Talk:Wikidata

From Meta, a Wikimedia project coordination wiki
Jump to: navigation, search

Contents

Data added from wikis[edit]

Will Wikidata allow data to be added from/through individual Wikimedia wikis, either by allowing some JS tool to modify data through an accessible API, or (preferably) a way to edit data through the wikitext in the form of magic words/parser functions/lua functions? --Yair rand (talk) 23:47, 5 March 2012 (UTC)

Not through the wikitext, as that would defy the very purpose of wikidata: centralizing the storage and history of the data records. The central idea of wikidata is to avoid representing such data as wikitext, or to store it on individual language versions. Can you explain why you would want to do this?
We will allow editing via the API, and we do plan to provide JS widgets for inline editing of infoboxes and interlanguage-links. However, these features are not priority. Since we have a pretty tight schedule, it is possible that they will be pushed back. Eventually though, users should be able to just click an "edit" button in the infobox and change the data, without caring where it is stored. This should work at least if the user is logged in with a global account. I'm not yet sure how we will handle anon edits and locaql accounts in this context. -- Daniel Kinzler (WMDE) (talk) 06:47, 6 March 2012 (UTC)

WikiData logo candidate[edit]

Ambox important.svg
We're doing this all properly and for real now \o/. Check this for details please! --Lydia Pintscher (WMDE) (talk) 17:53, 9 June 2012 (UTC)
Information icon4.svg
Please cast your votes at Wikidata/Logo voting.
My candidate 1
My candidate 1 (icon)
My candidate 2

Hello, I'm IWorld and I have created a logo candidate for WikiData. The rectangle is the database and and wikis (the 10 lines) can access to them. In the top the logo uses the original Wikimedia Foundation colors. I hope you like this logo but I'm open for suggestions. --IWorld (talk) 17:47, 21 March 2012 (UTC)

Hi there. I appreciate your effort, but I think your work does not meed the basic criteria for a logo. I propose to use some kind of an icon for a "database" or "storage" and remake this into WMF collors.--Kozuch (talk) 19:49, 21 March 2012 (UTC)
So I can upload an icon of this logo but how I must remake that? I use the WMF colors in it. --IWorld (talk) 06:03, 22 March 2012 (UTC)
I have created a 2nd candidate and an icon for the 1st candidate. --IWorld (talk) 13:20, 22 March 2012 (UTC)

Why are they so black? Bináris tell me 07:50, 4 April 2012 (UTC)

Hi IWorld, thanks for your work. I'm somewhat bothered by the similarities between the logo you propose and original Google+'s logo. They have the same shape, pretty much the same colors, even the order and the position of the colors is the same. Best regards — Arkanosis 08:47, 4 April 2012 (UTC)


Hi all, here my proposals for the Wikidata project logo (sorry I'm not yet authorized to upload images here):

  • wikidata1.png: this logo is a database symbol with a bottom border including the Wikimedia colors (blue/red) and the sphere of the Wikimedia logo.
  • wikidata2.png - wikidata3.png: the 2 logos have the sphere incapsulated between two disks of a database symbole also with the Wikimedia colors.

Benbois (talk) 14:12, 4 April 2012 (UTC)

Do you have created these logos in SVG, too? --IWorld (talk) 15:13, 4 April 2012 (UTC)
Hi IWorld, they are made in SVG with Inkscape, here they are:
- 20120404_wikidata1.svg
- 20120404_wikidata2.svg
- 20120404_wikidata3.svg
Benbois (talk) 18:11, 4 April 2012 (UTC)
I like the second and third versions of the above post. I am not entirely sure about the significance of the colours in different directions. ·Add§hore· Talk/Cont 03:25, 5 April 2012 (UTC)
Hi, following the discussion in the mailing list, I made 4 new variations in a new theme: distributed data to the wikis
- 20120405_wikidata4.png - 20120405_wikidata4.svg
- 20120405_wikidata5.png - 20120405_wikidata5.svg
- 20120405_wikidata6.png - 20120405_wikidata6.svg
- 20120405_wikidata7.png - 20120405_wikidata7.svg
Benbois (talk) 09:21, 5 April 2012 (UTC)
And finally a last one, more in the Wikimedia logo spirit: 20120405_wikidata8.png - 20120405_wikidata8.svg
Benbois (talk) 14:52, 5 April 2012 (UTC)
I already posted them on the mailing list; here are mine: File:Wikidata logo idea.svg and the one with not-so-many vertices: File:Wikidata logo idea2.svg. --Joancreus (talk) 09:43, 5 April 2012 (UTC)

I propose this one. Mahadeva (talk) 12:42, 5 April 2012 (UTC)

Posted on the mailing list by Rebecca Jordan: http://www.sociableblog.com/wp-content/uploads/2012/04/wikidata-logo.png --Lydia Pintscher (WMDE) (talk) 19:11, 7 April 2012 (UTC)

I would prefer a solution that is similar to the Meta-Wiki-Logo. Because the Commonslogo also includes the same colors so there would be a better connection. All three wikis are global Wikis. so my personal favorite would be wikidata1.png. --Sk!d (talk) 15:33, 8 April 2012 (UTC)

Yes, but only Commons and Wikidata are repositories + Wikispecies. Meta-Wiki is rather backstage project. Przykuta (talk) 15:06, 12 April 2012 (UTC)

Hi, I've just finished a new (way) logo: 20120407_wikidata10.png - 20120408_wikidata10.svg Benbois (talk) 20:00, 8 April 2012 (UTC)

Category:Wikidata logo proposals[edit]

Finally, all logos are uploaded in the wiki: Category:Wikidata logo proposals. Benbois (talk) 08:45, 9 April 2012 (UTC)

Just came across this (image here) wikidata related logo from 2007. I'm guessing, based on the names on the slide, that Danny knows who made it. Assuming that it didn't end up getting used for another project it might be of interest, as a starting point if nothing else. /129.215.73.144 11:09, 9 April 2012 (UTC)

IMO "flower" logos and "diaphragm" logos would be better as proposals for new Wikimedia Commons logo, especially these: File:20120405 wikidata4 derivative 3.svg, File:20120405 wikidata4 derivative 4.1.svg :) Przykuta (talk) 18:58, 12 April 2012 (UTC)

I personally like the "logo8" proposal above, because it shows how different projects will source data from a single repository. It seems simple, intuitive, more universal than the "cylinder" symbol for a database, and uses the WMF colors pleasant and effectively:

File:20120405_wikidata8.svg

--Waldir (talk) 13:05, 13 April 2012 (UTC)

As a sidenote, it may be a good idea to take a look at this. --84.41.86.38 23:34, 13 April 2012 (UTC)
I made a new version less Chaos like, more organic File:20120414_wikidata12.svg Benbois (talk) 10:00, 14 April 2012 (UTC)
I applaud the change, but I have two remarks: 1) the gradient makes the symbol more complex and less likely to be reproduced exactly in all media; 2) the thickening of the arrows near their endings doesn't strike me as particularly pleasant. The curved shape is welcome in the connection to the central part, though. Alternatively, you could go ahead with the look the thumbnail above hints to, which is using diamond-shapes instead of the classical triangular arrow endings (if this is unclear let me know and I'll try to explain better what I mean). --Waldir (talk) 11:20, 17 April 2012 (UTC)
How about something like this? File:20120405 wikidata8 derivative.svg --84.41.86.38 11:45, 20 April 2012 (UTC)
  • File:20120405 wikidata8 derivative-2.png ? Maybe little green dot in the centre of logo could be better, than big green hexagon. Przykuta (talk) 12:18, 20 April 2012 (UTC)
  • I think the more we move away from the arrow symbols, the more unclear the "data sourcing" meaning becomes. This does surely look good, but unfortunately it loses that meaning, IMO. The diamond-ended protrusions as I suggested (I might do a demo myself to make it clear what I mean) could still resemble arrows, but getting too far from that disables that effect. --Waldir (talk) 14:26, 20 April 2012 (UTC)
To be honest, that symbol looks way too orderly and regular for a symbol of chaos, IMO. Something like this would make more sense. Otherwise, unless people already largely associate this "arrow star" formation to chaos (I had never heard of this association, for instance), I don't see how it would cause any serious intuition problems to this logo. --Waldir (talk) 11:25, 17 April 2012 (UTC)
Here are some variants on the proposal 8 that hopefully look less like the chaos symbol:
File:Wikidata benbois8 variants.svg
--Waldir (talk) 15:22, 20 April 2012 (UTC)

New try to replace the chaos logo in the middle of my logos, now with compass rose File:20120420_wikidata13.svg Benbois (talk) 13:59, 20 April 2012 (UTC)

Looks nice but my first intuition is that it represents geographical information... --Waldir (talk) 14:26, 20 April 2012 (UTC)
well well, it's not wrong :) then another way, more in technical view (under the cover) with a gear style File:20120421_wikidata14.svg Benbois (talk) 13:00, 21 April 2012 (UTC)

My two cents I am also in favor of something like 8, 12, 13 or this which all use arrows or compasses pointing outward--it implies that the goal of WikiData is to generate content to be used elsewhere and acts as a kind of inverse of the Commons logo, which is nice in my mind. Koavf (talk) 19:46, 21 April 2012 (UTC)

Again, a new version.. File:20120501_wikidata15.svg Benbois (talk) 07:18, 1 May 2012 (UTC)

Not bad. Unfortunately, like with the gears variant above, the data extraction/distribution sense doesn't come as intuitively as in the chaos-like #8. Actually, I came to grow fond of the 6th proposal from my 6-version file above (based on your original #8). What do you think of it? --Waldir (talk) 15:19, 2 May 2012 (UTC)
I agree, but since I know chaos logo, I'm not more happy with centric arrows. I finished a new version to be more extraction/distribution oriented. File:20120508_wikidata16.svg Benbois (talk) 11:39, 8 May 2012 (UTC)
flower :) Try with more "tech" lines. Przykuta (talk) 11:59, 8 May 2012 (UTC)
I happen to like it a lot :) Less round shapes would probably remove the flower shape, but the general concept seems great IMO: it does provide an immediate meaning of centralized distribution, and unlike the other logos, it also conveys a sense of tight integration. --Waldir (talk) 13:59, 8 May 2012 (UTC)
By the way, the association with a flower might not be a bad thing, considering the MediaWiki logo! File:MediaWiki.svg. In fact, looking at the concept from this angle, this one might also be an interesting option: File:Wikidata logo idea2.svg
Hmmm, right! :) Ok, flower with [[ ]] or {{ }}? {{ }} IMO cold be better (data for infobox, data including in other Wikimedia projects). Przykuta (talk) 10:22, 12 May 2012 (UTC)
I think, this one is still the best one, as I look on it and know what it means.
File:20120405_wikidata8.svg
All others are somehow too complicated. Kersti (talk) 10:01, 20 June 2012 (UTC)

Even if I have a favorite (which is File:20120508 wikidata16.svg) I added my own proposals. What do you think of them? @Lydia: Some of the logos shown here are not in the category! --TMg 18:36, 20 June 2012 (UTC)

I like them! re missing category: yeah :/ I am currently super busy. It would be awesome if someone could remind the artists of these logos to add the category. --Lydia Pintscher (WMDE) (talk) 00:23, 21 June 2012 (UTC)
I double-checked. All logos are in the category. Sorry for the confusion. --TMg 09:18, 21 June 2012 (UTC)
Thanks a lot for checking! --Lydia Pintscher (WMDE) (talk) 23:13, 21 June 2012 (UTC)
I took one of your logos and did some violence to it. What do you think? File:Wikidata Logo TMg Hexagon Brackets derivative.svg -----84.41.86.38 11:27, 24 June 2012 (UTC)
To be honest I prefer my version. But yours is also very nice. Good work. File:Wikidata Logo TMg Hexagon Brackets.svgFile:Wikidata Logo TMg Hexagon Brackets derivative.svg --TMg 18:24, 24 June 2012 (UTC)

Some might say that File:20120405_wikidata8.svg looks a little like the Star Wars Imperial emblem (though certainly in more cheerful colors...) -- AnonMoos (talk) 20:35, 24 June 2012 (UTC)

It's the Symbol of Chaos. Some more logo proposals look like this. I don't think it would be a good idea to use one of these. --TMg 15:12, 26 June 2012 (UTC)

Issue with licensing[edit]

I'd like to draw your attention to a requirement for exlusive rights on the new logo:

“(...) you acknowledge and agree that by submitting your proposed logo design that you grant to Wikimedia Foundation, Inc. an exclusive (...), perpetual, irrevocable, worldwide, fully paid-up license to use, reproduce, and exploit in any way without limitation all copyright, trademark, publicity, and any other intellectual property or other proprietary rights thereto. (...)" [1]

Any logo submitted under a fee license (such as GPL or CC-BY-SA) does not and will not meet this requirement as a free license can never be revoked. This applies to most posted proposals. --Spischot (talk) 04:52, 4 July 2012 (UTC)

Yes this problem is also something I stumbled upon but we've got confirmation from a lawyer that dual-licensing is fine apparently for this all. --Lydia Pintscher (WMDE) (talk) 04:59, 4 July 2012 (UTC)
Sounds intresting. So WMDE is going to hold an exlcusive(!) right and everybody else hold rights, too. I guess the chances to take legal action against non-authorised use later on could be quite disputible. Hopefully this will remain a theoretical case only. --Spischot (talk) 05:55, 4 July 2012 (UTC)

Existing wikidata project[edit]

Are people aware of the much older application called "wikidata" that is the basis of http://www.conceptwiki.org/ ? It lives at https://trac.nbic.nl/wikidata/ .

