Reviewing the ontology item at the link provided to AmiGO seems not to conflict with the common (English) use of the term. --Izno (talk) 17:18, 19 January 2016 (UTC)
This is one example, there may be more. The issue is to what extend do we trust imports like "gene ontology" who has its own vocabulary to define Wikidata. Thanks, GerardM (talk) 15:00, 20 January 2016 (UTC)
Wikidata is inclusive as it has to reflect major POVs, and gene ontology is a major one. So there is no problem to import them. The mapping to Wikidata concepts may be however different, as their concepts might be tighter that the definition of our items. I guess there is no real obvious answer and way of doing, as wikidata items might themselves in a lot of case have a lot of hidden conflicts and not so precise definitions, conflictual in articles (one big example is "chemical elements" : http://goldbook.iupac.org/C01022.html where enwiki is def 2, frwiki and some others are def 1, not solved yet) There is probably a lot of cleaning work to do currently hidden that might need to be enlightened. I don't think we have a better option than import the datas to make someone realize one way or another a mismatch in the item usage ... author TomT0m / talk page15:27, 20 January 2016 (UTC)
Symmetry
Hello everyone, I am not a regular contributor on Wikidata, so I hope the question has not been asked before. Will there be anything done in Wikidata in order to symmetrize the relations between elements? For instance, let's say that an element 1 has the property "son of" element 2, is it planned in the future that the element 2 gets then automatically a property "father of" element 1. As there is the property P1696, I suppose it is somehow planned, then when? Psemdel (talk) 21:57, 19 January 2016 (UTC)
I guess the short answer would be “Yes, but there is a lot of important things to do.” ;) As for the details ask someone else because I don't know what the status is on the matter. —Tinm (d) 00:12, 20 January 2016 (UTC)
The last sentence of paragraph four, see below, directs the reader to another Wiki site. The site does have songs for the album that can be listened to, but Big Railroad Blues is not one of them
"Modern listeners can hear Cannon's Jug Stompers recording of "Big Railroad Blues" on the compilation album The Music Never Stopped: Roots of the Grateful Dead." – The preceding unsigned comment was added by207.174.243.181 (talk • contribs). 07:26, 11 January 2016 (UTC)
Currently, related properties tool shows, that there are only 50,000 items with P31=Q5. Is it using very, vey old data; there is too big dataset or something else is wrong? If somebody could write SPARQL equivalent that doesn't get to timeout error, I would be thankful (I'm still learning SPARQL).
@Edgars2007: You need to put the LIMIT statement on the innermost query, to put a limit the number of statements looked at, otherwise the query will try to look at them all and won't complete. Here's a version of your query set to look at 4 million statements, which is about as many as can be asked for without hitting the query time out. It turns out that that corresponds to statements on 55,623 items -- i.e. more or less the number that Magnus's tool is limited to, although his WDQ is a bit faster than SPARQL.
Actually, it turns out to be faster and more efficient to limit the number of items looked up, just like Magnus does, rather than the number of statements.
Thank you very much. Learned some new things (putting limit somewhere in the condition, not simply at the end of query), that will be helpful in future. --Edgars2007 (talk) 08:42, 21 January 2016 (UTC)
Defining which parts of an external dataset are suitable for Wikidata
Hi all
I'm looking for some guidance on what the process is for deciding which parts of an external dataset are suitable for Wikidata. E.g importing the education data from UNESCO's UIS database, there are over 1500 indicators, which have values for each indicator going back to 1999. My assumption is that not all this data is suitable for Wikidata but its unclear to me the process that decides what is suitable and what isn't. At the moment please ignore questions of copyright of the data.
This is mostly solved through the WD:Property proposals process. If there is already a property for some data, then you can implicitly assume it to be data suitable for inclusion. If there is not a property, and you pick some data that needs a new property, then you should propose it and see if the rest of the community agrees. --Izno (talk) 16:27, 20 January 2016 (UTC)
That said, Wikidata handles large datasets poorly, which is why we use IDs to the degree that we do--so that other consumers can go and digest those datasets themselves. There's been some discussion of that topic on Phabricator and here at WD:PC but I don't know that I could identify where. --Izno (talk) 16:30, 20 January 2016 (UTC)
I don´t know the process for deciding which parts of an external dataset are suitable for Wikidata, but I remember Wikidata:WikiProject Freebase. In my view the point with UNESCO-data is, that I don´t know if there is someone using this data now or in the future. I have looked around and maybe data like "total population" will be used? Another problem is that there are lot of years for each dataset. We see this in country items already that the software is not written to work with many data. Therefore it would be better to store UNESCO-data not in the country items, but maybe in something like Q327213 (demographics of Germany). With Wikidata:Arbitrary access it is possible for a Wikipedia article to get data from any item. Another option is to use this data from outside Wikimedia. In this case more data could be transferred, I guess. Another point is the question if these data will be updated sometime. As far as I know, there is no discussion anywhere about that future problem. --Molarus19:12, 20 January 2016 (UTC)
@John Cummings: You won't find a lot of help here because wikidata is not the final user of the data. Uploading data to increase the size of Wikidata is not interesting. So the question is who can use that data. My proposition is to find a very small set of data (2-5 properties) and to work on the use of that data in Wikipedia articles by creating tables or graphics there. You will get more support by having a project especially if you are able to connect wikidata and wikipedia.
Define which data can be used in the article of the countries as tables and graphics. Discuss there to see if you can get some feedback from wikipedia contributors about use of UNESCO data.
Create 2-5 properties
Upload the corresponding data with a higher time interval (try to cover 20 years with time interval of 4 or 5 years to catch long term trend with a small amount of data)
Create the lua module which will display the data in tables/graphics in Wikipedia articles (and try to create a multilanguage module which can be applied on all wikipedias)
Apply it in wikipedia articles and try to add some comments and some references about analysis of these data (I hope UNESCO provides some reports which discuss the trends)
This is a very challenging goal but I think this can be a very nice example of how external databases can share data and how Wikidata can support Wikipedia. I can only encourage you to get in touch with user:Jérémy-Günther-Heinz Jähnick for the creation of lua module for tables in Wikipedia because he had a lot of experience. Snipre (talk) 20:27, 20 January 2016 (UTC)
From a short look I can propose you to start with
life expectancy at birth (most articles about countries have a demography section)
GDP per capita (most articles about countries have an economic section or a whole article about economy and I think we already have the property)
GNI per capita to create a graphic with GDP per capita (one graphic with 2 curves).
along these lines I've been wondering if we should try to get a "time series" datatype added to the basic datatypes wikidata supports - that way you can have one claim that includes the entire series of year, quantity values, rather than having to enter multiple claims with timestamp qualifiers. There's a general phabricator epic covering new datatypes but it may be a long time before developers can get to something like that. I think it would really be useful particularly for the economics-related properties that really don't make a lot of sense to present single values. ArthurPSmith (talk) 21:02, 20 January 2016 (UTC)
An approach that might work in this case could be to create one or several projects (e.g.) "Country Indicators" and try to build a mapping of fields for the data that could be imported and existing properties and/or newly required properties. This makes it easier to view the context of possible proposals and avoids that one ends up uploading PDFs to Commons instead. --- Jura06:34, 21 January 2016 (UTC)
Psychiatric treatment
Hoi, I am concentrating on psychiatry lately and one of my stumbling blocks is that there are methods and interventions. One "method" is FACT and ACT. They are an approach to care in the community. They use interventions and as such they differ. What is missing is the ontology stuff for this.