Yes and unfortunately it's not the only one. If you look at Wikidata/Archive you'll find a few more. It's unfortunate but I hope it will not be much of a problem. --Lydia Pintscher (WMDE) (talk) 20:52, 30 March 2012 (UTC)

References[edit]

Hello,

Is a central storage of references planned for Wikidata? By this I mean not references for the data stored on Wikidata itself (which, as I understand, is already planned). Currently, many resources are cited multiple times in the same wiki, and on different-language wikis, with citations varying in completeness and format. A central storage of references that can be used on any Wikipedia by just entering some identifier would be very useful; it could lead to more consistent citations and make sure that more parameters are filled out.

Thank you. InverseHypercube (talk) 01:00, 31 March 2012 (UTC)

We will collect references on a publication level. So say a book for example. We'll not do this for pages. --Lydia Pintscher (WMDE) (talk) 16:14, 2 April 2012 (UTC)
That's great news! Really looking forward to this. InverseHypercube (talk) 21:49, 3 April 2012 (UTC)
How about individual articles in journals? See for example this discussion for a justification. Boghog (talk) 19:40, 24 October 2012 (UTC)
Would be interested in this as well. For example, we now have many template pages like enwiki:Template:Cite pmid/12421101 in use on the English wikipedia, which would, in my opinion, be more appropriately stored on wikidata. --Wouterstomp (talk) 08:39, 30 May 2013 (UTC)

DBPedia / Sparql / RDF / Skos[edit]

I thought there is already a system data extraction of the Mediawiki based on the W3C Standard Protokoll in DBPedia. Starting some years ago with the 30million facts of Wikipedia and now having already 40 billion facts available and already linked through different Factbased Domain. Including the most benefitial for Media sites from BBC (all songs / songwriter / video and so on) and all medical study results also already linked together.

I assume Wikidata will only add additional data or create an updateprocess, but hopefully do not create an new standalone one!

There are some notes about the relationship in Wikidata/Notes/DBpedia and Wikidata. Hope that helps. --Lydia Pintscher (WMDE) (talk) 16:16, 2 April 2012 (UTC)

Is CC the right license for data?[edit]

The OpenStreetMap-Community realised after several years, that the CC license doesn't fit for data very well and therfore switches to a new license right now (the ODbL), which they elaborated together with some layers. Now all users at OSM have to accept or decline the new license - which means, the project will "loose" a certain amount of data from those users who are meanwhile inactive or declined.

Here are the main issues (according to the OSM FAQ) why they are changing the license:

  • The current license uses only copyright law. This clearly protects creative works such as written documents, pictures and photos. It does not clearly protect data, particularly in the US.
  • The current license is not written for data and databases. It is therefore very difficult to interpret. If someone uses your data in a map in a book and the map has several layers, what should be placed under CC-BY-SA? Just the OpenStreetMap layer and any enhancements? The whole map, including any unconnected layers and markers? The whole book?
  • It is difficult or impossible to ask questions about what can and cannot be done, as this means asking all the thousands of contributors individually to give their permission.
  • This means that “good guys” are stopped from using our data but the “bad guys” may be able to use it anyway.
  • It is difficult or impossible for folks to mix our data with data under other licenses.

IMO, data usually is for creating derived works from it, making mashups, mixing it with other data (which doesn't necessarily comply with CC). A CC license, that has been thought for creative works like books or images would be obstructive to the creativity of what can or cannot be done with this data. Maybe you should consult some legal advice or speek with the OSM-folks before we make the same mistake as OSM. --Alexrk2 (talk) 13:57, 1 April 2012 (UTC)

The WMF people should take note of this. Did you post this in a mailing list or something. Or at least link this post? --Locos epraix 18:54, 1 April 2012 (UTC)
We are aware that most CC licenses are not suitable for data. CC-0, though, is. And for the textual content (discussion pages, project pages) the most consistent license would be CC-BY-SA. --Denny Vrandečić (WMDE) (talk) 14:34, 2 April 2012 (UTC)
Yes, CC0/PD has also been discussed in the OSM-Community as another option. The community was ambivalent regarding PD vs Share-Alike. At the end PD has been discarded, - Here a detailed explanation - in short:
  • [..] a significant proportion of contributors are vehemently opposed to this and we would like to keep the project unified.
  • There is also a fear that large organisations could take the data and release a better product that ours. [..] if we go Public Domain, it would very difficult to reverse course.
And IMO, lots users wantend to keep the attribution condition for publicity reasons and to attract other users - OSM maps are widely used in newspapers, websites, television etc. --Alexrk2 (talk) 15:03, 2 April 2012 (UTC)
PS, I guess there is another problem specific to Wikidata: if Wikidata would be CC0, IMO it would not be possible to transfer content from Wikipedia articles into this database. I think of several lists (eg. lists of cultural heritages) along with short description texts for each item. --Alexrk2 (talk) 15:10, 2 April 2012 (UTC)
The latter is not the case. Copyright always protects a certain text. Since Wikidata will provide a form based interface, a list like the list of cultural heritages could not be simply copied into Wikidata from Wikipedia, but it needs to be reentered, fact by fact. And facts are not copyrightable anyway.
Regarding CC0 vs ODbL vs CC-BY-SA 4.0 (which seems to be able to deal with data, haven't looked into it yet): as per discussion on the wikidata-l mailinglist, the second and third licenses are mutually non-contributable. So an early decision for one of both would be effectively a decision for them forever. Whereas an early decision for CC0 can be easily changed by a relevant community, once it exists. So I am more convinced then ever that CC0 is a good license, for starters. --Denny Vrandečić (WMDE) (talk) 21:37, 3 April 2012 (UTC)
And facts are not copyrightable anyway ? It depends what you mean by copyrightable and under which law you are talking. Under US law, your statement is almost correct. Under European Union law (directive) databases are protected by sui generis right (en:EU Database Protection Directive#Sui generis right) prohibiting substential extraction of a database. In plain language, substantial extraction means "copying a lot of it". In plain language it means "copying is forbidden". So it must be clear if wikidata will adress the concerns of the European users or if it is a United-States law based project, like Wikimedia Commons.
  • I am against and I Oppose Oppose CC0. As CC0 requires users to renounce to some of their fundamental rights (moral rights, privacy rights), CC0's philosophy is similar with slavery. These rights are inalienable. By promoting the alienation of inalienable rights, the Creative Commons Foundation is on a very slippery slope, providing for the domination of human beings by trademarks, whose rights are upheld, while human beings are left with no tool to defend themselves. I believe that collaborative work in human societies must be done within the frameworks that require no more than minimal limitations of human rights. The maximalist approach of CC0 is evil. By the same token that slavery was abolished, the Creative Commons Foundation must abolish CC0. Teofilo (talk) 15:57, 12 April 2012 (UTC)
A recent resolution of the Wikimedia foundation (foundation:Resolution:Images of identifiable people, 29 May 2011), saying "We [...] value the right to privacy" is heading in the right direction. Teofilo (talk) 16:07, 16 April 2012 (UTC)
I would strongly support making Wikidata data available under CC0. Unlike prose, data represents work but does not contain an original or creative element of presentation, which is why it's not protected by law in the US. Moreover, it is reused in many contexts (such as aggregate statistics) which synthesize large amounts of different data, leading to an attribution burden that is far too excessive to be practical. CC0 is simply the only option that makes any sense in this context. The fears of an organisation displacing Wikidata are silly, since their commercial offering will not be integrated with WMF projects and so be essentially unable to compete. Dcoetzee (talk) 00:02, 10 June 2012 (UTC)
Read the above.. at least under European Union law databases are protected by copyright. CC0 won't be compatible with other projects like OpenStreetMap or Wikipedia. This means a CC0-WikiData won't be allowed to import content from Wikipedia, OpenStreetMap or any other share-alike data source. The worst case IMO would be if WikiData extracts content out of Wikipedia and release it as CC0. Under EU law this would be illegal. As a contributor in DE Wikipedia I would feel like being expropriated somehow. This is not acceptable! --Alexrk2 (talk) 15:32, 16 June 2012 (UTC)
I agree with Dcoetzee. @Alexrk2: I'd like you to consider the context. To put it to the extreme: I say here, on this talk page "Berlin. 3,501,872 people (source: statistik-berlin-brandenburg.de) –Krinkle". Now, next time someone looks up the population on statistik-berlin-brandenburg.de and uses it in a paper, they have to attribute me, because I said it first. No, of course not. Its the same for extracting data from Wikipedia. We've got Berlin and a certain number of people live in it but it's not in Wikidata yet. So lets create an entry for it, where we are presented with 2 fields: Value, and Source. So, lets go to the Wikipedia article on Berlin, find the population and the reference related to it. Then enter the value and the source. I'm pretty sure this requires no attribution whatsoever to the Wikipedia contributor who entered it there. For one because of the obvious lack of connection. There is no way to proof that I got it from Wikipedia. I could've found that source anywhere. After all, "Wikipedia is not a source". If what I just said wasn't true, then we all owe Google attribution, because there is a big chance that Wikipedia contributor, in turn, found that source by looking it up in Google. So, references and facts themselves are not copyrightable in that context. CC0/PD is the only logical license in this case. –Krinkletalk 13:09, 5 July 2012 (UTC)

Alexrk2, it is true that Wikidata under CC0 would not be allowed to import content from a Share-Alike data source. Wikidata does not plan to extract content out of Wikipedia at all. Wikidata will provide data that can be reused in the Wikipedias. And a CC0 source can be used by a Share-Alike project, be it either Wikipedia or OSM. But not the other way around. Do we agree on this understanding? --Denny Vrandečić (WMDE) (talk) 12:39, 4 July 2012 (UTC)

So, Wikipedia content (in this case, data) will just be thrown away and people will need to start from scratch? That doesn't seems good... Helder 12:42, 5 July 2012 (UTC)
I would agree with Krinkle in that if we extract (data + source) from Wikipedia, there should not be any need to cite Wikipedia.But the question is also: should we allow data-extraction from Wikipedia infoboxes, even when they are poorly sourced ? It is probably a community decision but it may impact the license. Of course: "source=wikipedia" should be given low rank, and even be removed as as soon as we have something better. But disallowing them completely probably means that it will take much longer before we can make use of infoboxes in Wikipedia, and make cause social problems if Wikipedians feel frustrated in the process.
Licenses like Open government license and "Licence ouverte" allow about anything, but the source is supposed to be stated, is it compatible with CC0 ?--Zolo (talk) 16:38, 5 July 2012 (UTC)
@Denny Vrandečić: I agree. But I thought, the aim (or one aim) of WikiData would be to draw all the data out of Wikipedia (infoboxes and such things).
@Krinkle: Database right is pretty complex and I'm far away from beeing an expert into that. Especially because there is a large grey area around the definitions of what a database is and what a substantial extraction is (in terms of quality and quantity). (A single value for the population of Berlin is of course not protected by copyright.) The problems you've mentioned are IMO a part of the problems I've discussed above regarding the OpenStreetMap project and their decision to switch from CC-BY-SA to ODbL. Becaus it's impossible to attribut millions of contributors for a single map. --Alexrk2 (talk) 22:37, 14 July 2012 (UTC)

The statement "at least under European Union law databases are protected by copyright" is incorrect. A database right (http://en.wikipedia.org/wiki/Database_rights) is not a copyright, it is an independent property right granted on non-copyrightable materials. Since only the database as a whole achieves protection (not individual contributions), it is unclear whether the Wikimedia Foundation could claim that right, or whether in the present situation of no-one owning the entire database, that right is actually unclaimable. However, I believe this is irrelevant. Creative Commons version 3 does include database rights for those localized licenses in legislations implementing database rights (e.g. http://creativecommons.org/licenses/by-sa/3.0/nl/legalcode http://creativecommons.org/licenses/by-sa/3.0/de/legalcode). They are included in the works, while at the same time, the license does waive any specific database rights to avoid database right proliferation. -- G.Hagedorn (talk) 17:05, 4 August 2012 (UTC)

  • I'm not familiar with this project, but this development sounds appalling. In the U.S. databases isn't copyrightable, period, and that's how I want things to stay. I reserve the right, as a reader, to suck up and use the source data for any table I want to make. (Not just theoretical - I've done this) It sounds like this OSM "share-alike" license would effectively claim that, by engraving the right legal incantations on a website, it's no longer legal for me as a reader to do that. And I object to that altogether, because the point of open source is to free the data, not to devise some clever new legal tactic designed to lock it up. I do not want to be signing a contract every time I view a web site or a page therein. Wnt (talk) 21:45, 27 August 2012 (UTC)

Data dictionary for Wikidata[edit]

I'm not sure if these items have been identified and if not, perhaps food for thought moving forward.