Another issue is that the implementation of FACT is done by the CCAF. How do I indicate the certifying organisation? Thanks, GerardM (talk) 09:45, 21 January 2016 (UTC)
Official name (P1448)
What is an official name? And can an official name be added without an reference? For example, if someone translates an official name of a state to another language, is it still an official name or is just a popular name of some kind? I believe the name must be verified by an authority to be official, and because of this an official name must have a reference to that authority. My gut feeling is that this is an example of a property that should not be used without references. Jeblad (talk) 15:09, 21 January 2016 (UTC)
The property description clearly excludes translations. Also the addition of a reference can be desirable but in many cases the value can be “general knowledge”. In the case of France for instance I am not sure the official name “République française” is defined anywhere, e.g. it seems the constitution just uses it. Interestingly though, the Code officiel géographique (Q2981593) defines the official names of many cities and villages, the usage names of which are sometimes different (but the property was probably just designed to describe basic information, not to handle this kind of cases). —Tinm (d) 16:06, 21 January 2016 (UTC)
Let's try something for ordering statements
Hey folks :)
I am on a quest to make Wikidata easier to understand at the moment. One thing we'll do to achieve that is separating out identifiers into their own section. The next thing I would like to do is tackle ordering of statements on an item. At the last Wikimania we've discussed this topic in the Wikidata meetup and it seemed to me there was strong agreement to go with a globally defined order. What I'd like to try therefore is to come up with an ordered list of properties by which all statements would then be sorted. This would have several major advantages: The biggest one for me is consistency across all of Wikidata's items. The other one is no edit clutter in the edit history of items for ordering related edits. I also realize that coming up with this one global list might be challenging. However I'd like us to at least give this a try. Are you on board? If so I'd start a page where we can start coming up with this list.
Ah good link! Thanks. Yeah the list needs to be editable for new statements and agreed upon ordering changes. I need to look into the performance aspect with the developers to see if we can actually make it work on-wiki or if we need to find some other mechanism. Either way the authority for the order would be the editors. --Lydia Pintscher (WMDE) (talk) 21:23, 10 January 2016 (UTC)
People are going to have very different priorities regarding what statements are important. Statementsort prioritizes social media accounts over residence and religion, and sorts Barack Obama (Q76) so that his various past employers and assorted occupations (along with his associated Wikimedia category) are above position held (P39)President of the United States (Q11696). Ideally, individuals would see different orders, with positions being determined by things like Wikiproject membership and past editing activity. Different items having different sorting orders depending on statements and statement values would also be helpful. But since all that would probably take more effort than it's worth...
This should probably have a full RfC, since hammering out a list, or even basic principles for writing it, is probably going to take a while. Once the ordering is actually live, it's going to be difficult to change things since people will quickly get used to looking for properties above/below/near other ones, and changing it later will be disorienting. One rule I'd like to see would be that properties about similar areas/topics get placed near each other, even if it means stretching "importance" a bit. I'd also suggest that Wikimedia internal stuff be placed near or at the bottom. --Yair rand (talk) 22:45, 10 January 2016 (UTC)
Yeah this is basically just about getting an ok to start such a list. We can do that as an RfC or something then if people are ok with at least trying the general idea. --Lydia Pintscher (WMDE) (talk) 00:44, 11 January 2016 (UTC)
Could you imagine to move our "classifying" properties instance of (P31) and subclass of (P279) to another sidebar section? This could be placed above the identifier datatype section. Another step to a cleaner UI whould be to move wikimedia related internal stuff (eg. topic's main category (P910)) to the sidebar. I think the main section should be reserved for the "real" facts about an item and ordering them is then easier. --Succu (talk) 20:10, 12 January 2016 (UTC)
Wouldn't it be possible to only establish and maintain a partial list ? That way we would have tree “sections”, ordered-unordered-identifiers ? I assume that it would be technically (almost) the same, and it would lessen the amount of maintenance and create a way out for disputed cases such as the ones discussed above. Regarding P31 and P279, in my opinion creating another sidebar would introduce unnecessary complexity. Giving them the first and second ranks on the property list would be simple, clear and sufficient. —Tinm (d) 16:16, 20 January 2016 (UTC)
Hmmmm I'd like to try us to get a full list. And if we fail to get to consensus on the list we can digg into the areas where people disagree to see how we can solve those. Because I am not sure which issues specifically we'll run into and solutions might look different for each of them. Sounds good? --Lydia Pintscher (WMDE) (talk) 11:14, 22 January 2016 (UTC)
I agree that a full list would in principle be better. ;) I was mentioning the partial list option as a (IMO good) fallback. —Tinm (d) 17:35, 22 January 2016 (UTC)
"Class" for a cycling race
Hi everybody. We start to display tables in the body of articles. I prepare the work for future tables, and I have a question to know if a good property exist or if I must ask its creation. We have cycling races, these races can be a part of the UCI Europe Tour 2016, UCI World Tour 2016, or UCI Oceania Tour 2016... Inside these continental circuits, there are different level : HC, 1 and 2. There is also the nomber 1 if it is a single day race or 2 if it is a stage race. So I will create items for races WT, 1.HC, 1.1, 1.2, 2.2, 2.1, 2.HC... but I don't know what property to use. In France, we called this a "classe", so it gives "class" in English. Jérémy-Günther-Heinz Jähnick (talk) 16:40, 20 January 2016 (UTC)
difficulty <HC> -- it seems that neither the property or the item exist, you probably can create items for “difficulty” (I'm not sure if we can reuse difficulty (Q16515105)) and “UCI difficulty”
Bot proposal to revert contribs made by a given user over a given period of time
Hello !
After a few months of (some) disability, I've finally brought to life a project I've had in mind for a long time.
Editing Wikidata is somewhat easy. Editing it fast isn't exactly hard either. For instance, Quick Statements can mow through a large amount of data and import it into Wikidata in no time. Same can be done with malfunctionning authorized bots. And of course, with unauthorized bots, whether the intent is good or malicious.
Reverting a large amount of edits is tedious, at best.
It is also not always possible to rollback, for at least two reasons :
someone else (bot or human) edited the page afterwards
legitimate edits by the same users were made immediately before the undesirable edits
As a way to improve data quality (or, more accurately, revert to a state where data was considered of quality), and with suggestions at various development phases by Harmonia Amanda, Ash Crow, and GZWDer, I created RollBot, which will mow that large amount of data back to how it was.
It works on a per-request basis (a human operator - myself - has to start it for every mass-revert request), and has two modes :
one that reverts what it can without overwriting legitimate edits, and reporting the rest
one that agressively reverts every page or entity to its former glory, reporting on overwritten legitimate edits
I'd gladly accept feature requests and bug reports (I am reasonably confident there shouldn't be any of those in the bot's current version, but bugs are pesky little creatures that exist when you don't expect them to), not to mention comments and feedback.
For simple mass-rollback there're Hoo man's smart rollback script. But this bot is useful if it can revert edits which are not the most recent ones (other users edit these items later), and other users' edits will not be reverted together.--GZWDer (talk) 19:02, 21 January 2016 (UTC)
For now, it can't. I'd much rather have a first working proof of concept and attempt to implement that non-trivial functionality in it, than having nothing and spending time developing what is really a secondary feature. Alphos (talk) 11:54, 22 January 2016 (UTC)
Hi, I'm new at Wikidata and I wanted a little of help. I have readed Help:About data and Help:Statements and I wanted to know what's the property of the province of a city and the Autonomous Community of a province. I found Property:P131 but it returns all administrative territorial entities where the city is located, for example, P131 of Valladolid, Q8356, returns "Provincia de Valladolid", "Castilla y León" and "Campiña del Pisuerga", but I wanted only the value "Provincia de Valladolid". How can I do? Thanks! --Elisardojm (talk) 13:51, 22 January 2016 (UTC)
No, these source statements are terrible and incorrect. (They're also constraint violations, but for some reason those still aren't showing up in the reports.) Jura1, I thought you had some plan to switch those to Q20651139? (I don't approve of that being used either, tbh, but it's better than this mess.) --Yair rand (talk) 23:37, 18 January 2016 (UTC)
There are various possible approaches and I tend to follow whatever is suggested at the moment. I think the bot request to change them may have failed as people couldn't agree on a practical solution. The last values you mentioned seem comparatively stable given earlier revert issues.
I think what we found is that: (1) people don't add missing statements manually. (2) users don't use reasonator to find missing inverse statements when doing a 10-page analysis of what the capital may be. (3) An indication on the statement is needed to ensure that people are aware why they are there and possibly fix both if they are incorrect. --- Jura
What are these statements supposed to express? Wikidata can hardly be a source for itself. This is utter nonsense and counter-productive, since it does not allow us to easily find unsourced statements. --Srittau (talk) 11:06, 19 January 2016 (UTC)
This is a list of all (item, property) pairs that include a stated in (P248) source referencing the property value. Please help clean that up:
PREFIXprov:<http://www.w3.org/ns/prov#>PREFIXp:<http://www.wikidata.org/prop/>PREFIXr:<http://www.wikidata.org/prop/reference/>SELECT?s?pWHERE{?srcStmtr:P248?src.# all source statements?stmtprov:wasDerivedFrom?srcStmt.# statements these source statements are used in?stmt?stmtProp?src.# statements where the object references the source?s?p?stmt.# subjects/predicates of these statements}LIMIT100
If someone looks for a template that makes good use of Wikidata, I don't think there is an easy way to find that yet. One or two badges could be help doing that. --- Jura14:36, 19 January 2016 (UTC)
On the items for the template, as other badges. Maybe we could have one for LUA templates, and another one for #property uses; both with some sort of Wikidata icon. --- Jura08:31, 23 January 2016 (UTC)
Wikisource data to feed StrepHit
Hi everyone,
I'm currently investigating reliable sources to use as input for StrepHit (see here).
It seems there are interesting data in Wikisource, like [6], but I was wondering whether it is fine to consume them.
My concern is that Wikisource is not a third-party (i.e., non-Wikimedia) source, but still contains lots of data coming from external resources.
What do you think?
no wikisource is not a third party. But it was created to provide sources on other wikiprojects.
You should be able to use all valuable biographic info (and other encyclopedias, or books, or reviews), because those should have a corresponding item, that you can simply use as "stated in".
in wikisource you will (almost) only find reliable and faithfull text-version of previously published books or journals - in fact, almost all wikisource items should be "edition", "periodical article" or "dictionary/biography article"… and have (or will have when it will be completed) a complete biographical notice ; those are probably the most reliable references you can use ;D
This is an edition from the year 1900. I would say it is a good source, because there where some proofreading against facsimile (see discussion page) and there is the possibility for the reader to look into Appleton's cyclopaedia itself. But there are fictitious and suspicious entries listed by user Bob Burkhardt. I don´t think this speaks against using Wikisource as a source, because there are lot of hoaxes out there. --Molarus00:53, 22 January 2016 (UTC)
@Hsarrazin: nice, I totally agree, both regarding Wikisource and the property to use, namely Property:P248 (while I would use Property:P854 for completely external references).
@Molarus: got it, I understand that extra care should be taken when digging into those data.
Some days ago I discussed with Magnus on his talk page whether he could provide a tool for complex references. He’s running a bot that can add such refs, but this functionality is not yet available as a user tool. The source code to do that is not very complex (and available in Magnus’ public code repository), but I think the idea needs some support before he’s going to implement it in a tool. So feel free show up in this topic to second my request to him … ;-) —MisterSynergy (talk) 07:46, 18 January 2016 (UTC)
WE-Frameworks gadgets remembers last used sources, so they can be used to quickly add duplicated sources. But it is not integrated with Wikidata interface. -- VlSergey(трёп)09:27, 18 January 2016 (UTC)
@Mippzon: also duplicating P2334 value in each reference to movie is redundant. Since one should specify SFDb for movie, you can omit it for all references in this movie entity. -- VlSergey(трёп)09:31, 18 January 2016 (UTC)
Thanks for your comment! I do not plan to develop scripts/modules/templates for data usage in Wikipedias, I just want to add properly referenced data here :-) Anyway, I think it should be stated clearly at Help:Sources how to reference data within databases. I fully agree with you that we should avoid redundancies, but I also see the comfort of having a click-able link to the source directly at the statement, not just somewhere within the item. If we just add stated in (P248) and publication date (P577)/retrieved (P813), one has to look around for the click-able link first… —MisterSynergy (talk) 11:10, 18 January 2016 (UTC)
MisterSynergy By adding reference to wikidata you should foreseen the fact that this information will be used in Wikipedia for citation purpose. Adding pieces of references is useless for Wikidata and for Wikipedia.
For now Help:Sources was defined based on the idea that no requirement is made for the item of databases. If you want to reduce the number of information for each statement you should define where the information will be stored. If I take the example of language of work or name (P407), if we don't provide it in the statements, we should specify that this information is in the item of the database. This is a small change but this should be specified. For me, it is a little difficult to define if a database can have a language.
Then for the title, this is a consequence of the current design of citations in Wikipedia: without title, citations in WP display the raw URL which is really ugly and prevent a first filtering of the references by the readers (if I have a reference titled "Organic acids" and another one "Acetylsalicylic acid" on page Aspirin, I think I can get more detailed information on the second reference).
I prefer to put too much information now than receiving later complaints that WD is not providing enough data to create clear citations. Snipre (talk) 12:55, 18 January 2016 (UTC)
To define the information we have to provide in WD, we have to look at the minimal requirements of the templates used for citation purpose and to the current habit in term of data added by contributors when filling these kind of templates. Snipre (talk) 13:33, 18 January 2016 (UTC)
Snipre, I understand the problem. In my opinion it does not really matter whether we talk about how references appear in the Wikidata web frontend, via API access or for Wikipedia modules. There’s the conflict of these two options where it’s not clear which one to choose:
(a) providing as much information as possible in each individual reference for easy access and processing (stated in, database ID, title, language (of database interface and/or content), and time of data retrieval), and
(b) providing only essential information (stated in, and time of data retrieval) to reduce redundancies. Other information could in principle be pulled from the database property (database language) or a combination of the database property and the item (database title and item label), or a statement of the item (database ID).
At the moment, none of the options appears optimal, I see advantages and disadvantages in both approaches. More important is that I don’t really know which option is generally preferable, and this is indeed something that needs to be clearly stated soon at Help:Sources. Wikipedia modules will be constructed along these guidelines, and external data access would possibly also rely on what’s stated there.