  • Any plans for a Wikidata dictionary so that there will be consistency in what folks will call their entry property/attributes and allowed values? A data dictionary can both serve as a reference as well as a place to discuss topics perhaps using Liquid Threads.
  • What portions of Wikidata be managed similar to what's already available in Semantic Mediawiki (SMW)? What would be different?
  • What are your thoughts regarding how type data properties (as used in Semantic Mediawiki) and what they will be called for a specific language will be maintained?
  • Would there be a #Redirect entry for similar property names?
  • What happens if someone renames the property name or allowed values?
  • Would the data be stored within a regular wiki page similar to how SMW stores its data using template syntax?
Hi!
  • Wikidata will have a separate namespace for defining properties. Properties have a unique, canonical ID as well as labels in different languages and a data type. Every page on Wikidata, incuding those descibing properties and those describing items on wikipedia, will have a talk page for discussion, as usual. Whether or not LiquidThreads will be used on these will be up to the community to decide.
  • Wikidata is expected to re-use some of SMW's code base, e.g. data type formatters and result printers. Our data model however differs sufficiently from mediawiki to warrant developing the storage layer from scratch. The main differences concern the multilingual nature of Wikidata entries, and the possibility to provide sources for every statement. Both founders of the SMW project are taking part in Wikidata, Denny Vrandecic as the project directory and Markus Krötsch as a consultant.
  • We will likely use the same data types as SMW.
  • Aliases for property names are an undecided issue. Different possibilities for providing access to individual properties using local property names are currently under discussion.
  • Renaming of properties is an unsolved problem. It my be that it will not be possible to change the canonical identifier of a property. It will however be possible to change the localized labels.
  • Data will be stored using the same mechanism currently used for wiki pages, but it will not be embedded in wikitext. MediaWiki will be handling it as a different page type, with specialized code for editing and displaying it.
HTH -- Daniel Kinzler (WMDE) (talk) 17:30, 3 April 2012 (UTC)

Thank you Daniel! One more question: From what I've seen so far, it looks like Wikidata will use as much of Mediawiki as a base. Is the plan to write the Wikidata code as plugin/extensions or will it be a fork of Mediawiki? --LionAlex 20:43, 3 April 2012 (UTC)

It will be implemented as a set of extensions (at least one for the repository and one to be used on client wikis). We will modify the mainline core a bit to allow extensions to actually do what we want to do, see Wikidata/Notes/ContentHandler. -- Daniel Kinzler (WMDE) (talk) 08:21, 4 April 2012 (UTC)

Try[edit]

to explain things more simple. Noone wants to know how data types will be defined and how they will be generated. My simple question as a user is: ur trying to rebuild commons, but not with pics, but text modules, right? And authors or machines then can build articles out of that modules, is that right, or am I wrong informed?--Angel54 5 (talk) 02:09, 8 April 2012 (UTC)

Replace text modules with data and you are right, yes :) The FAQ is probably right for you. If something is unclear in there please let me know. --Lydia Pintscher (WMDE) (talk) 10:03, 8 April 2012 (UTC)
What then 'data' means? Could u explain that somewhat further? Data is everything, from a pixel of a pic, to a bit and a byte (building a letter). Btw. I know [[2]] him.--Angel54 5 (talk) 23:19, 8 April 2012 (UTC)
Data in this case means things like the birthday of a person or the length of a river. --Lydia Pintscher (WMDE) (talk) 17:10, 9 April 2012 (UTC)
Here's a question - does data include things like videos, images or zips? Or is wikidata for data that can be expressed as a non-encoded string? (of course, with the right encoding videos, images, and zips can all be stored as text, which is why I make the distinction). Jztinfinity (talk) 00:10, 10 April 2012 (UTC)
No. Wikimedia Commons is the place for those. --Lydia Pintscher (WMDE) (talk) 08:08, 10 April 2012 (UTC)
Could I reframe that last question slightly, then? What if data relates directly, technically, to an image or set of images, or to a sound file or set of sound files? Is there any scope for including images and/or sound files directly on a data page? Tony (talk) 05:30, 12 April 2012 (UTC)
As far as I know no. That's what Commons is for and duplicating that would be a bad idea in my opinion. Linking to things there is of course another thing. --Lydia Pintscher (WMDE) (talk) 11:11, 12 April 2012 (UTC)
The media would just be references pointed to for data about their contents, if I understand correctly. So, the photograph may be on Wikimedia Commons (eg, old photo of street with cute building); Wikidata would point to that as evidence of a fact (eg, a building used to be on the street); Wikipedia would point to Wikidata as reference (eg, discussing the history of the street, "a cute building used to be on the street"). It would be interesting to be able to embed a referenced item, via pointer from Wikidata, while editing Wikipedia. --Roket (talk) 20:47, 18 April 2012 (UTC)

Man against machine[edit]

Collecting the datas is IMO the way to "free all knowledge". Still there is the risk, that we'll lose the much more important "real" knowledge as we push out the people who have it. If our texts are starting to look like that (no offence againt that user!) we will loose people who have knowledge but are not able (or willing) to learn "IT language". The loss could be higher than the gains. ... I dont want to speak against Wikidata, just please put very high on the agenda the usability to "normal" non-Wikifreak people ...Sicherlich Post 16:51, 10 April 2012 (UTC)

Yes that is high on our list. In some cases it will probably even become easier for people to contribute who are not as you say Wikifreaks. For example the plan is to have forms for the data in infoboxes. It'll be much easier for people to handle these than large templates hopefully. --Lydia Pintscher (WMDE) (talk) 20:08, 10 April 2012 (UTC)
Will it be possible to integrate those forms into Wikipedia proper, or will users have to come to the Wikidata platform in order to edit data there? I already adressed this issue in the German office hour, and I think it would make a great difference for users to be able to edit data on their native platform, viz. the localised Wikipedias they are used to. Compared to Wikimedia Commons, it is a bigger step to come to Commons to make a contribution than doing it right away on, say, de.wp.--Aschmidt (talk) 10:06, 11 April 2012 (UTC)
I agree that this would be good. We still have to figure out how we can do this. --Lydia Pintscher (WMDE) (talk) 10:14, 11 April 2012 (UTC)

Diversity in data[edit]

I would like to hand on a hint to this piece in The Atlantic that comes in via wikimedia-l:

...should the population of Israel include occupied and contested territories?...
If we start to rely on a singular source for our truths, it will undoubtedly in most cases make most articles more accurate and current. But it also means that in contested cases, we will likely see an even more vivid reinforcement of existing core/periphery inequalities of knowledge production. A disagreement over facts or data would no longer be confined to a specific article and language, but would most likely have to be conducted in English to an unfamiliar community of editors.
Without social structures and technologies specifically dedicated to maintaining diversity in Wikidata, the almost certain outcome is that the truths and worldviews of the dominant cultures in the Wikipedia community will win out...

There should be a way to mark disputed data, and there should be a way to also add alternative data. Communities should be able to choose between the different versions.--Aschmidt (talk) 12:52, 11 April 2012 (UTC)

It will be possible to have different statements about the same thing. It will also be possible to chose from them. Denny explains this best in his comment on the article in The Atlantic. How and if we are going to mark disputed data still has to be figured out. --Lydia Pintscher (WMDE) (talk) 13:36, 11 April 2012 (UTC)
I think the best way is that Wikidata shouldn't rate data at all but let the wikipedia communities decide.--Trockennasenaffe (talk) 13:45, 11 April 2012 (UTC)
Well, data is disputed if there is more than one version of it. This might be enough.
Thanks to Denny for elaborating in his comment on the Atlantic article. I would like to dwell on his final remarks there: I am sorry for this long answer, but since I consider your concerns would be very valid if Wikidata was done in a more naive way, and since I understand that many people will think that Wikidata is being developed in such a naive way, I took the liberty to expand more on our current thinking of how Wikidata could work, and some of the design decisions in building Wikidata. – I think the discussions about Wikidata would probably be improved if we knew more about the design of Wikidata. Such points, viz., that there will be alternative values for describing the same thing, really matter, as you can see from the Atlantic article and from the recent discussion in Wikipedia Kurier. This is why I think you should publish more about the proposed design of Wikidata in order to develop it together with the community. At present I think the impression prevails that we will be presented a Wikidata sometime we will have to live with, no matter if we like it or not. However, we can only provide input on development if we know what you are up to implement in detail.--Aschmidt (talk) 14:28, 11 April 2012 (UTC)
You are of course right. We'll publish more over the next days/weeks. We've only had a short week of work so far and most of it was spent on getting started. The data model and api draft are published already and those are the things that matter the most right now and need feedback. Someone will work on setting up a demo system over the next days. This should improve things a lot as well. I am also working on a blog post to address the concerns of the discussions on The Atlantic and Kurier. --Lydia Pintscher (WMDE) (talk) 11:20, 12 April 2012 (UTC)

Report errors in data[edit]

It might not look like the most urgent issue, but I am mentionning it here in case it has structural implications.

Several major institutions have partnered with Commons to upload content. In order to remain in syhcn with the institution database, editors were sometimes discouraged from directly editing the description. Separate error reporting pages were created instead (eg Commons:Commons:Bundesarchiv/Error reports).

Similar situtations are likely to arise with Wikidata. In some areas, even authoritative databases can contain a substantial number of errors. Would it be imaginable to provide a way to easily find all data from a particular source that are disputed on Wikidata. That would make collaboration with external institutions easier. - I would imagine it would entail more carefully designed talk pages, or maybe supplement to entity talk page with a comment section for each attribute.--Zolo (talk) 06:38, 12 April 2012 (UTC)

That's a very good point. Thanks for bringing it up. Do you have other examples beyond the one you linked to already? --Lydia Pintscher (WMDE) (talk) 11:23, 12 April 2012 (UTC)
This could be in contradiction with the non negotiable requirement No. 7 at Wikidata/Notes/Requirements : "Wikidata will not be about the truth, but about statements and their references. These can be contradictory.". If you find an error, you should input another alternative data alonside the error (if this non negotiable requirement is to be taken seriously). I am opening a new topic below about deleting data. Teofilo (talk) 12:13, 12 April 2012 (UTC)
@Lydia: There are other similar /error report pages, most notably Commons:Commons:National Archives and Records Administration/Error reporting, but I am not quite sure that all partner institutions took them into account...
@Teofilo. The idea is not to change the info on Wikidata directly: if the the source provides wrong info, we keep it. But it would be great if we could report errors to the source's authors to see if it they can be corrected. Done in a clean way, I think it would be a win-win solution. --Zolo (talk) 12:40, 12 April 2012 (UTC)
if the the source provides wrong info, we keep it. I think there should be a deletion policy to allow deleting a wrong data, even if the source provider wants to keep it (or fails to answer E-mails). I think the user community should have a deletion policy allowing to delete contents, even if the source of the data (like the uploader on Wikimedia Commons) wants to keep it. Teofilo (talk) 13:12, 12 April 2012 (UTC)
I think there is a misunderstanding here. There is a difference between correcting clearly wrong data and having different sources having conflicting data which both are to some extend right. The former should obviously be corrected and this will be possible. The requirement you mention is about things like the size of a disputed territory. Each side has a different "right" number for the size of the territory. They will both be able to be recorded in Wikidata with their respective sources. --Lydia Pintscher (WMDE) (talk) 13:56, 12 April 2012 (UTC)
Correcting a wrong data without providing a source for the data you provide as a correction would be akin to original research. What I mean is that there should be a way to remove one datum provided by an otherwise trusted source, while keeping all the other data provided by that source. For example if a source providing a set of geolocations for cities happens to locate London in the middle of the Atlantic Ocean, we should either provide the geolocation for London from an other source, or leave that field blank. If data are provided by sources in bulk, I am not sure if it is possible to remove one datum without removing the whole bulk. Teofilo (talk) 14:13, 12 April 2012 (UTC)
This should be possible. --Lydia Pintscher (WMDE) (talk) 14:15, 12 April 2012 (UTC)

Deletion policy[edit]

Will it be possible to delete data ? Will there be a process similar to Commons:Commons:Deletion to delete data ? Will it be possible to delete a data on the basis that it is wrong ? As I said above, non negotiable requirement No. 7 at Wikidata/Notes/Requirements : "Wikidata will not be about the truth, but about statements and their references. These can be contradictory." can imply that deleting a data on the basis that it is wrong is not possible. This would mean that deletion is possible only in the case when the data is infringing the project's license. On the other hand, I am not sure if keeping deliberately wrong data is a good ethics. This takes me to the question below about whether wikidata will be a wiki. Teofilo (talk) 12:19, 12 April 2012 (UTC)

Yes it will be possible. Also see my answer above. --Lydia Pintscher (WMDE) (talk) 13:56, 12 April 2012 (UTC)
In some (not all) cases, provably wrong data may be useful, provided they are not too far off the mark. Suppose that for consistency reasons, en.wikipedia decides to use IMF data for all its economic statistics about a country. Then suppose that from some country, the IMF directly uses the GDP figure provided by the national statistical agency. Three months later, and without warning the IMF, the national agency makes some correction to the GDP (it happens fairly often). Even if the IMF fails to update its database for some time, we should keep the IMF data on Wikidata for en.wikipedia.
That said it is certainly a worthwile effort to try to mitigate as much as possible the effect of bad data on Wikipedia articles. Wikidata/Data model#Ranks_of_Statements says that incorrect data can be given a lower rank in the data hierarchy. As far as I can see nothing is foreseen to associate an individual statement with a comment about its accuracy. If it does not raise to many technical difficulties, it would certainly be nice thing to have. We could then imagine bots tracking accuracy disputes and leaving a note on Wikipedia pages that use the incriminated data.:) --Zolo (talk) 14:28, 12 April 2012 (UTC)

A deletion policy could adress promotional contents (whether data covering a specific brand are allowed or not). Teofilo (talk) 13:47, 12 April 2012 (UTC)

Yes something like that will have to be figured out in the community when it becomes relevant. --Lydia Pintscher (WMDE) (talk) 13:56, 12 April 2012 (UTC)

Will wikidata be a wiki ?[edit]

A wiki software supposes that contents are editable. Editable means a possibility and a freedom to delete contents (subject to community consensus). As Wikidata/Notes/Requirements non negotiable requirement No. 7 says Wikidata will not be about the truth, but about statements and their references. These can be contradictory., it remains to be seen whether editing is allowed. It is possible to understand non negociable requirement No. 7 as meaning that editing by removing information is not allowed at all. Hence wikidata is not a wiki. I wish some clarification on whether editing will be possible. Teofilo (talk) 12:31, 12 April 2012 (UTC)

Yes it will be possible. Also see my answer above. --Lydia Pintscher (WMDE) (talk) 13:56, 12 April 2012 (UTC)
It is very important, because it seems that the ultimate goal is to interconnect Wikidata with Wikipedia. If Wikidata is not flexible enough, then Wikipedia will in turn lose flexibility and will no longer be the encyclopedia "anyone can edit". Only big data providers (institutions) will be allowed to edit some selected data within Wikipedia, and correcting mistakes will take the time it takes with big institutions: always longer than "one click". Teofilo (talk) 14:27, 12 April 2012 (UTC)
Yes :) That'll not be the case thankfully. --Lydia Pintscher (WMDE) (talk) 14:29, 12 April 2012 (UTC)

Data sorting[edit]

On File:Wikidata UI mockup.png, population figure "3,490,445" is located higher than "3,500,000". "Calling code" is located higher than "Area". Will it be possible to switch "3,490,445" and "3,500,000", or to switch "calling code" and "area" so that the data are shown in a different order ? Will there be a sorting function similar to en:Help:Table#Sorting ? Teofilo (talk) 13:29, 12 April 2012 (UTC)

The mock-up is just a mock-up. We're starting to figure out the UI. Things like this will likely be done. And if the initial development team doesn't get to it it should be trivial to add it later or in parallel. --Lydia Pintscher (WMDE) (talk) 13:56, 12 April 2012 (UTC)

Writing text about variable data[edit]

Suppose someone adds to an article from ABC.wikipedia a statement of the form

"Between the years of ... and ... the ... (increased OR decreased) from ... to ...",

and for that purpose it fills the "..." with data from WikiData (whatever syntax it is used for this).