Finally I want to mention that this is one of the two reasons why I do not add the huge pile of references that is currently sleeping on my local hard disk: I do not know how the references should look like (type (a) or (b) from above, or even something else?). The second reason for me to wait is the lack of a suitable tool to process thousands of references in a single batch (or a couple of batches) without having to use Wikidata’s web frontend. —MisterSynergy (talk) 15:20, 18 January 2016 (UTC)
@Vlsergey: Oh, I thought each and every statement should have at least one source, even if it's the same source for multiple statements. Can someone point me to an items that does the sourcing thing in a more correct way? //Mippzon (talk) 17:02, 23 January 2016 (UTC)
@MisterSynergy: Thanks for clarifying! I guess the help section on sources can be improved regarding this as well, and I guess thats a bit of what has been discussed previously in this tread. //Mippzon (talk) 16:30, 24 January 2016 (UTC)
@MisterSynergy:@Vlsergey : By the way, can you try to explain why it's good to reduce the reference information on each statement? What is the advantage of doing so? Trying to learn something here :) //Mippzon (talk) 16:39, 24 January 2016 (UTC)
See my comment on 15:20, 18 January 2016 in this thread regarding the two models. I think it is too early for changes on Help:Sources, although the current situation is complicated. IMHO we would need at least an RfC to discuss any changes. —MisterSynergy (talk) 16:59, 24 January 2016 (UTC)
They seem to be different things. One a township in Camden County, the other a census-designated place in Franklin Township. (Whatever a "township" and a census-designated place is.) --Srittau (talk) 23:57, 23 January 2016 (UTC)
They're not the same. Some of the statements and sitelinks were on the wrong items, but I've fixed those now. - Nikki (talk) 17:10, 24 January 2016 (UTC)
Hope you all take the time to share what you have done with Magnus tools. Most recently I have been able to add references to over 1000 Wikidata statements thanks to quick_statements. --Jane023 (talk) 23:54, 24 January 2016 (UTC)
Jeah, we owe him and his tools a lot. By the way, I started writing code three years ago as I wanted to improve one of his js-tools. --Molarus11:42, 25 January 2016 (UTC)
Wikidata weekly summary #193
Here's your quick overview of what has been happening around Wikidata over the last week.
Events/Blogs/Press
Past: office hour on IRC with review of the past quarter, outlook on the next and discussion (log)
Right now: Magnus Manske Day! Thank you for all you are doing to make Wikidata truly shine! We owe you a ton.
Changes in how authentication in MediaWiki is handled are being rolled out (and more coming). [7] Ordinary users and gadgets / user scripts should not be affected but bots may need to be adjusted. pywikibot already supports [8] OAuth authentication, though bot operators need to set that up for their bot.
Investigated language support in MediaWiki and Wikibase to allow more languages in monolingual text and disallow some non-sense for label/descriptions/aliases (phabricator:T78006)
Fixed missing license info and broken links for Wikibase libraries on Special:Version
Working on improving scoring of search results on Wikidata, on Special:Search
Preparing to take the "in other projects" sidebar out of beta features, with the Wikimedia Commons link going to the Commons category (phabricator:T103102, phabricator:T94989)
Amir and Aaron are turning ORES into an extension to make it easier for you to spot bad edits \o/
Final touches on putting identifiers into their own section (phabricator:T117421)
Switching the in other languages box to be expanded by default for new people (phabricator:T92387)
More experimenting with adding an image to the header area (phabricator:T119493)
According to WD:L, a label «is the smallest unit of information that names an item». So, the label “municipality” refers to municipality (Q15284), a type of administrative territorial entity, whereas “municipality of Spain” refers to a specific type of municipality only seen in Spain. To my understanding, the “of Spain” is part of the definition itself of that particular item (municipality of Spain (Q2074737)) , so it must be on the label, which is currently the case in most languages. Some users of the Spanish Wikipedia decided to change all labels in Wikidata to just “municipality” in all cases in order to fit the design of a new infobox (instead of fixing the infobox in Wikipedia). This massive change does not make much sense to me, but i would like to hear other opinions. Also, how should we handle decisions affecting Wikidata that do not take into account Wikidata policies (or are discussed elsewhere)? Finally, one of the users has made an appeal in the Spanish Wikipedia for more users to show up at my talk page here, so I was wondering if we are ready to discuss turning Wikidata:Canvassing into a policy. Andreasmháblame / just talk to me02:35, 16 January 2016 (UTC)
I see on svwiki the problem with local discussions like: "I do not find any other article in any language who fits mine article, so it risks to stand alone without interwiki. But there are articles with a similair content. Let's decide [here] to link the article to that WD-item instead." To solve that, I have designed templates that adds interwiki from other items than the directly linked item. Maybe we can do the same thing here, help them code the infobox that it show "municipio" no matter what country affected. The problem, I'm afraid, is that we often have to hard-code this into the infobox, since I do not think we can rely on such statements as "subclass of municipio". -- Innocent bystander (talk) 08:16, 16 January 2016 (UTC)
Question is that what I understand of a label «is the smallest unit of information that names an item» means that Municipalty of Spain is not the smallest unit, the smallest unit is "Municipalty", imagine tomorrow they change municipio into a different and unique name, you won't use "of Spain" in that new label. I don't care if Spanish Wikipedia template needs to be changed or not, that's another problem. And sorry if it's not correct to ask other users to participate, in fact I proposed you to open the discussion with more users so we can arrive to a good decision.--Pertile (talk) 21:30, 16 January 2016 (UTC)
The name of the smallest unit of information that names the item is "municipio" (see for example: Spanish Constitution: [9]: El Estado se organiza territorialmente en municipios, en provincias y en las Comunidades Autónomas que se constituyan., The State is organised territorially into municipalities, provinces and Autonomous Communities that may be constituted). Or the Local Entities Register: [10]. "de España" is just a disambiguation formula used in es:wiki. My question is: does Wikidata have any sort dissambiguation policy? I see having aliases and the item being univocally determined by its "Q" it is not needed.--Asqueladd (talk) 20:14, 17 January 2016 (UTC)
That statement is only valid within Spain, but elsewhere that entity is known as “municipality of Spain”. According to WD:L, «The label is the most common name that the item would be known by», so “municipality of Spain” and not just “municipality” is the most widespread name around the world. Another example: for Americans, Barack Obama (Q76) is their president, but for the rest of the world he is the “President of the United States”. Accordingly, the label for President of the United States (Q11696) uses the most common name. So, the “of Spain” of “municipality of Spain” is not a disambiguation, but an intrinsic part of the concept: by using only “municipality” as a label, nothing would suggest that it is referring to a specific type of municipality, but rather to the general term municipality (Q15284). Andreasmháblame / just talk to me03:40, 19 January 2016 (UTC)
US Constitution's article II says "The executive Power shall be vested in a President of the United States of America", and POTUS (President of The United States) is often used, so President of the United States is a correct label for the US and the rest of the world. In my opinion for the rest of the world Spain has "municipalties" not "municipalties of Spain", Madrid is a municipalty or a city, you say that it is a city or a municipalty of Spain because you want to give a geographical context, not because it's part of the name "city" or "municipalty".--Pertile (talk) 23:33, 19 January 2016 (UTC)
municipality of Spain is a subclass of municipality. of Spain is not a disambiguation formula, but is a part of the full designation of the concept. Most people around the World, if they say or hear „Valldemossa is a municipality“ they do not think about legal definitions, and only if they know Valldemossa, they will think about Spain. They will not ask, what kind of municipality it is (Is it a municipality of Spain or maybe a municipal corporation in the United States?). Maybe they will ask, where it is, but that's another question. --Diwas (talk) 14:17, 23 January 2016 (UTC)
Al parecer estos aprendices de editores de wikidata no tienen nada para hacer, de ahí la ocurrencia de andar destruyendo cuanta plantilla exista en Wikipedia, y encima tienen la caradurez de lavarse las manos echándole la culpa a Wikipedia. Causar desórdenes para estos aprendices está justificado. Vaya tontería!