At the time the paragraph was written, the resulting text could look like this:

"Between the years of 1800 and 2000 the mortality rates decreased from 50% to 10%".

The data from 1800 could have been informed by someone which was reading a random article of XYZ.wikipedia, and wanted to provide this missing information. Unfortunately, it could happen that this was based on a reference which contained wrong information (even if it is just a typo).

The time goes on and this wrong information is added to more articles, on different Wikipedias, until a reader of PQR.wikipedia notices a phrase such as

"On 1800 the mortality rate was 50%..."

is incorrect in some other article (the reader may be an specialist in this area), and then uses some super cool interface which allows the interaction with the WikiData database without leaving PQRwiki, and fix their article to show the correct value of 5%.

Great! This article is now more correct.

On the other side of the world, people now read an article of ABC.wikipedia which says

"Between the years of 1800 and 2000 the mortality rates decreased from 5% to 10%".

although it should say "increased".

I tried not to focus on specific details, and in this (simplistic?) example the only thing which would need to be updated is "one word", but in other situations, or depending on the language, the article may need to be changed substantially (e.g. if another part of the same article refers to the data implicitly).

Some points I think it would be good to discuss inspired by this hypothetical example are:

  1. When all the data of an article is in the article itself, someone changing a value of 50% to 5% will likely update the wording of any phrase which refers to it as well, but this do not happens if people are not seeing all the other contexts in which a piece of information is been used;
  2. In case they did have a tool to see the "global usage" of a data, wouldn't the need for updating zillions of articles (possibly in languages the reader doesn't know) make them afraid of fixing the wrong data because they don't want to/can not fix all articles which use it? (so, instead of having at least one article with the correct information we would instead have zero!);

How to deal with this kind of situation? Ideas? Comments? Helder 20:02, 12 April 2012 (UTC)

I think this is one of the cases where using data from Wikidata is probably not a good idea in this combination. That's why we're focusing on infoboxes for the beginning. --Lydia Pintscher (WMDE) (talk) 13:16, 16 April 2012 (UTC)

Permanent (or permanently broken?) links[edit]

Once people start changing Wikipedia articles to use WikiData information instead of putting the data in the articles themselves, how useful will it be to cite old versions of articles? Will the OLD TEXT be displayed together with the NEW DATA?

That would make things even worse than what they are now: old version of articles should transclude the version of the templates used at the time of that revision, but:

Helder 20:55, 12 April 2012 (UTC)

Yes unfortunately this will likely be the case. I don't think this makes the situation worse than it is right now though. As you pointed out this is already happening with templates. --Lydia Pintscher (WMDE) (talk) 07:38, 13 April 2012 (UTC)
I think it does: once WikiData is in use, the data of e.g. the infobox templates (which currently are on the articles, and as such doesn't suffer this problem) will be outside of the wiki, and this will cause an increase in the amount of information of the articles which we won't be able to cite by means of the permanent link feature. Compare:
Currently In the future
{{infobox
| field1 = permanent value1
| field2 = permanent value2
}}
{{infobox
| field1 = (non permament) value1 from Wikidata
| field2 = (non permament) value2 from Wikidata
}}
Helder 17:37, 14 April 2012 (UTC)
But infoboxes included as templates also give you the current data in this case. Or am I missing something here? --Lydia Pintscher (WMDE) (talk) 12:51, 16 April 2012 (UTC)
What do you by infoboxes included as templates? Do you have any example?
I was considering something like this: the current wiki code of w:Brazil has (as of 17 April 2012) the text "|GDP_nominal=$2.517 trillion" (this in in the main namespace, not in the template). On the source of an old version from 22 February 2012, the value was "|GDP_nominal=$2.422 trillion" and someone can safely link to that old version, because the old value will be shown as expected. This wouldn't happen if the code e.g. was like |GDP_nominal={{<template storing the GDP value>}}" or, with Wikidata, |GDP_nominal={{#data:<some identifier related to GDP nominal>}}" (I'm just inventing a syntax here, since I don't know how exactly the central data will be inserted in the articles).
Currently, I think most of the data which is displayed inside of a infobox is defined in the the article itself (in Main namespace) and not transcluded to it from somewhere else (e.g. Template namespace). So, most of the factual data is not affected by this limitation of MW. On the other hand, with Wikidata, there will be a lot of information defined outside of the article, and then the permanent link will be less useful because the factual data would not be permanent anymore. Helder 12:28, 17 April 2012 (UTC)
I suppose wikidata will (naturally) have a versioning system, so for changes in values stored in the wikidata system, you would want a wikidata diff, not a wikipedia diff. Now, having the data separate from the article might make it difficult to see the article+data in sync as they were in a previous state. It might be an interesting problem to consider, but I'm sure a solution can be thought off. For instance, the wikidata extension (if that's how it'll be integrated, I don't know) could insert a hook at the oldid display action (pardon any technical inexactnesses) to fetch the data from that oldid's timestamp from the wikdata API. But this is just an idea from the top of my head. It doesn't sound like a terribly hairy problem to solve, but of course it should be kept in mind and definitely solved in some way in the wikidata system. --Waldir (talk) 22:24, 17 April 2012 (UTC)
By the way, you make an excellent point: mediawiki itself should have a mechanism to see a page as it was at a certain point, including the templates' rendering at that time (or that any transcluded material, such as other pages, images, etc). If this were solved, wikidata would just need to hook into that mechanism. Sounds like a worthy feature to request in bugzilla :) --Waldir (talk) 22:28, 17 April 2012 (UTC)
I agree with that, but the bug(s) (linked in my first comment) were closed as WONTFIX... =( Helder 17:27, 18 April 2012 (UTC)
I'm so sorry; My comments above were completely redundant. I should have read the discussion above more carefully. Indeed this looks like a quite hairy problem to solve across mediawiki (templates, magic words, image versions, category memberships, red links, etc), so perhaps the pragmatic approach in this case would be to actually implement this on Wikidata as I described above (or some way similar to that). --Waldir (talk) 14:35, 19 April 2012 (UTC)

What does this mean for Wikispecies?[edit]

Redundant? Does this project make Wikispecies redundant? Will it be maintained? What if users only want to be able to download species-related data into their own database? Will other individualized databases be created (e.g. only baseball players or only modal jazz albums)? Koavf (talk) 18:08, 13 April 2012 (UTC)

Wikidata itself will do nothing to Wikispecies. The Wikispecies community can decide whether to use Wikidata, and if so, how and to what extend they want to use it. I don't know enough about Wikispecies to be able to say how much of it's functionality could be covered using Wikidata itself, and which things would need a dedicated, separate Wiki... but whatever the answer may be, I expect that there will always be things to do for a community committed to biological taxonomies.
As to other "specialized" databases: Wikidata itself will for the foreseeable future not provide any filtered views like this, but I expect an ecosystem of tools and services to grow around Wikidata that will allow such extractions. -- Daniel Kinzler (WMDE) (talk) 13:01, 16 April 2012 (UTC)
As I see it, the immediate change will be: wikispecies and wikipedia can stop using two different redundant sets of data -- both the wikipedia infoboxes and the wikispecies infoboxes can draw data from Wikidata, and communities who care about species on both wikis will be able to share a single workspace for updating data. SJ talk  06:25, 21 April 2012 (UTC)

Research data[edit]

Will this project be looking at hosting anonymized research data from clinical trials? I am associated with a group that is looking into the possibility of creating a database for this sort of content. The issue is that human data is frequently not released so that results of trials cannot be verified. Thus we do not know for example if the medications used for this last influenza epidemic actually worked as the company that did the trial will not share them. We need an easy way to share this sort of content. But is would be more for researchers rather than for other WMF projects.--Doc James (talk · contribs · email) 13:08, 14 April 2012 (UTC)

What specifically will be in Wikidata will be up to the community to decide. Should they decide not to allow this kind of information it'll be possible to run your own instance of the software running Wikidata where you can do this. --Lydia Pintscher (WMDE) (talk) 13:08, 16 April 2012 (UTC)

Language issues?[edit]

  • Will the Wikidata wiki be multilingual? How will it allow non-English speakers to participate on its organization?
  • Once someone build a tool to allow people from non-English wikis to edit the central data without leaving their own wiki (possibly integrated with the VisualEditor, using some API), people from different wikis may happen to be editing the same piece of data without being aware one of each other (nor of the existence of the central wiki)
    • How could people from different wikis interact when a central piece of data is being contested (like in an edit war)?
    • Will there be a way for new users from local wikis to get notified about new changes (e.g. reversion, or vandalism) on central data they fixed from their wikis?
    • Will there be any way to compensate the possible extra difficulty in getting consensus about a given information?
    • If not, wouldn't this reduce the number of contributions to fixing wrong data?

Helder 14:38, 14 April 2012 (UTC)

1. The UI will be translated, yes.
2. I imagine this to work similarly to Wikimedia Commons works right now unless we find better ways.
3. We're going to try to make this happen.
4. Do you have suggestions?
5. Possibly some. We'll have to work on making this as painless as possible. But we expect a large increase as well due to Wikidata. --Lydia Pintscher (WMDE) (talk) 12:56, 16 April 2012 (UTC)

Values like names?[edit]

How about values? I have seen examples of data-values, which are language dependent. OK Mozart is Mozart (I think), but how about people from Russia or Asia? Their names are often transcribed in western european languages. So the name of a person can be language dependant. Likewise other text-values, I fear.

That will work too. --Lydia Pintscher (WMDE) (talk) 12:45, 19 September 2012 (UTC)

Namespaces and localization[edit]

On the IRC log I saw the question "will everything be localizable?" was answered as "absolutely". So, one more question on this: Will there be namespaces on the wiki? Will their name change depending on the user language? (currently MediaWiki doesn't do that!) Helder 17:23, 14 April 2012 (UTC)

I've been told Wikidata won't use namespaces. What Denny meant in the office hour was that the content and interface will be localized. --Lydia Pintscher (WMDE) (talk) 13:13, 16 April 2012 (UTC)
Ok sorry. This was a misunderstanding. Wikidata will use namespaces but localizing/translating them would be hard to impossible and it's therefor not on the things we plan to do during the initial development. --Lydia Pintscher (WMDE) (talk) 15:33, 16 April 2012 (UTC)

Other open data projects[edit]

I read a bit about predecessors of Wikidata and that their work should be kept in mind. But I assume Wikidata is not the only open data project out there. So how is Wikidata‘s relationship to other open data projects? Are there cooperations (planned)? Is their structure evaluated? Just wondering, --Alexander Sommer (talk) 19:08, 16 April 2012 (UTC)

We're working together closely with some of them and are generally familiar with at least the major ones, yes. Freebase, DBpedia come to mind. --Lydia Pintscher (WMDE) (talk) 08:49, 17 April 2012 (UTC)

Geographic shapes and OpenStreetMap[edit]

Wikidata/Data model says that Wikidata will accept "geographic shape" data, which is cool. However, such data are virtually absent from Wikipedia and serious efforts to add them to Wikidata would obviously create a massive overlap with OpenStreetMap. Is cooperation planned ? I would be nice to synchronize both projects early on than to let both projects develop separately and inconsistently.--Zolo (talk) 07:06, 27 April 2012 (UTC)

We have Katie on the team who is also involved with OSM. So yes we're keeping all this in mind :) The goal with the geographic shape data also isn't to compete with OSM really. --Lydia Pintscher (WMDE) (talk) 13:50, 27 April 2012 (UTC)
The scope of OSM is to make a street map. The OSM community says that they don't want to include any data beyond that scope - for example species distributions or historical data. --Alexrk2 (talk) 15:03, 9 June 2012 (UTC)
I think that Zolo was talking about cooperation specifically regarding the data types that both projects share. That seems perfectly reasonable, in fact; duplication of work is rather inefficient especially in volunteer-curated data projects. For instance, country boundaries and locations of cities and various landmarks are simple examples (from the top of my head) of data for which there is a clear motivation to exist in both projects, and therefore some setup for regular data sync between the projects would be ideal. --Waldir (talk) 20:24, 9 June 2012 (UTC)