Por favor Indrus, cuando digas tales afirmaciones sin fundamentos al menos firma tus comentarios, para que se sepa quien es el autor. No creas que por estar en un mar de letras se pierde el origen, porque para eso está el historial. Si quieres echar agua sucia, bien puedes ir aquí y explayarte todo lo que quieras. --Shadowxfox (talk) 16:54, 26 January 2016 (UTC)
Extract statistics from wikipedia pages
Hi everyone,
I am new to the community and I wanted to ask if there is a way to extract from Wikidata's SPARQL endpoint statistic on an entity's page. For example, I would like to choose a specific entity (e.g. city of a country) or a group of entities (e.g. all cities in a region of a country) and obtain:
- number of words in that entity's wikipedia page
- number of pictures
- stats on edits and revisions
- stats on visits to that page
I know strat.grok.se and X!'s Tools for statistics on page visits and edits respectively, but I am not sure whether their information is in Wikidata's database (or perhaps dbpedia's?).
Has anyone tried to do something similar? Any suggestion?
Some of Magnus's tools allow a SPARQL query as input. But I haven't seen anything to use SPARQL output as a condition to extract a group of edit stats, or visit stats, or stats about the page make-up.
It may be that given a list of pages from SPARQL, you then have to write a script yourself to look up the results for each of those pages.
I think it's something that must eventually come, the ability to write queries that can combine information both from SPARQL and the standard Mediawiki SQL databases, all accessed in a single query. (Which of the pages in this class have been edited recently? Who by? What super-categories of this category can't be explained by Wikidata relationships? etc). But so far it's self-written scripts, I think, with the exception of Magnus's tools. Jheald (talk) 15:46, 24 January 2016 (UTC)
Thanks to both for the advice. I will try to understand more about it and perhaps come back with some results (and more likely further questions). --Federicopvs (talk) 16:17, 26 January 2016 (UTC)
Question
Does anyone know how to get P31 for each of 1000 shortest pages in some wikipedia, to make something like a report with page_title + P31 of the item of this page? --92.115.106.25015:49, 25 January 2016 (UTC)
If you can turn that list of pages into a Page Pile, then you can use Magnus's Taberancle tool to list the P31 values (put 31 into the box on the right).
Finding the 1000 shortest pages may be possible using the Quarry front-end to MediaWiki's SQL tables, which PagePile can then access; or some other method, and then drop the list directly into PagePile, or via PasteBin. Jheald (talk) 21:00, 25 January 2016 (UTC)
Capitalizing information retrieved from Wikidata in Other Projects
I am trying to make a info box heavily utilize Wikidata. However, some items are always fetched lower case. Here is an example of what I am referring to: https://en.wikipedia.org/wiki/User:Acebarry/sandbox (I am specifically referring to 'public domain' in the license list). Is there any way to force the proper case?
Lower case *is* the proper case. MediaWiki has the magic word {{ucfirst:}}, in Lua it is either lang:ucfirst( s ), or mw.ustring.upper(mw.ustring.sub(s, 1, 1)) .. mw.ustring.sub(s, 2). Matěj Suchánek (talk) 07:32, 26 January 2016 (UTC)
I've been asked by some of you to investigate how we can allow people to confirm that a dataset they own or parts of it is ok to be released under CC-0 on Wikidata. I discussed this with people at the developer summit and here is the outcome: We should adapt the waiver that Commons already has in place. We can then give it to the legal team for review to make sure it's ok. After that I can poke Quiddity to get the necessary OTRS queue created. Anyone up for taking a stab at modifying the Commons text? --Lydia Pintscher (WMDE) (talk) 11:30, 18 January 2016 (UTC)
So Wikidata:OTRS or do we need to create a page on Wikimedia (wikidata.wikimedia.org/wiki/Wikidata:OTRS) ? If the last solution is cosen, please indicate where we can adapt the text. Snipre (talk) 12:28, 18 January 2016 (UTC)
Yes, I think defining which data within a dataset or within several datasets on a website is important. An organisation may be willing to openly license a particular subset of data Wikidata needs but not the whole dataset/s. John Cummings (talk) 13:16, 19 January 2016 (UTC)
@Lydia Pintscher (WMDE): Something is not clear for me in that process: if X want to release a dataset to wikidata, he should send a copy of the data he wants to donate, isn't he ? This means we have to store that data set in a secure place until someone upload the data into wikidata. What's about the format ? What's about temporary storage ? Who will do the upload job ? Do we need to save an original copy of the data set in case of vandalism or wrong action (modification or deletion) on the uploaded data ? A description of the process should be provided. Snipre (talk) 16:06, 19 January 2016 (UTC)
Ok the reply I got so far is that we should go ahead with the drafting. The "donor" needs to specify what they grant us. Once we have something they're happy to look it over. --Lydia Pintscher (WMDE) (talk) 09:29, 27 January 2016 (UTC)
Future IdeaLab Campaigns results
Last December, I invited you to help determine future ideaLab campaigns by submitting and voting on different possible topics. I'm happy to announce the results of your participation, and encourage you to review them and our next steps for implementing those campaigns this year. Thank you to everyone who volunteered time to participate and submit ideas.
I think that maybe "documentary" is a genre also, but the fact that's it's a film of that genre can as well lead to consider that that the genre is mostly "documentary" and that the format is "film". Documentary film can as well considered as a class of films. There is also radio documentary (Q2125867). Ontologically, what's the difference beetween a genre and a class ? author TomT0m / talk page16:21, 21 January 2016 (UTC)
@Danrok: I do agree that "documentary" is a film genre (and it didn't take me long to find supporting references: 1, 2), but Queryzo thinks otherwise. @TomT0m: A genre could be a type of class limited to creative works, but it would be great to hear a more documented opinion on the matter. Anyway, it would be wonderful if we agree to delay performing massive edits that involved deleting claims until after opening a discussion in a relevant place, like here or a Wikiproject. Andreasmháblame / just talk to me17:09, 21 January 2016 (UTC)
Maybe it is german point of view, but as it is, I may have to translate the linked text from deWP:
No, that is different, I thought we’re talking about films. Dokumentarreihe is more for tv series. documentary films are a sub class of film, so there are not about a specific theme or motive (f.e. war, love or science fiction) but technically specific like short film or silent film. Is this clear at least? So is there a definition here in wikidata about that difference? Queryzo (talk) 23:55, 21 January 2016 (UTC)
From Russian point of view I suppose that "documentary film" is not a genre (so as "animation film") because it can have a genre by itself (drama, action, sport and so on). --Infovarius (talk) 05:34, 28 January 2016 (UTC)
Textual source "Колькасьць насельніцтва РБ у 2009 годзе"
Playing Addlingue I found a lot of claims sourced by "Колькасьць насельніцтва РБ у 2009 годзе", for example Pyeruch 1 (Q21282176). @Infovarius, Ivan A. Krestinin: could you explain to their author that there should be an item for this instead of a string ? and solve this by bot if possible ... author TomT0m / talk page17:09, 22 January 2016 (UTC)
Hi. It seems that User:Infovarius is again edit warring with me, last time we had such a conflict (about different item) in 2013. This time the item is skier (Q4270517). He is not willing to discuss with me, so I write here. Infovarius is saying that "skier, person who performs in skiing" is a subclass of "athletics competitor, sportsperson that competes in athletics (track and field, running, walking)". Please check what is correct... Thanks. --Stryn (talk) 17:03, 25 January 2016 (UTC)
@Stryn:, Sorry for delay, I'm a bit busy with my watchlist... I've overviewed definitions and must admit that you are right. Skiing is not light athletics because the last is strictly defined. I was misguided by the thought that everything not being heavy athletics is light athletics. --Infovarius (talk) 05:05, 28 January 2016 (UTC)
It seems that the value of these two properties are always the same. Is it a coincidence? Otherwise we only need one property.--GZWDer (talk) 19:09, 28 January 2016 (UTC)
Due to the forbidding interface of MycoBank I use it only when absolutely necessary. I do know that it does happen (if rarely) that MycoBank differs in opinion from IF. MycoBank also offers a wider range of information than IF. - Brya (talk) 04:33, 29 January 2016 (UTC)
Wikilinks and redirects
Hello !
As you may know, having two items point at the same page is frowned upon.
And as you may know too, it is not possible to add a wikilink that points to a page already in use in any item in Wikidata - the interface prevents it, except for a bug I've been told of that occurred some time ago.
However, it is possible to have an item link to a redirect pointing to another page with its own item, for instance by creating an item pointing to a page, and to turn that page after the fact into a redirect to another page.
While testing RollBot, I stumbled upon a few of them, and after I mentionned it in the #wikidata IRC channel, @Addshore: suggested a tool be built to find them.
For now, it's nothing but a collection of reports for such conflicts, with the wiki and date in the filename. The tool works in the background on the Tool Labs when I run it. Hopefully I'll be able to run it as a cron job, perhaps every month ; I'm currently getting to request permission to do so :
periodically
on bigger wikis, given that it did take an hour to complete on arwiki.
I however intend on puffing the interface up a little, possibly with a wiki selector and a date selector to make it easier to browse through…
Each report is either empty (indicating no conflict was found) or a TSV file with rows of 5 columns (in the current version of the tool) :
originating Wikidata item
wikilink for the current wiki
page and section the wikilink redirects to (should include the section as URL fragment if one exists, but my SQL-fu ain't what it used to be ; this may be dropped in later versions)
wikilink of the target of the redirection (pretty much the same thing, without the section/URLfragment)
which can be cut like this : 12181506Επάνω Αρχάνες ΗρακλείουΕπάνω_Αρχάνες#Επάνω Αρχάνες632418,
which means that "Q12181506 has the elwiki wikilink pointing to Επάνω Αρχάνες Ηρακλείου, which redirects to Επάνω_Αρχάνες, which is linked from Q632418".
I'd love any input you can give me, except asking for other wikis for the time being. The Tool Labs database replica already took a pounding with my heavy requests, I'd rather avoid pounding it again without full approval from our DBAdmin ^^
I'd also love it if you used the reports to fix these pesky wikilinks conflicts
Is a separate wikidata entry for a redirect always a bad thing? Wikidata is structured data, and wikidata items usually have a definite "type" (via instance of (P31)). In contrast a wikipedia page often covers several different things with differing types, and so would naturally correspond to different wikidata id's. One way to handle that is through alternate labels for the wikipedia page - the "redirect" mechanism. It seems to me quite useful to allow this. Although I've noticed that it's not possible to add new wikipedia links from redirects - or at least I haven't figured out how to do it. ArthurPSmith (talk) 20:27, 25 January 2016 (UTC)
@ArthurPSmith: Edit the redirect on the wiki so that (temporarily) it isn't a redirect any more; create the wikilink; then turn the article back into a redirect.
As for "using the reports to fix these pesky wikilinks conflicts" (ping @Alphos:) -- please don't. As Arthur says above, they're a useful workaround for the limitations in the software. Jheald (talk) 20:46, 25 January 2016 (UTC)
Uh, no. Really, don't. That's precisely one of the ways to cause trouble when editing the items afterwards. And it's more or less what wrcr (hey, it's catchy after all) detects now that the other way (failing check) is fixed. Alphos (talk) 23:45, 25 January 2016 (UTC)
I think it's a good thing to clean them out. They keep generating duplicate items, keeping people busy adding twice the same information to Wikidata .. as if there weren't already other things to fix. --- Jura20:56, 25 January 2016 (UTC)
In those cases where they are being used as workarounds, I have no problem with keeping them. That however is not always the case. Sometimes there are pages that have been merged, but should not have been, so this task is not just a matter of removing links. I have allready cleared out the iswikipedia links (it was quite an miniscule list anyway).--Snaevar (talk) 21:03, 25 January 2016 (UTC)
The "workaround" is creating at least as many problems as it is solving. Let me explain :
Wikidata items refer to specific entities. And wikilinks should link to pages relating to specific entities ; if you don't have an article on Zeus on your wiki, you shouldn't link that entity to the section in your article on the twelve olympian deities in Greek mythology : Wikidata is supposed to link entire pages to entities, not small sections of pages to entities. Internal section links are useful because they can refer to notions in a page, they don't have to refer to the entity defined by the page.
Interwikis could get incredibly tricky to read on pages that have multiple Wikidata wikilinks pointing back to, should they all be included in the listing.
And to be honest, some operations on Wikidata items get denied because the interface detects the conflict.
The situation is not really desirable.
I get that the concept of "1 Wikidata item ⇔ 1 entity ⇔ 1 wiki page" is uneasy to accept, and bruises some people's feelings ; but that constraint shouldn't be something to be "worked around", for multiple reasons :
Wiki pages that get renamed or deleted are noticed. Sections that get renamed or deleted, not so much - and a tiny change can make all the difference.
Assuming direct links to sections were a good thing, would you like the only interwiki for an entire list of entities in a small wiki to point to a page about only one of them on a bigger one ? Or possibly worse, multiple interwikis on that small wiki to the same big wiki pointing each to a different page ?
I know constraints are sometimes hard to work with. But they've been implemented for a good reason, and they really make the entire dataset we're working on valid and ultimately easier to work with
Just on one thing : Wikipedia articles do not necessarily correspond to specific entities. There are loads of legitimate examples. —Tinm (d) 00:52, 26 January 2016 (UTC)
Many WP language edition have rejected "1 topic : 1 page" as causing poor reading and maintenance experience. What I envision the user workflow as: They open a modded Steam, it searches the app id (80340) and find out wikidata item (Q2905613), then displays a link to the article (in this case a redirect marked for possible expansion). That last bit is broken for no good reason (your concerns are bollocks. Wrote a tool for the first one, the other one were able to do before Wikidata) and I can't read French. Dispenser (talk) 04:50, 26 January 2016 (UTC)
A list (in the wider sense, a "group of things") is an entity of its own too, an entity distinct from the entities it lists. See for instance Q34768.