Species name templates - datasource for the translations used in commons in wikidata?[edit]

In Commons:Commons_talk:WikiProject_Tree_of_Life#Kersti_Nebelsiek_idea_2:_template_for_langage we diskuss templates for internationalising vernicular names of animals. There are two parts of the system: first the Commons:Species name templates which contain the translations, 2nd the template Commons:Template:VNIncluded. There Zolo suggested, that it may be better to store the list of translations and interwikis here. But one question remains: Is it possible that the translations show up in the search results in Commons for the corresponding category?--Kersti (talk) 11:06, 29 May 2012 (UTC)

Hi Kersti. Sorry for the late reply. I missed your question. I can't give you an answer to that unfortunately. It's a bit too early in the project. This will become interesting in the second phase starting in some weeks. --Lydia Pintscher (WMDE) (talk) 14:43, 18 June 2012 (UTC)
Vielleicht interessiert dich ja auch diese Idee: Multichill fordert hier internationalisierten Kategorien für Commons. Er hat dort leider keine Lösung vorgestellt, sondern nur eine Forderung aufgestellt. Aus meiner Sicht wäre es praktikabler wenn die Kategorien weiter auf englisch blieben, aber die Übersetzungen hinter der Kategorie in Klammern gezeigt würde, wenn man in die Mutterkategorie oder Suchergebnisse schaut. Als Quelle für eine solche Übersetzung könnte Wikidata dienen und die Vorlage zum aufrufen der Interwikis würde festlegen, welcher Datensatz (Item) aus Wikidata genutzt wird. Das wiederum könnte man dann auch für andere Wikipedias anschaltbar machen. --Kersti (talk) 17:49, 12 June 2012 (UTC)
English: Maybe this idea is interesting too Multichill wants internationalisiced categories for Commons. He doesn't provide a solution for this problem. It would be easier, to leave the categories in english and add the translations in brackets, while looking in the mothercategory or in search results. Datasource would be Wikidata and the interwiki-template would tell which Item in wikidata should be used. This could be used in all wikipedias, if the user doesn't speak the language really good and does work like searching for pictures or similar. --Kersti (talk) 09:42, 20 June 2012 (UTC)
I think Commons could tremendously benefit from Wikidata but that it would require a major overhaul of the whole site, switching from something based on templates and categories to something using Wikibase and calling Wikidata entries for internationalization and short descriptions. It could have major impact on internationalization, searchability, maintainability ... I guess it is too early to discuss that, but perhaps an interesting use case to keep in mind. --Zolo (talk) 10:18, 20 June 2012 (UTC)
Oh I know that such changes take years, and that first the interwikis should be in place and that I second will make up my species name templates. And when this will be done, two parts of the system are in place and the third thing could be implemented as soon as the wiki-software is written in changing the template wich is used to include interwikis from wikidata in each wiki (that is one change per wiki), to implement it. Kersti (talk) 10:36, 21 June 2012 (UTC)

For the record[edit]

For the record, this would be roughly equivalent to Microdata in HTML5, something we're discussing here and in English Wikipedia's Pump. 68.173.113.106 00:20, 11 June 2012 (UTC)

From looking at the page this project aims to embed the information in the page. This doesn't seem to include the part where we store the information in a central repository and have the ability to share this between the Wikipedias themselves and others. —The preceding unsigned comment was added by Lydia Pintscher (WMDE) (talk) 08:23, 11 June 2012
My idea was to embed Wikidata as microdata in infoboxes. 68.173.113.106 22:06, 24 June 2012 (UTC)

Infobox example[edit]

 <table class="infobox" itemscope itemtype="//bits.wikimedia.org/wikidata/infobox-country">
  <tr>
   <td>Full name</td>
   <td itemprop="full-name">United States of America</td>
  <tr>
   <td>Population</td>
   <td itemprop="population">300,000,000</td>
  </tr>
 </table>

How do we go between this and Wikidata? 68.173.113.106 22:49, 24 June 2012 (UTC)

As far as I understand this is not a problem. You will be able to do this using the current template syntax. What Wikidata does is adding a few more parser functions you can use in your template. --TMg 16:44, 18 July 2012 (UTC)

Wikicite[edit]

This project seems to be perfectly suited to Wikicite. Has a requirement for use for references been considered? ··gracefool | 16:52, 20 June 2012 (UTC)

This should indeed be doable with Wikidata. As for requiring references: The software will not require them (but it will be possible to enter them). However the community will decide if and how they require them later. --Lydia Pintscher (WMDE) (talk) 16:57, 20 June 2012 (UTC)


Interwikis for Commons[edit]

In looking on the demo system I notice, that commons doesn't exist there. Commons should have Interwikis as they are needet to maintain Commons. But it should not look up IN the interwikis. And Olt-style-Interwikis won't be maintained, if other Wikipedias don't have them, as there is no longer an easy copying of the whole list. We simply don't have the manpower to do the work there. --Kersti (talk) 07:36, 22 July 2012 (UTC)

What do you mean with "But it should not look up IN the interwikis."? --Lydia Pintscher (WMDE) (talk) 08:58, 23 July 2012 (UTC)
For example, the page commons:Tree contains links to all Wikipedia languages but the Wikipedia pages don't link to the Commons page. --TMg 16:38, 23 July 2012 (UTC)
Yes, because the Commons link usually is listed as link in the page itself and not as interwiki on the side, below the logo. Kersti (talk) 19:07, 23 July 2012 (UTC)
Ah ok that makes sense now. Thanks for the input. I will bring that up. --Lydia Pintscher (WMDE) (talk) 12:41, 24 July 2012 (UTC)
See also bugzilla:708 (interproject links). Helder 15:43, 24 July 2012 (UTC)
That's covered by Bug 35960. Merlissimo (talk) 10:31, 26 July 2012 (UTC)

Interwikis for Wikisource and Wikiquote for example[edit]

An author may have an interwiki list in Wikisource (for example wikisource:Author:Jean Racine) in Wikipedia (for example Wikipedia:Jean Racine) and in Wikiquote (for example wikiquote:Jean Racine). These three lists must be stored in one item, as it is the same person. Each Wiki will usually display somewhere in the page the own language link to the other projects, and the interwiki list of the own project. Is the storing of three or more interwikilink-lists already part of the current software? Kersti (talk) 19:32, 23 July 2012 (UTC)

Yes and no. We want to support this at some point, but do not plan to have this fully implemented in the first version. We are however already building the notion of "site groups" into the code, so it will be easy to add support once we have covered the minimum functionality we promised.
Please also have a look at the Wikisource and Wikiquote use cases on Wikidata/Notes/Future. -- Daniel Kinzler (WMDE) (talk) 10:46, 26 July 2012 (UTC)
Please also note than Wikisource interwiki is more complex than interwiki in Wikipedia. First, the interwiki are not equivalent: for each text we always have one language, that is tha original language of the text. Second, the non-original language versions of the text are translations, posibly many different translations of one text. So few Wikisources support one-to-many interwiki links. Ankry (talk) 17:07, 10 September 2012 (UTC)
In the case of author pages the interwikis are equivalent. And these are the only existing interwikis in source, as far as I know. The book translations are another problem, which is in fact much more complicated. Kersti (talk) 16:08, 1 November 2012 (UTC)

Missing list item in section Further Information of the Wikidata page[edit]

Hello, in section Further Information of the Wikidata page a list item is missing in the English text that is in the German translation, beginning with:

* [[Wikidata/Technical proposal/de|Technischer Vorschlag]]

It seems to be necessary because otherwise the procentual value for the translation is 93 %. The item should be added to translation message Wikidata/4/xx. Otherwise the translation of this string is missing in other languages. Regards, --Michawiki (talk) 17:51, 25 July 2012 (UTC)

PS: Don't forget to change the translation messages when you add or remove text to resp. from the Wikidata page. Otherwise the translations won't be up-to-date. For instance the first item in the box This needs your input the first item has been changed from Vote for a logo to Future use cases on the Wikidata page. But the corresponding translation message has not been changed yet. Besides a logo has been added to the page with the caption The Wikidata project logo. The caption is missing among the translation messages. --Michawiki (talk) 18:21, 25 July 2012 (UTC)

[edit]

What will happen to the CC license on the logo when it is finalized? (Hint hint: it is irrevocable.) 68.173.113.106 00:05, 26 July 2012 (UTC)

It'll stay. I am not sure what your question is trying to imply. --Lydia Pintscher (WMDE) (talk) 10:06, 26 July 2012 (UTC)
I'm pretty sure they're referring to this:
"...by submitting your proposed logo design you grant to Wikimedia Foundation, Inc. an exclusive (even as against you), perpetual, irrevocable, worldwide, fully paid-up license to use, reproduce, and exploit in any way without limitation all copyright, trademark, publicity, and any other intellectual property or other proprietary rights thereto. (...) if your logo design is chosen as the winning adopted logo for the Wikidata project, you will be required to enter into a contract assigning and/or licensing all of your rights thereto to Wikimedia Foundation, Inc."
That section's language and length was conveys an overal slightly too defensive (maybe not the best word here, I hope it gets the idea across) position from the Wikidata team, so I understand 68.173.113.106's concerns. --Waldir (talk) 11:43, 26 July 2012 (UTC)

Data relationships[edit]

Hello,

I was wondering whether it will be possible to have relationships between data on different pages. For example, if one adds Jeremy Bentham to Peter Singer's "Influenced by" entry, will Peter Singer automatically be added to Jeremy Bentham's "Influenced" (see [3] for an interesting application)? Or will this have to be done by bots? InverseHypercube (talk) 19:27, 29 July 2012 (UTC)

From what I know this will not happen automatically. You can not connect different properties. --TMg 02:11, 30 July 2012 (UTC)
That would be a real shame. At least I hope this is in the plans for future improvement. --Waldir (talk) 13:37, 30 July 2012 (UTC)

We do not plan to support semantics for properties or relations, and thus will not support the notion of one relation being the inverse of another. So, "A influenced B" will not imply "B was influenced by A".

However, we will support backlinks, and I could imagine showing backlinks with their respective properties directly on the page. That is, if Peter Singer has an "influenced by" property pointing to Jeremy Bentham, then perhaps on Jeremy Bentham's page this could show up in a "References" section where it says that "Peter Singer was influenced by Jeremy Bentham".

The difference is that we may support showing the same relation on both pages involved (the subject and the object), but we will not have one property imply another. We'll have to find a way to make this feasible for items that are linked thousands of times, though. Anyway, this is an important distinction: we can add features aiding with structure based navigation, but we are not going to build any semantic relationships into Wikidata. -- Daniel Kinzler (WMDE) (talk) 14:35, 30 July 2012 (UTC)

Why not? I assume this has been explained somewhere, so if you could, just point me there, please. --Waldir (talk) 17:42, 30 July 2012 (UTC)
I think sort of that functionality is absolutely needed for Phase3 (lists/tables). Different tables will need to be built from one dataset, for example "winners at the olympic games 2012" on one page and "winners of badminton in olympic games" on an other one. I guess this will need to be built as an own set of objects, as it could be done with any relation between objects (we know this schema with an extra table from n:m-relations in relational database systems). This could be done in the above example, too, by having a set of objects representing influences in the creative business.
However, it would be very nice to have such relations as properties on objects as well. The "Usain Bolt" object would have a field "wins", one of them beeing his result from London 2012. -- (talk) 15:18, 10 September 2012 (UTC)
I think this functionality is absolutely needet for phase two too, als the main content of taxoboxes are relations to other taxa. Like "This species belongs to family xy" - and as sometimes there are two genuses with the same name - one for a plant and one for an animal, there must be mentioned which item is the correct one.
Furthermore the current scientific name must be handled in a different way compared to the name in one language, als it is relevant for each language version. --Kersti (talk) 19:28, 11 October 2012 (UTC)

Data communication between Wikipedia and Wikidata[edit]

I'm just wondering: did somebody an estimation of the time and of the bandwidth necessary to exchange data between a single article with a single infobox of wikipedia and the future wikidata repository ? And then multiply that numbers with the daily traffic on all wikipedias ? I'm not an expert but I think we can forget to build a normal infobox in the way explain in these pages. I think it is not necessary in most cases to update the data of a wikipedia article everytimes someone opens the page too.

My proposition is to use bots on each wikipedia which will update every day, every week or every month the different infoboxes by sending a request to wikidata repository which will extract the information and will format the infoboxes in the normal wikitext before sending them to the bots. These latter will delete the previous code of the infoboxes and replace it with the updated wikicode.