A disambiguation page is also a list, distinct from the entities it disambiguates. See for instance Q35027.
How should any automated process behave when looking for a specific entity in a given language, if it gets a list of entities (which is in fact a different entity) instead ?
Wikidata isn't just an extension to Wikipedia, anymore than Commons is. It is a project of its own. I understand the need to have interwiki links to sections ; but Wikidata cannot currently serve that purpose, not with plain wikilinks which are links to pages about the entity they're linked from.
Are you asking a design question? Just stick a language selector next to the section [edit]. And use an icon to indicate partial matches. Dispenser (talk) 22:18, 26 January 2016 (UTC)
Maybe it helps, if we look at some of the files and the issue redirect generate for Wikidata.
Here a few samples from the dawiki file. Not that I think that it's a specific issue to dawiki, I'm sure we can fine similar problems in all files.
There are two items about two persons, but both link to the same article at dawiki
From a dawiki perspective, these redirects are probably fine. It's just that we can't have them on items at Wikidata. --- Jura07:12, 26 January 2016 (UTC)
Indeed. One entity must (in the RFC 2119 sense) only have one item. Redirects are absolutely fine per se. They just don't mix well with Wikidata… Alphos (talk) 10:13, 26 January 2016 (UTC)
Look, maybe a lot of the cases where there are two wikidata items effectively pointing to the same thing via a redirect are like the above examples where it's a duplication or error of some sort that needs to be fixed. Fine, I don't have a problem with that. The issue I have, which is maybe a conflict between the wikidata data model and how the wikipedia's want to work, is when different aspects of a real-world concern are best addressed by a single wikipedia page, but really need two or more wikidata items. An example here (where redirects are NOT currently being used) is Meta Platforms (Q380) vs Facebook (Q355). There is no enwiki entry for the first, but there is this redirect for the company that points to the page about both the service and the company. In fact the enwiki page there seems to be more about the company than the service. In any case, there are clearly two distinct entities recognized in wikidata here, but only one enwiki page covering both. If we can't point to the redirect page here then the pages on other wiki's about the company are left without suitable translation links. I don't think the current situation is optimal at all. I'm not sure what the best solution is, but linking to redirects in a case like this does seem a helpful approach. ArthurPSmith (talk) 14:55, 26 January 2016 (UTC)
Well, for the data in Wikidata, it's actually saner to really have two separate pages for Q380 and Q355. This also allows people who might be interested in an english article on Q380 when looking at the german article to notice there isn't one specifically about it in English. And yes, that's really a good thing !
This is not an attack on redirects themselves - individual wiki communities can decide what they want in that regard, and that's perfectly normal -, but a comment on wikilinks in Wikidata pointing to redirects to pages about different entities. Alphos (talk) 16:25, 26 January 2016 (UTC)
Alphos - please explain to me, why is no link to a perfectly acceptable English-language article about the company (and its service) a good thing? I fully accept that wikidata should have two separate items here, I'm fine with that. But there is just one enwiki article that both items are covered by. Why would it be wrong to link both items to that one article (in one case via a redirect)? ArthurPSmith (talk) 18:26, 26 January 2016 (UTC)
Because the article isn't specifically about one or the other, it's about both. Conceptually, it's closer to a list ([[List of things called Facebook]] if you want ) than it is to an article : it talks about two different things, Facebook (the service) and Facebook Inc. (the company). There's nothing wrong with a redirect, but a wikilink in Wikidata shouldn't be one. Alphos (talk) 23:58, 26 January 2016 (UTC)
No, in the interest of Wikidata the subject of a Wikipedia article can typically be decided by its first sentence and en:Facebook is clearly mainly about "an online social networking service", not a corporation, which is why it is a sitelink on Q355. Most Wikipedia articles are about one main entity, but may also handle other related ones, so your reasoning is impractical.
Wikipedias may for whatever reason decide that related entities should be described in one article, but unless the article is badly written you can typically make out which is the main subject. A typical example would be a museum where the Wikipedia article about the museum may also include a section on its purpose-built building. In that case, the article is linked to the Wikidata item about the museum, while the Wikidata item about the building gets linked to a redirect.
The idea that only entities with their own dedicated Wikipedia articles should have Wikipedia sitelinks seems odd to me. Wikidata should obviously include Wikipedia identifiers for as many entities as possible, even if not all of them have their own articles. Redirects are a good way to achieve that and to thereby structure Wikipedia content, linking it to the greater knowledge structure and creating a form of index. This has obvious advantages and is quite useful in many ways, not least in allowing users to find more information. Väsk (talk) 07:57, 27 January 2016 (UTC)
Except of course if they're looking for an article about the building in another language, with your example. Alphos (talk) 20:04, 27 January 2016 (UTC)
If another Wikipedia has an article about that particular building, then that article should be linked to the building-item on Wikidata. It's far better to have an interwiki link that points to an article with information about that building rather than no link at all. That information isn't worthless just because it doesn't have a dedicated article. Väsk (talk) 23:00, 27 January 2016 (UTC)
Except there won't be a link back from the article about both the museum and building to the article in the language about the building. Again, this is not an attack on the information, but a comment on the way it's linked with Wikidata. Alphos (talk) 19:39, 28 January 2016 (UTC)
The museum article in language 1 wouldn't have an interwiki link to the building article in language 2 if we follow your opinion either, so your objection there doesn't make sense. Again, the original (language 1) museum article is considered to be an article about the museum (P31:museum, not P31:Wikipedia collection of assorted stuff). It should therefore be linked to the Wikidata object about the museum and the article on language 2 about the museum. It shouldn't be linked to the Wikidata object about the building or the language 2 article about the building, because that is not its main subject.
The language 1 article about the building is non-existent, so you can't link directly to that, but adding a sitelink for a redirect to the museum article where the building is covered allows you to find that information via an interwiki link from the building article on language 2 as well as many other places that use data from the Wikidata item about that building.
To bring it back to a more general point: Sitelinks to Wikipedia redirects are a good way to bring structure to Wikipedia content. That is absolutely within the remit of Wikidata. It also makes Wikidata content more useful when we reuse it on Wikipedia, and other places. Removing these links would destroy the work people have put into adding them, while making Wikidata less useful in the process. Väsk (talk) 06:59, 29 January 2016 (UTC)
Is it possible to do a Cirrus search like intitle:/abc/ hier in Wikidata? I want to have a list with all Labels matching that pattern: /%s\([a-zA-Z0-9]*\)/Queryzo (talk) 16:07, 29 January 2016 (UTC)
Using the regular search (probably Cirrus), insource:/regex/ kinda works indeed, but after some testing I’m not sure whether it really finds all matches. I’m also not sure whether you can restrict searches to labels only. Special:ItemDisambiguation apparently does not support regular expressions. —MisterSynergy (talk) 17:28, 29 January 2016 (UTC)
SPARQL can't traverse all labels with 30 seconds limit even for counting ([14]). However Quarry works well for counting and limited queries ([15]). --Lockal (talk) 18:36, 29 January 2016 (UTC)
similar query could probably be done with incompatible dates for mather/child (a mother has to be alive to give birth…) ; father/child is probably a bit trickier, for obvious reasons ;)--Hsarrazin (talk) 18:55, 28 January 2016 (UTC)
The sitelinks should be removed, but I would expect the items to be merged since they're essentially duplicates. - Nikki (talk) 13:27, 29 January 2016 (UTC)
Sounds like a lot of maintenance for all that category stuff that is difficult to make use of. --- Jura08:07, 30 January 2016 (UTC)
help needed with classifying a bunch of edits
Hey everyone :)
Amir, Aaron and others are working hard to make it easier to detect vandalism on Wikidata with the help of ORES. It classifies edits to figure out for example which are the ones that are likely to be vandalism so we can concentrate on those when doing recent changes paroling and so on.