This method allows each wikipedia to select the data and to organize the data according to their own criteria. This will allow a reduction of the pressure on wikidata to organize a complex system of translation by transfering the charge of the translation of the request to each wikipedia through the bots.
We can imagine to do the same for data upload: each wikipedia can take the charge of an interface to export data to the wikidata repository.
That's just an opinion but I'm interested in all comments. Snipre (talk) 19:39, 10 August 2012 (UTC)

There will be a lot of caching and stuff like that. The bot proposal has come up a few times but it would kind of defeat the point of having one central data repository and make things quite a bit more complicated because you'd have to merge changes for example. --Lydia Pintscher (WMDE) (talk) 13:10, 13 August 2012 (UTC)
There is some discussion on that at Wikidata/Notes/Caching investigation. I don't think there are any problems with translations, most of them should be covered by the WikiData infrastructure itself.
However, I think you are right when saying that each Wikipedia might need some time to select and reduce the data they need for their single project. --Bergi () 15:28, 10 September 2012 (UTC)
I support this project, as it seems beneficial to the Wiki-community as a whole (though this will, of course, be dependent upon how it is implemented). My main concern, and why I would oppose it at the present time, is that the Wikimedia servers seem to be getting slower compared to when I started contributing to Wikipedia four years ago. Back then, I never came across the 'Wikimedia error' page, but now I come across it more and more frequently. The more often a user comes across the Wikimedia error page, I would guess that they are more likely to stop contributing to Wikimedia projects all together: Why spend time editing or writing articles, if all you get is that annoying blue page saying there is a server error? The BBC recently told us that once the failure rate of a CAPTCHA reaches 75%, people stop trying to accessing the site, and that is for a task where one is required to type merely two words! Much less than writing or editing an article.
By all means, continue building this database, as it definitely will be useful (even as just a source that users can refer to when writing articles). However, given that the Wikimedia servers seem to be slowing down, and are returning error messages far more often than they should, I do not support any further implementation of such a project that would make pages heavier and more difficult to load. I cannot support any further implementation of this project till the incidence of the blue page proclaiming 'Wikimedia error' has been reduced significantly. V85 (talk) 18:20, 7 October 2012 (UTC)

Using wikidata to co-ordinate image use[edit]

Sorry if this has been posted before, but once phase two is completed, would it be possible to use Wikidata to monitor the presence of commons files on individual Wikipedia articles?

For instance, the footballer John Barnes has roughly 20 articles across various Wikipedia, assuming that his page on the English Wikipedia contains every interwiki. Commons currently has one good image of the subject, but at the time of his post only 15 of his articles use it. If it helps, the vast majority of his articles have infoboxes, including those without images.

The above is just one example, but I have done large scale image runs maybe 5 or 6 times, and it's pretty tedious work when you don't speak the languages. If it were possible to do that sort of work by only editing an English-speaking page, I would do far more of it. WaitingForConnection (talk) 18:49, 18 August 2012 (UTC)

In commons there is the project Commons:Commons:Valued images - if these pictures were listed here, the most valued picture could be found automatically and added to taxoboxes and other templates on a page. Kersti (talk) 20:43, 18 August 2012 (UTC)
It should be possible to connect items (like the football player in this case) to images or pages on commons and then build templates that use them. --Lydia Pintscher (WMDE) (talk) 10:45, 20 August 2012 (UTC)
What if we just add a property called "Image" or something like that? It may be useful also for coats of arms and flags/banners of administrative divisions all around the world (not to speak of countries). Sannita - not just another it.wiki sysop 10:10, 22 August 2012 (UTC)

Use DataMining[edit]

Please use various Data Mining techniques in large extend on databases for making Wikidata as Knowledge Source rather than Content source. Sujitkumarpatil (talk) 14:51, 6 September 2012 (UTC)

This is not the aim of WikiData (see also Wikidata/Notes/Requirements). We want to build only the infrastructure for collecting and maintaining data in a wiki way - making it usable on other projects -, and expect others to build cool apps that mine interesting "knowledge" from it. --Bergi () 15:33, 10 September 2012 (UTC)

User:DrTrigonBot subster as data source[edit]

Hello all!

Lydia announced the demo roll-out today at commons. I think I mentioned it already some other place and some time ago; What about or how to setup DrTrigonBot as a bot feeding to and updating data in wikidata? Or will there be a more sophisticated way to do this? At the moment DrTrigonBot subster has to be setup for every single wiki that wants to use its service, which does not make much sence, e.g. maintenance, cpu time, ... In my oppinion either my bot should be completely replaced by wikidata or become an essential addition to it. What do you think? Thanks and Greetings --DrTrigon (talk) 13:45, 10 September 2012 (UTC)

Hey :) I am having a hard time finding out what exactly the bot does and how it works and therefore can't really answer your question at the moment. Can you give me the short version or point me to a page with some more information? If you want we can discuss this on IRC in #wikimedia-wikidata. (But here is fine too.) --Lydia Pintscher (WMDE) (talk) 14:00, 10 September 2012 (UTC)
Yeah... sorry for that ;) Please confer w:en:User:DrTrigonBot/Subster or w:de:User:DrTrigonBot/Subster. Basically the bot enables you to link data from external pages here to wikipedia. The bot runs once a day and checks if the data have changed and if needed updates the ones in the wiki. Greetings --DrTrigon (talk) 14:59, 10 September 2012 (UTC)
Thanks! Ok so if I understood this correctly it basically keeps things like the number of inhabitants in an article up-to-date whenever the place where they're originally published changes them. Then there are two answers to your question. One: There will still (probably for quite a while) be articles that have not moved to using Wikidata yet and who will continue to use the service the bot provides. Migration will take a while obviously. Two: We (the team doing the initial development of Wikidata) will not write bots or anything to update data in Wikidata. It'd be very cool if you would adapt your bot for this. The necessary API will be available later. Merlissimo already worked on a bot to handle language link migration and wrote Pywikidata. I think you two should talk. All in all the Wikidata part of this should be easier since you'll have a proper API and everything and don't need to do anything fancy. But it will obviously require changes to the bot if you're willing to make them. I am happy to help with whatever you need. Just let me know. --Lydia Pintscher (WMDE) (talk) 15:11, 10 September 2012 (UTC)
Aaa - that sound great - whenever things are getting interesting, Merlissimos name appears on the screen... good thing to hear! ;)) I am interested in adopting the bot and thus will contact him! What would be the best place to create a page in order to keep track of this small "sub-project"? Greetings --DrTrigon (talk) 17:29, 10 September 2012 (UTC)
Sweet! How about Wikidata/Bots? --Lydia Pintscher (WMDE) (talk) 17:31, 10 September 2012 (UTC)
Perfect! Thanks for getting it started :) --Lydia Pintscher (WMDE) (talk) 19:01, 10 September 2012 (UTC)
Your welcome! I hope you can help to make popular and more useful...? ;) To me the most useful found at the moment were; Wikidata/Data collaborators#User:DrTrigonBot (which I forgot myself) and Wikidata/Volunteers#Adapting bots ... so I hope the page can help us to keep track and simplify the process! Greetings --DrTrigon (talk) 19:09, 10 September 2012 (UTC)
I posted a note to identi.ca, twitter and the mailing list. --Lydia Pintscher (WMDE) (talk) 19:16, 10 September 2012 (UTC)
Nice job! :)) I had a look into Pywikidata and am starting to understand, but there are open questions, please have a look at Wikidata/Bots#User:DrTrigonBot (SubsterBot). Thanks again and greetings --DrTrigon (talk) 20:36, 10 September 2012 (UTC)
Please update message Translations:Wikidata/16 in the Translation Tool that the new entry for the bots page will be translated into other languages as well. Regards, --Michawiki (talk) 23:49, 10 September 2012 (UTC)
Done :) --Lydia Pintscher (WMDE) (talk) 12:55, 11 September 2012 (UTC)
You'll no longer need a template like that. The API will come together in the next weeks. I think the rest of your questions can only be answered when that's in place. --Lydia Pintscher (WMDE) (talk) 12:55, 11 September 2012 (UTC)
Could you may be keep me up-to-date, e.g. by pinging me on important events? That would be cool... ;) ...otherwise I will try following all changes and updates by myself... (e.g. is there a maillist? or could you post to pywikipedia-l?) Thanks a lot and greetings --DrTrigon (talk) 14:27, 11 September 2012 (UTC)
I will try to, yes. https://lists.wikimedia.org/mailman/listinfo/wikidata-l is the mailing list. And you can subscribe to the on-wiki newsletter. That'll give you weekly summaries of everything that's important. These summaries are also posted to the mailing list. --Lydia Pintscher (WMDE) (talk) 14:32, 11 September 2012 (UTC)
Thanks a lot!!! I'm on the maillist now, let's see how that works... ;)) Greetings and thanks! --DrTrigon (talk) 21:59, 11 September 2012 (UTC)

I tried[edit]

Yet, 16+5=21 and 49-8=41 in my opinion. So, the system did not accept my comment which was: if the title of the page should be the same as the label, I don't see how I can link an article Гелій to Helium data. Regards. --93.73.19.163 14:52, 10 September 2012 (UTC)

Are you trying this on the demo system? To link an article to the Helium data you go to the repo part of the demo. Then you search for the page on Helium. There you see a table of languages and links. If what you want to add is not already there you can add it using the link at the end of the table. Does that answer your question? --Lydia Pintscher (WMDE) (talk) 15:15, 10 September 2012 (UTC)
No, I want to create a new article entitled Гелій. And I don't want to copy all wikilinks from Helium data to Гелій data. This would be unreasonable. I want my article entitled Гелій to use Helium data. Otherwise, this is just an English project and you lose me - over 500K people will cope. Is this exclusively an English language project? And the system really has problems with 16+5. --93.73.19.163 15:36, 10 September 2012 (UTC)
Right. That's the whole point of Wikidata in the future. If you go to the demo and add a new link to the Helium page on the repository it will show up in the client. Keep in mind though that this is a demo with a few limitations so what you are trying to do might not work properly. I'd need to know more to tell though. --Lydia Pintscher (WMDE) (talk) 16:35, 10 September 2012 (UTC)
Hi I think what 93.73.19.163 means is that he landed in the english localised version of the Wikidata repository were the page (not article that is part of the confusion I think) is entitled Helium and followed by an english short description. 93.73.19.163 if you look in the upper left corner were normaly your userstuff is you see an link english there you can change to русский and refresh. Now the page title should be Гелий because you now see the Rusian Version. Hope this is what was meant. I somehow already got the German version in the first place does it use the language of my browser? This leads to the question how you can add localizations. --Saehrimnir (talk) 14:48, 11 September 2012 (UTC)

Central Babel repository?[edit]

What about a central repository for language skills indications? --Hellsepp (talk) 00:00, 11 September 2012 (UTC)

Interesting idea. I am not sure if Wikidata is the right place for this though. I feel this should be handled in the user's preferences. But why don't you add it to Wikidata/Notes/Future and see where it goes? --Lydia Pintscher (WMDE) (talk) 10:43, 11 September 2012 (UTC)
If you mean a way to provide user language skills without having to create local templates, we already have mw:Extension:Babel.
If you mean a place to store all user names + associated language skills, we should probably have global user pages for all Wikimedia projects. The current Wikiverse structure seems way too complex. To some extent, this may just be a naming problems. For instance Meta, Commons, Wikidata, and even local village pumps are all to some extent "meta" and/or "commons" things. And I don't even mention the "MediaWiki"/ "Wikimedia" conundrum (neither of them being related to media). Streamlining that should be on the agenda, but I am not sure it can have anything to do with Wikidata. --Zolo (talk) 12:36, 11 September 2012 (UTC)
See mw:GlobalProfile/design. --Yair rand (talk) 16:50, 11 September 2012 (UTC)
Sounds great thanks. --Zolo (talk) 07:08, 12 September 2012 (UTC)

TrustedXFF: hosts file missing on wikidata-test-client.wikimedia.de[edit]

Some files for TrustedXFF seem to be missing on the test client, on http://wikidata-test-client.wikimedia.de/wiki/Main_Page I get (I replaced the IP of my proxy with x.x.x.x)

TrustedXFF: hosts file missing. You need to download it.

Backtrace:

#0 /var/www/wikidata-test-client.wikimedia.de/w/extensions/TrustedXFF/TrustedXFF.php(66): TrustedXFF->getCdbHandle()
#1 /var/www/wikidata-test-client.wikimedia.de/w/extensions/TrustedXFF/TrustedXFF.php(39): TrustedXFF->isTrusted('x.x.x.x')
#2 [internal function]: TrustedXFF::onIsTrustedProxy('x.x.x.x', false)
#3 /var/www/wikidata-test-client.wikimedia.de/w/includes/Hooks.php(216): call_user_func_array('TrustedXFF::onI...', Array)
#4 /var/www/wikidata-test-client.wikimedia.de/w/includes/GlobalFunctions.php(3775): Hooks::run('IsTrustedProxy', Array)
#5 /var/www/wikidata-test-client.wikimedia.de/w/includes/ProxyTools.php(72): wfRunHooks('IsTrustedProxy', Array)
#6 /var/www/wikidata-test-client.wikimedia.de/w/includes/WebRequest.php(1088): wfIsTrustedProxy('x.x.x.x')
#7 /var/www/wikidata-test-client.wikimedia.de/w/includes/User.php(1723): WebRequest->getIP()
#8 /var/www/wikidata-test-client.wikimedia.de/w/includes/db/Database.php(828): User->getName()
#9 /var/www/wikidata-test-client.wikimedia.de/w/includes/db/DatabaseMysql.php(158): DatabaseBase->query('SET NAMES binar...', 'DatabaseMysql::...')
#10 /var/www/wikidata-test-client.wikimedia.de/w/includes/db/Database.php(631): DatabaseMysql->open('localhost', 'root', '', 'my_wiki')
#11 /var/www/wikidata-test-client.wikimedia.de/w/includes/db/Database.php(681): DatabaseBase->__construct('localhost', 'root', '', 'my_wiki', 16, 'get from global')
#12 /var/www/wikidata-test-client.wikimedia.de/w/includes/db/LoadBalancer.php(711): DatabaseBase::factory('mysql', Array)
#13 /var/www/wikidata-test-client.wikimedia.de/w/includes/db/LoadBalancer.php(586): LoadBalancer->reallyOpenConnection(Array, false)
#14 /var/www/wikidata-test-client.wikimedia.de/w/includes/db/LoadBalancer.php(507): LoadBalancer->openConnection(0, false)
#15 /var/www/wikidata-test-client.wikimedia.de/w/includes/objectcache/SqlBagOStuff.php(117): LoadBalancer->getConnection(-2)
#16 /var/www/wikidata-test-client.wikimedia.de/w/includes/objectcache/SqlBagOStuff.php(177): SqlBagOStuff->getDB()
#17 /var/www/wikidata-test-client.wikimedia.de/w/includes/objectcache/SqlBagOStuff.php(165): SqlBagOStuff->getMulti(Array)
#18 /var/www/wikidata-test-client.wikimedia.de/w/includes/cache/MessageCache.php(332): SqlBagOStuff->get('my_wiki:message...')
#19 /var/www/wikidata-test-client.wikimedia.de/w/includes/cache/MessageCache.php(723): MessageCache->load('en')
#20 /var/www/wikidata-test-client.wikimedia.de/w/includes/cache/MessageCache.php(661): MessageCache->getMsgFromNamespace('Gadgets-definit...', 'en')
#21 /var/www/wikidata-test-client.wikimedia.de/w/includes/Message.php(655): MessageCache->get('gadgets-definit...', true, Object(Language))
#22 /var/www/wikidata-test-client.wikimedia.de/w/includes/Message.php(533): Message->fetchMessage()
#23 /var/www/wikidata-test-client.wikimedia.de/w/extensions/Gadgets/Gadgets_body.php(581): Message->exists()
#24 /var/www/wikidata-test-client.wikimedia.de/w/extensions/Gadgets/Gadgets_body.php(38): Gadget::loadStructuredList()
#25 [internal function]: GadgetHooks::userGetDefaultOptions(Array)
#26 /var/www/wikidata-test-client.wikimedia.de/w/includes/Hooks.php(216): call_user_func_array('GadgetHooks::us...', Array)
#27 /var/www/wikidata-test-client.wikimedia.de/w/includes/GlobalFunctions.php(3775): Hooks::run('UserGetDefaultO...', Array)
#28 /var/www/wikidata-test-client.wikimedia.de/w/includes/User.php(1244): wfRunHooks('UserGetDefaultO...', Array)
#29 /var/www/wikidata-test-client.wikimedia.de/w/includes/User.php(2230): User::getDefaultOptions()
#30 /var/www/wikidata-test-client.wikimedia.de/w/includes/context/RequestContext.php(274): User->getOption('language')
#31 /var/www/wikidata-test-client.wikimedia.de/w/includes/Message.php(327): RequestContext->getLanguage()
#32 /var/www/wikidata-test-client.wikimedia.de/w/includes/context/RequestContext.php(353): Message->setContext(Object(RequestContext))
#33 [internal function]: RequestContext->msg('pagetitle')
#34 /var/www/wikidata-test-client.wikimedia.de/w/includes/context/ContextSource.php(166): call_user_func_array(Array, Array)
#35 /var/www/wikidata-test-client.wikimedia.de/w/includes/OutputPage.php(833): ContextSource->msg('pagetitle')
#36 /var/www/wikidata-test-client.wikimedia.de/w/includes/Article.php(560): OutputPage->setPageTitle('Main Page')
#37 /var/www/wikidata-test-client.wikimedia.de/w/includes/actions/ViewAction.php(37): Article->view()
#38 /var/www/wikidata-test-client.wikimedia.de/w/includes/Wiki.php(427): ViewAction->show()
#39 /var/www/wikidata-test-client.wikimedia.de/w/includes/Wiki.php(304): MediaWiki->performAction(Object(Article))
#40 /var/www/wikidata-test-client.wikimedia.de/w/includes/Wiki.php(536): MediaWiki->performRequest()
#41 /var/www/wikidata-test-client.wikimedia.de/w/includes/Wiki.php(446): MediaWiki->main()
#42 /var/www/wikidata-test-client.wikimedia.de/w/index.php(59): MediaWiki->run()
#43 {main}

--Schnark (talk) 10:22, 12 September 2012 (UTC)

Thanks for letting us know. This should be fixed now :) --Lydia Pintscher (WMDE) (talk) 12:38, 12 September 2012 (UTC)
Yes, it now works as expected. Thanks for fixing this issue so quick. --Schnark (talk) 07:15, 13 September 2012 (UTC)

Suggestions for the respository[edit]

I've created a page for an Italian poet and I've added some links, but can't this be done automatically? For example: one copies the list of interwikis from the English Wikipedia, pastes it into a text box, adds the English-language article itself to the list, clicks a button and the page is generated... It would be easier this way.

Another suggestion: the links should be sorted alphabetically regardless of when they were inserted.--Erasmo Barresi (talk) 07:56, 13 September 2012 (UTC)

We won't do that but I am pretty sure that this will be done by someone very soon. Someone is already running a bot for example who can do most of this automagically. As for sorting the links: They are sorted when you reload the page. It'd be confusing if they get added somewhere in a long table immediately after entering them. --Lydia Pintscher (WMDE) (talk) 09:24, 13 September 2012 (UTC)
Thanks! Your work is fantastic!--Erasmo Barresi (talk) 12:10, 13 September 2012 (UTC)
Awww thank you! --Lydia Pintscher (WMDE) (talk) 12:29, 13 September 2012 (UTC)

A few thoughts about the demo[edit]

I know things are still a work in progress, but a few thoughts as I understand first real deployment is coming soon ;)

  • When the label is not filled out in the user's language, there is an "enter label" message. Shouldn't it somehow also display the label in a fallback language ? I find it a bit disturbing to have pages with no explicit title (plus for places, people etc., the label is often the same in many languages).
  • Apparently, aliases are just very simple properties with no kind of qualifiers, but how exactly are they supposed to be used ? Is it to help the search engine ? But searching for "He" should probably not redirect to "Helium". Or is it to inform people about synonyms ? In that case, qualifiers like "obsolete" or "scientific name" may be useful so that people do not believe that "Daddy can we get a dog ?" can be equivalently rendered as "Daddy can we get a canis lupus familiaris ?".
  • Weren't the language names for interwikis supposed to be in the user's language ? It would be prettier and more consistent with the stated goal of full internationalization. --Zolo (talk) 07:46, 14 September 2012 (UTC)
Yes this is still coming. See https://bugzilla.wikimedia.org/show_bug.cgi?id=36430
Yes it is to help people search. If there is more than one hit there will be a list of them of course.
That's part of the work of the universal language selector which we're using as far as I know. --Lydia Pintscher (WMDE) (talk) 11:57, 14 September 2012 (UTC)
Thanks, but whatever the language I choose I get "Deutsch: Helium", rather than say: "German: Helium". --Zolo (talk) 14:09, 14 September 2012 (UTC)
Actually I guess that what bothered me about the "aliases" thing it the word alias. Saying that "He" is an alias of "helium" seems to suggest that "He" redirects to "Helium". Since that is not the case, isn't there a better word than alias ? --Zolo (talk) 09:13, 27 September 2012 (UTC)
Hmm in a sense it is a redirect. If you search for He for example then Helium will show up as a result. But if you have ideas for a better wording please let me know. --Lydia Pintscher (WMDE) (talk) 14:48, 27 September 2012 (UTC)

Adding to the interwiki sets locally[edit]

If I created the article nn:Fisk, I would like it to be possible type in en:Fish somewhere while I am creating it, and this way add nn:Fisk to the same data set (item) that en:Fish belongs to, once I save the page. Has there been any talk about/work on something like this? Right now, it seems that adding interwiki links via the common repository website is the only way possible.

The actual edit on the repository site would be under the name of the person who added the link from the local wiki, just as if he had visited the global repository and added the link via the interface there. Editing interwiki locally could also be restricted to adding interwiki only, which is to say that the option only exists for pages that have no interwiki; if you want to remove/replace the current interwiki set, you need to visit the common repository.

More specifically, I would want e.g. two input fields (one for the language code, and one for the page name) either below or above the edit box. If the article already belongs to an interwiki set, the input fields could e.g. be replaced with a link to the interwiki set on the common repository. Njardarlogar (talk) 11:13, 14 September 2012 (UTC)

One more thing: will there be a local notice on recent changes when an article is either added or removed from the global interwiki set? This would be beneficial in defeating interwiki vandalism. If there are no local notices, then only "repository guards" can spot vandalism like this, and they may not be the best equipped otherwise they understand the relevant language(s). Njardarlogar (talk) 11:27, 14 September 2012 (UTC)
About your first point: Yes there are already mock-ups for this. Feedback welcome!
About your first point: There will be notifications like this. They still need to be worked on though. --Lydia Pintscher (WMDE) (talk) 11:52, 14 September 2012 (UTC)
Right, that looks good. Two other things, then: has there been any communication with the Pywikipedia team regarding moving interwiki that has no conflict? It could be moved pretty fast to the repository (and the old interwiki lists be removed from the articles) by all the bot operators if a standard script was released as a part of Pywikipedia. And, when this stuff is activated, if an article is connected to an interwiki set but also has a local old-style interwiki list, will the two lists merge/create duplications, or will only one list be displayed? Njardarlogar (talk) 13:43, 14 September 2012 (UTC)
Jep there is already Pywikidata for example :) Coordination is happening on Wikidata/Bots. --Lydia Pintscher (WMDE) (talk) 13:49, 14 September 2012 (UTC)

Preferred method for bulk import into wikidata?[edit]

Is there already a use case for taking an external database or scraped web data, creating statements, and uploading in bulk to wikidata? Could the data be prepared in an RDF file and then just uploaded and, if necessary, merged with existing data? --Robert.Baruch (talk) 14:45, 30 September 2012 (UTC)

They'd need to be entered through the API. But please have a look at the FAQ and the Data collaborators page. --Lydia Pintscher (WMDE) (talk) 08:45, 1 October 2012 (UTC)

Integration into aggregators and metadata platforms[edit]

Are there data aggregators that are already incorporating WD material and dialect into their collections? Is WD considering being an active part of such efforts? Taking part in discussions about how to adopt uuid and cross-ref standards that would minimize unwanted duplicates and support annotation and clustering across different data holders?

Cheers, SJ talk  14:24, 15 October 2012 (UTC) (thinking about the dp.la content hubs, hypothesis, and global equivalents)

Since WD hasn't launch yet, "no" to your first question :) Also, we are not currently planning to actively push in this direction. But yes, we plan to support anyone who wants to use as an Identity-provider i.e. as an EntityBank, this is one of the external effects that we explicitly hope to achieve. But we need to launch at least Phase 1 and 2 for that, in order to push that further. --denny (talk) 15:24, 15 October 2012 (UTC)

Please, mark for translation[edit]

Please, mark for translation the subpages of the project!

Eloy (talk) 23:30, 22 October 2012 (UTC)
Which ones specifically? I'd like to not overload translators with stuff that doesn't really make sense to translate so I am not marking everything as for translation. --Lydia Pintscher (WMDE) (talk) 05:28, 23 October 2012 (UTC)

Supporting categories in wikidata[edit]

Hi,

In many cases categories can be similar in all languages.for example: Article en:Lionel Messi has more than 20 categories that we can divide them in to stage

1-Categories that need to assume feminine or masculine in languages : this cases can use in many languages without any modification (only translation) such as Farsi or English that they don't have different verb or adjective for feminine or masculine e.g. en:Category:Footballers at the 2008 Summer Olympics or fa:رده:بازیکنان فوتبال بازی‌های المپیک تابستانی ۲۰۰۸ but in some languages according to feminine and masculine they have different interwikis.

2-Categories that they don't need to assume feminine or masculine in languages:these cases can be used in any languages and they will have only one translation on other wikis. e.g. en:Category:Football at the 2008 Summer Olympics or en:Category:2008 in sports or en:Category:2008

My suggestion

In my opinion we can add possibility to wikidata for adding translated version of categories number 2 to all articles that are listed in wiki data like what it will do for iterwikis and for categories number 1 users can add it directly to locale wiki.

Benefit

I will help all wikis to have good categorizing their articles and using other communities categorizing works.

p.s.
Now in fa.wikipedia we developed a bot that adds related categories to related pages automatically ,according to en.wikipedia's pages, it helps us to categorizing our articles, templates also categories in precise way.adding cat to category, adding cat to articles, adding cat to templates Aliaram (talk) 09:17, 25 October 2012 (UTC)

Where is the place to discuss the project as a whole?[edit]

Hi!

There are many places to discuss wiki-data, but i didn't found the place to discuss the goals of the project or new features etc. I am looking for something like Projektdiskussion on de.WP.--Svebert (talk) 16:15, 1 November 2012 (UTC)

Come here :) Przykuta (talk) 16:29, 1 November 2012 (UTC)

Source mirror: Instead of only pointing on the source, copying it and saving it on wiki-data[edit]

moved to [4]--Svebert (talk) 17:38, 1 November 2012 (UTC)

Move to Wikidata[edit]

The pages about Wikidata should be moved to Wikidata. Is there a possibility to keep the history with the content? --Kolja21 (talk) 14:59, 19 November 2012 (UTC)

Definitely! To Wikidata:Project_chat? Mange01 (talk) 00:07, 19 January 2013 (UTC)
Please coordinate with me before moving more pages. --Lydia Pintscher (WMDE) (talk) 09:21, 19 January 2013 (UTC)

Metrics of content stability[edit]

Hi! I'm new around here and I'm not sure where's the best place to get feedback, but this looks a promising place :-)

I've just added a proposal at Wikidata/Preventing unwanted edits about ways to allow programmers to assess the rate at which some resources are edited. The idea is basically that edit wars should not lead bots and scripts astray; for that, programs using Wikidata should be able to detect values that are changing quickly or recently, and review the level of trust that they place on them.

The idea of ranks is a good and simple base for this goal, but might not be enough if their values can be manually changed by editors at any time. Having some stability metrics so that programmers can develop their own ranking systems would be preferred. What do you think? Diego Moya (talk) 19:13, 14 December 2012 (UTC)

Phases hard to find[edit]

I was looking for the definition of the three phases and I had quite a hard time to find it, even though I had already read it in the past. Looks like it's on Wikidata/Technical proposal and that one can arrive to it only through a link buried in the end of Wikidata/Introduction, which is linked from the main page. --Nemo 10:55, 13 January 2013 (UTC)

Demo does not accept login from other Wikimedia projects?[edit]

I can't get the "Demo system" to accept my standard Wikimedia login. Is that intentional or accidental? I like having my work on Wikipedia and Wikinews (in various languages) being connected, and I'm not eager to do things without that central connection. Thanks, DavidMCEddy (talk) 09:12, 22 January 2013 (UTC)

Yes that demo system is not connected to central auth so you can't sign in with that account there. You can use it without signing in however. --Lydia Pintscher (WMDE) (talk) 19:38, 22 January 2013 (UTC)

Removing interwiki links and there reversal[edit]

There has been some great progress in removing interwiki links because of the new format used. However many many of these edits are being reverted by people not aware of the situation - is there a link to the topic (a talk) that can be added to the edit summary to guide our editors to a page that explains whats going on? We cant have editwars over this all we need is an explanation for all to see.Moxy (talk) 23:42, 14 February 2013 (UTC)

Are they humans or bots? --Rschen7754 23:50, 14 February 2013 (UTC)
You can point them to this blog, d:Wikidata:Introduction and meta:Wikidata/Deployment Questions for example. Improvements to the latter two are very welcome. --Lydia Pintscher (WMDE) (talk) 10:48, 15 February 2013 (UTC)
I think you mean Wikidata/Deployment Questions. --Michawiki (talk) 20:39, 15 February 2013 (UTC)

Wikidata and WIkimedia project[edit]

Wikidata currently supports only Wikipedia. Will it support other projects in the future? I'm guessing the format will have to be changed, like items will be called WikipediaItem:Code (or some short for it) or by the letter which appears before the code (like Q for Wikipedia). Galzigler (talk) 12:51, 15 February 2013 (UTC)

We have sister projects in mind. It still needs to be thought through though. I hope we have more details about that soon. --Lydia Pintscher (WMDE) (talk) 13:51, 15 February 2013 (UTC)

What happens with old data?[edit]

Until now, I reused the old data in the article, while replacing with new one. For example population or mayor of a town: I put it in a table and added the year (mostly in african villages and towns, information hardly to find). So Wikipedia is not only up to date, but also a book of history. When data will be updated by machines, some data can get lost. Or has Wikidata also a history channel or an archive? --Eruedin (talk) 15:59, 22 February 2013 (UTC)

Wikidata will be even better than the current system ;-) It does keep a history just like the article history on Wikipedia (in fact it is the same thing). At the same time it is able to hold historic values on the page along with the current ones. We're still implementing the details of this but it should be able to handle everything you want there. --Lydia Pintscher (WMDE) (talk) 17:21, 22 February 2013 (UTC)

Why implementation in phases[edit]

Has anyone thought that the implementation in phases (first en-wiki then this, this and that) is a very bad idea. Now all wikis not using Wikidata have the problem that they don't find interwiki links anywhere. First rule of software logistics is that if you upgrade one component, you have to upgrade at the same time all components that are using the upgraded component. Impementation of Wikidata in phases causes a significant loss of information (and lots of addtional effort afterwards) because users cannot create interwiki links or don't know how to mix wikidata links with traditional iws (which again causes a ot of clean ups).--Nedergard (talk) 16:21, 26 February 2013 (UTC)

It will all clear up soon when the other Wikipedia follow on March 6. The old system continues to work. We have specifically not rolled it out on all Wikipedias at the same time in case something breaks or we made mistakes in terms of the load it would create. This way at least not everyone would be affected. The same thing is done for deployments of other extensions like the visual editor for example. --Lydia Pintscher (WMDE) (talk) 16:28, 26 February 2013 (UTC)
Before that happens, how are you supposed to create interwikis for example for [5]? A recent significant event for which now all editors in wikis not using either don't add iws or try - with considerable effort - to create some kind of links. It doesn't help if it works March 6 if you need it now. I'm sorry to disagree but comparison to other extensions is a poor one. Interwikis are used by all wikis interactively and simultaneously at the same time and propably every second. Cutting off the most crucial wiki (= en.wiki) even for couple of hours is not really acceptable. Not speak of weeks. We are very unhappy that the functionality was not tested properly so that it could have implemented in every wiki at the same time. (as a PS I refer again to the general principles of software logistics: never let a new functionality go live, if you are not 200 % sure it's really gonna work;) )--Nedergard (talk) 18:10, 26 February 2013 (UTC)
I understand the problem. And I'm sorry this situation is not ideal. In the case of the papal conclave the interwiki links can still be added to the wiki text. They will still work and interwiki bots can continue to connect pages on these wikis. I am aware that this isn't perfect but this is a temporary thing that will be sorted out very soon. As for testing: The functionality was tested properly. That doesn't mean nothing was overlooked. You can never be 200% sure something works correctly. You have to get software out there to get it tested and get feedback. --Lydia Pintscher (WMDE) (talk) 18:57, 26 February 2013 (UTC)
Sorry to nag again - but now that WikiData is live there are lots of other problems. OK, the first big problem is that the project went live in en-wiki, which is not the primarily user of interwikis because it is the source of them. This leads to bigger problems as the Wikidata principles are now fixed with the feedback from en-wiki users (no blame intended but how users decide on this when 99.99 % of them don't even have to consider about iws?). The other problems follow from the first one. Of course an en-wiki user would say that redirects are not allowed in Wikidata because they think that of course you write in every language about this subject... What they fail to see is that these redirects are placeholders - in the hope that in the future someone will create an article about it but in them meantime it will redirect to concept one level up. One other big annoying thing is that Wikidata links don't update on real-time. There are even instructions to open the page for editing, then save without changes, in order to get the language links visible. Everyone unterstands that this is plain stupid and a waste of a lot wiki editors time. My wish is just that when Wikidata goes to the second phase (templates) you will do it better and include in the first testing phase besides en-wiki at least: one European language with special letters (look up to the Northen countries), Chinese, at least one RTL (right-to-left) language, and most importantly all countries that have non-Latin numbers (Arabic, Thai etc.). But WikiMedia should already know these things.--Nedergard (talk) 18:26, 9 March 2013 (UTC)
Phase 1 was tested on other languages before the English Wikipedia: Hungarian and Hebrew (LTR and special charachtera) were the first. For Phase 2 these two languages as well as Italian will be the first to use/test. HenkvD (talk) 19:25, 9 March 2013 (UTC)
I humbly apologize (this is NOT ironical), you did it by the rulebook (which corresponds to my own unofficial software implementation guide). For the next phase I would recommend fi-wiki for there are many - critical - users whose opinion would be valuable and who are sometimes even too direct about their opions. Just think about it.--Nedergard (talk) 17:29, 15 March 2013 (UTC)

Wikidata must use the users language[edit]

Wikidata always uses English when a non-registered user (an “IP”) clicks the “Edit links” button. I think this is a real problem now that all Wikipedia languages use Wikidata. Not all users understand the difference between the projects. They don't know they leave their local Wikipedia project. They don't understand why the language switches do English.

Commons uses this script but to be honest the code is bad. You should write your own script or use one of the existing extensions that use a cookie to store the currently selected language. --TMg 00:47, 8 March 2013 (UTC)

We've got a feature nearly finished that allows the adding of language links on the Wikipedia. So the user wouldn't even need to go to Wikidata. Will take a bit to polish it and roll it out though. --Lydia Pintscher (WMDE) (talk) 10:56, 8 March 2013 (UTC)
Yes, this would help a lot. But even with this "widget" (I think we called it "widget" earlier) there are still cases were an unregistered user ends on an English page he does not understand. This will become a lot more importand with phase 2, I think. I saw multiple extensions here at Meta that try to solve this issue. Why not review and use one of these? A "hackish" first step would be a very simple script:
  1. Check if the URL contains an uselang=. If yes, do nothing.
  2. Check the referer and redirect to the proper uselang=.
With this "hack" the page still turns to English when the user clicks a link. But it's a first step. --TMg 13:26, 8 March 2013 (UTC)

Page history mangled[edit]

Hi. Due to bugzilla:45823, I accidentally mangled the page history of Wikidata yesterday when trying to import w:en:Wikipedia:Wikidata. I've made a note to clean it up one day. --MZMcBride (talk) 03:13, 8 March 2013 (UTC)

I tried to disable translation and delete it, but the translate extension wants to delete all the language subpages. What's the best way to proceed? PiRSquared17 (talk) 03:39, 25 December 2013 (UTC)
mw:Help:Extension:Translate/Page_translation_administration, "Removal from translation". Why not revdel though? --Nemo 12:44, 25 December 2013 (UTC)
Nemo bis, I did remove it from translation [6], but it still wanted to delete all the translation pages. Is revdel better? "Deleted revisions and events will still appear in the page history and logs, but parts of their content will be inaccessible to the public" I'd rather remove them. PiRSquared17 (talk) 13:15, 25 December 2013 (UTC)
I have removed it from translation, then after a while it also stopped to show me the "delete translation subpages" stuff when trying to delete it. I have now deleted it and restored the revisions not imported from enwiki (i.e. not no. 5302205-5302354). --MF-W 23:42, 25 December 2013 (UTC)
Hm. Thanks MF-W! I guess I wasn't patient enough. PiRSquared17 (talk) 01:12, 26 December 2013 (UTC)

Filling red links with Wikidata[edit]

Not my proposal, but someone had to document it on this wiki... Please help link and summarise the past discussions, rationales etc. (especially if you like or dislike the idea). --Nemo 07:51, 24 April 2014 (UTC)

That's a terribly awful idea: "Instead of making bots create millions of entries with infoboxes and standard sentences filled by some data source, as was done for many years (since 2002) for municipalities, asteroids, forms of life etc., we could transparently show similar placeholder articles with data provided by Wikidata. A topic, when covered on Wikidata, would automatically be summarised on virtually all Wikipedias."
Someone actually did something like that on occitan Wikipedia, that is what it looked like:
This affected almost half of occitan Wikipedias articles (!) in 2013.
There was a discussion recently on German Wikipedia about redlink-pages and how to integrate WikiData. See:
The "italian model" for redlink pages with Wdsearch.js is what i find most interesting. See:
"Risultati da Wikidata", really nice. Can we have this without Wdsearch on Wikipedia searchresults too? See also related discussion at
--Atlasowa (talk) 22:09, 27 April 2014 (UTC)
Of course I agree the it.wiki model via wdsearch is interesting, as it was originally proposed by me. ;-) I certainly think at some point it should be default in all Wikimedia projects, possibly via CirrusSearch; that will be easier if more wikis get to enable wdsearc by default. Atlasowa, can you get it enabled on some German wikis? Can that be helped in some way, is it blocked on that silly bug?
However that doesn't provide human readable localised information on the topic on site, except the Wikidata labels: it just sends you somewhere else, which is very useful but won't stop the eager botpedia mass-creators. :) I believe it could be possible to do better than oc.wiki! --Nemo 17:11, 5 May 2014 (UTC)
Hi Nemo, of course we can do better than oc.wiki: Let the articles be created and written by humans with an actual interest in the subject, humans that take pride in giving this new article their best shot. :-) I don't believe in botpedia providing "human readable localised information on the topic on site" - that is awful and demotivating. Nobody likes to do cleanup after bots.
Back to the "italian redlink model": Let me explain what i would want it to be. Currently, on german Wikipedia and most other language versions, if a reader clicks on a redlink "salticoididae" he gets to the "salticoididae"-redlink-edit-page, where he can (a) try to create a new "salticoididae" article in freaky wikimarkup (a magnet for useless keyboard-test-text that will be speedy-deleted) or (b) search for "salticoididae" in special:search in this WP or in other WPs via links (good luck finding anything, making him search on several language WPs without success). The "italian redlink model" shouldn't fix (a), but it can fix (b) by providing on the "salticoididae"-redlink-edit-page the wdsearch "Risultati da Wikidata" for "salticoididae" with direct links to WP articles in different languages. So the reader gets direct links to read the relevant articles in other WP or reasonator. That's a huge improvement! And we wouldn't even need to change the redlinks in articles by adding templates! (What i don't want the "italian redlink model" to do on deWP is: Adding wdsearch to de:Special:search. Don't. It's overkill, lots of js on every search. Too unfocused. Opt-in for wdsearch is fine, but not default.) --Atlasowa (talk) 23:33, 8 May 2014 (UTC)
Hi Nemo, have you seen en:Template:Redwd, "which puts a Wikidata link next to red links for subjects which have no Wikipedia article in any language" [7]? See for example en:Ang Thong (disambiguation) and en:Robert Mitchell (engraver).
So some people are now adding wikidata/reasonator-links into article text, even though there is not a single Wikipedia article in any language for these...? The tail is now wagging the dog! (And making wikitext even less readable...) --Atlasowa (talk) 14:31, 14 December 2014 (UTC)

Disambiguation as perfect WikiData use case for language harmonization[edit]

Besides Info boxes, the Disambiguation pages are the most standardized formats in Wikipedia. I am always wondering why they differ - for the same term - in different languages. Of course, there are language specific disambiguations. Most most of them are names, abbreviations etc.