It is getting close to having this work in an extension we can deploy on Wikidata so it is all nicely integrated. One of the missing pieces is classifying a bunch more edits to train the classifier. If you can help with getting this done that'd be awesome. Here is the campaign: Wikidata:Edit labels and here you can see the progress.
labels.wmflabs.org is offline at the moment, causing slower loading of items (1.8 minutes vs 1.76 seconds). Please don't add the script to your global JS until this is fixed, please. Sjoerd de Bruin(talk)17:44, 29 January 2016 (UTC)
@Lydia Pintscher (WMDE):, I was not able to understand from the description what is going to happen if I install the gadget. I understand that someone would be collecting the info, but what is the info, how sensitive it is (for example, does it have access to my watchlist?), and who is going to have access to this info, remained unclear to me. Could you please clarify. Thanks.--Ymblanter (talk) 18:00, 29 January 2016 (UTC)
I'm interested in answer too... I installed the tool, and saw nothing. Is it only working on wp projects ? or can it be used from wikidata side ? --Hsarrazin (talk) 08:33, 30 January 2016 (UTC)
How to link article with exactly same content but in different script into wikidata?
For instance, Eastern Min (Q36455) is currently linked to the cdo:Mìng-dĕ̤ng-ngṳ̄ article on cdo wikipedia which is an article about the language written in latin script. On the other hand, there is a cdo:閩東語 article on the same wikipedia which have basically the same content as the previous Mìng-dĕ̤ng-ngṳ̄ article with the only different being this one is in hanzi script instead of latin script. As currently wikidata can only link one page instead of multiple per Wikipedia, the hanzi edition currently does not have any interlang link. Is there any way to resolve the situation? C933103 (talk) 13:23, 27 January 2016 (UTC)
Is said to be the same as (P460) really the right property for this? The English definition of the property suggests to me that it's only supposed to be used when two things are not the concept (and therefore should not be merged) but are sometimes considered to be equivalent. That also matches the way the interface behaves, you can't merge items which have a said to be the same as (P460) statement. When we want to mark duplicates though, we are saying they are the same concept and should be merged (but can't be right now because we can't have more than one sitelink to the same wiki), which seems to be the opposite of what said to be the same as (P460) means. - Nikki (talk) 12:50, 28 January 2016 (UTC)
@Nikki, Liuxinyu970226:: it's an extension of the current approach for cases where we find to items about the same (concept/person) and attempt to merge them, but can't given that one or several Wikipedia consider them as different. It might be cleaner to create one or two separate properties just for permanent duplicates. --- Jura12:15, 31 January 2016 (UTC)
No. A television film (Q506240) is a movie that's created for television and meant to be standalone. episode (Q1983062) isn't a movie. For some series watching an episode means you miss a part of the story because it's only part of a story. You need to watch multiple episodes to understand a story and often while 1 story unfolds another 1 has begun. Mbch331 (talk) 12:06, 30 January 2016 (UTC)
I would say also a television film (Q506240) can be part of a series. Tatort (Q689438) is a series with 973 episodes. Each episode is 90 minutes long and is standalone. Multiple broadcasters are producing episodes with different main characters. The only thing the movies have in common is the opening sequence and the broadcasting time. --Pasleim (talk) 14:08, 30 January 2016 (UTC)
I think if we begin with such exceptions it's difficut to tell where the episodes start to have no connection to each other. There are many series where it's no problem to just watch a single episode because they are standalone. But still there might be a loose connection to other episodes. --Jobu0101 (talk) 22:00, 30 January 2016 (UTC)
I think what's important is that they all do include the series property. The series items will the provide more details about the nature of the series. --- Jura10:33, 31 January 2016 (UTC)
In french, before recent times where the american model became the most known, there was a difference beetween a téléroman (Q2388283) where each episode follows the preceding one, and the "Série bouclée", where each episodes are independant, an other types such as the series analog of anthology (Q105420) (the only link is the subject, not the characters - for example the Twilight Zone or each seasons of True Detective), the Série-feuilleton, like a sitcom for example where episodes can usually be watched independantly but there is a progression in the story nevertheless, see fr:Série_télévisée#Typesauthor TomT0m / talk page10:41, 31 January 2016 (UTC)
One further question concerning episodes in Wikidata: How do you set the number of the current episode and the number of the season to an episode item? --Jobu0101 (talk) 22:00, 30 January 2016 (UTC)
Wrecking (Soviet crime) is misleadingly linked to Politisch-ideologische_Diversion
I'm new to the project and would like to fix this language link.
The article for "Wrecking (Soviet crime)" describes a different phenomenon than the German one linked in "Languages" section. Russian and English articles describe the same phenomenon, but the German is something related but different.
Hi, inverted winger game up in the gender round of Wikidata - The Game. I don't know how to assign that this position is not gender specific. Can someone please do this? Thanks! -- Zanimum (talk) 17:04, 30 January 2016 (UTC)
you may either go to the item and change its P31 (in fact there were 2 P31 statements on inverted winger (Q21707087)) or, if you don't want to leave the Game, click on the small Red flag in top right of your screen, which signals that there is something wrong on that item, here, where you (later) or someone else will correct the wrong claim :) --Hsarrazin (talk) 09:35, 31 January 2016 (UTC)
The second solution sounds great. I wouldn't inspired by (P941) in either case since it isn't a seperate text that is only based on or inspired by another text - the songs usually use the exact texts. DGtal (talk) 13:05, 31 January 2016 (UTC)
The header above the languages should mention that those links point to Wikipedia, according to the system messages and my personal experience. Sounds to me like someone who walks against a door, because he or she didn't read if they should pull or push. Sjoerd de Bruin(talk)14:54, 31 January 2016 (UTC)
It does, but those links aren't really about the same thing. Sometimes when one walks through a door one ends up in an adjacent building ;) --- Jura18:03, 31 January 2016 (UTC)
WikiProject Molecular biology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. so they can ask WikiProject Biology. 
is string datatype so unfortunately the order in sequence value might actually be "S1E1", which is imho a very bad idea since the all point of having an order is that it should be naturally orderable. This should be a real number so that we can use "order by" in sparql naturally. For "in beetween" numbers like 1b, we should use fractional values, for example 1.5 author TomT0m / talk page 20:10, 7 January 2016 (UTC)
Done --YMS (talk) 15:11, 4 January 2016 (UTC)
WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead..
– The preceding unsigned comment was added by TomT0m (talk • contribs) at 15:18, 10 January 2016 (UTC).
Done --Floscher (talk) 12:37, 11 January 2016 (UTC)
. @Infovarius, Ivan A. Krestinin: could you explain to their author that there should be an item for this instead of a string ? and solve this by bot if possible ... author TomT0m / talk page 17:09, 22 January 2016 (UTC)
