 We're going to talk about the Wikibase ecosystem indeed. We're going to talk about editable link data for everyone. It all started with Wikidata. Wikidata is Wikimedia's knowledge base, and it provides general structured data about the world to Wikipedia. So when you look at the Wikipedia article today, and you see the info box on the site with basic facts about the topic you're reading about, then that might come from Wikidata. It's a central place that stores all this data for Wikipedia in a language independent way so that all the Wikipedia language traditions can benefit from it. Wikidata has become rather successful. After seven years of existing now, we have around 12,500 people who edit at least five times a month. We describe almost 75 million things in the world. Like this could be this building, this could be Brussels, this could be a famous actor, this could be a movie, things like that. How does this look like in practice? When you go to Wikidata today, you might find pages like this one about Margaret Hamilton, and find out that she was an American NASA scientist and mathematician, that there are 41 different language versions of Wikipedia who have an article about her. If you didn't know that yet, she also was a human. But now you'll have machines who are also supposed to know this. For them, it's not quite so obvious, so we record stuff like that as well. If you go further down a bit, you will find things like that she was a developer on the Apollo space program, for example. And if you go even further, you will find links to a ton of other places on the internet, other websites, other catalogs, libraries, and so on, where you find more information about her. And I'll mention that all nearly 75 million times about many different things. So this is what it looked like at the beginning, right? You have Wikidata in the middle, and it provides data to these other Wikimedia projects like Wikipedia, Wikivoyage, Wikimedia Commons, and many others. And then with all these connections to these other databases, we build up a semantic web or part of the semantic web that connects all this information, both readable by humans, but also by machines. And this is then used by, for example, your personal digital assistant to answer your questions, or if you get a quick answer in your search results, that might be powered by that. Or your local library system might provide you with additional information about the author and their books, and then that might also come from there. Now, with Wikidata quietly gaining a lot of traction and attention, we had a problem because all this data is maintained by a community for the world. And we reached a point where lots and lots of people wanted to put more and more specialized data into Wikidata without necessarily being willing and able to maintain that over the long run. And we said Wikidata should be a very central place in the semantic web, but it shouldn't have all the data because that's not what the semantic web was about. It's about decentralizing and connecting and linking information. So lots of people approached us like, can't we run our own thing? Can't we run our own Wikidata? And that's what we made possible. So the software that's running Wikidata is called Wikibase. And over the last two years, we have put a lot of work into making it possible for other people, organizations, companies to run their own Wikibase with the idea that at some point, we have a picture that looks like this, which would be the Wikibase ecosystem, where all the data that we already have and all the data that we already connect to is further enriched and connected to many different individual Wikis that publish specialized data. So you could imagine, there's a famous German marmalade producer. They have lots of marmalades, gems, and chutneys. And they might have their own Wikibase about marmalade, gems, and chutneys, where they describe all the ingredients and how they connect to other varieties of the same marmalade, for example. And then if you're really interested in marmalades, you go to that Wiki and find all the information you could ever ask for. So since we've been doing this, there are quite a few installations out there already that do really cool stuff with Wikibase and connected to Wikidata and other Wikibase installations. This, for example, is Lingua Libra. Lingua Libra is a project to help people record pronunciations so that other people who don't speak their language can hear how a word is pronounced. And they structure their word lists and their connections to Wikidata through their own Wikibase instance, for example. Or we have FACRAT, which is a database that collects data about the humanities and they started out with data about the Illuminati. Yes, this is a real thing. So they recorded things like all these anonymous people wrote to other anonymous people. So who was behind this? When did they write to each other? What did they write about? And you can, once you have all that data connected and queryable, it becomes really interesting to ask questions about the Illuminati, for example. By now they expand it into other humanities data as well, but that's how they started. Or you have RISO, which is a Museum for Digital Born Art. You might remember some, at the beginning of the web, and as it grew, artists have played with, for example, technical clutches in browsers, exploiting bugs and so on to create art. And that art is preserved by this organization called RISO, who is now cataloging it in their own Wikibase instance to make it easier to find, to figure out which browser version and which operating system you need to use to actually be able to see what the artist had intended and so on. Or we have a Wikibase instance called personaldata.io, and they are trying to do something about surveillance capitalism, by helping people have conversations about it, by exposing how personal data is used, by connecting organizations that fight against and provide alternatives. Very worthy goal, in my opinion. And then to sum up the few showcases, we're currently working with a lot of national libraries, for example, the German National Library, the French National Library, as well as some library groups on trialing Wikibase as a system for their own internal cataloging systems, or as documentation to that. Now, you might think why are all these people using Wikibase, what makes it special? And there's many reasons for that, but let me highlight some. The first one is that it's actually editable link data, right? Like Wikipedia, you can enable your editors, your users to change your data, to contribute to it, to augment it. And a lot of the existing link data that is out there is read only, and this is a way to change that and give people the power to support your cause or your organization in contributing to it through additional data, to corrections and so on. Then, thanks to its foundation in Wikimedia, it has community at the core. So there are many features in Wikibase already that help you build a community and help that community organize. That is things like talk pages, user rights, and much more, which existing library systems, for example, have a very hard time providing. The other thing that's really important for people is multilinguality. So the screenshots I showed you were all in English, but with a click of a button, I can show you the very same data in German and French or in any other language that you would like and that someone has contributed translations for. But you're working on the same data. And another thing that is really intriguing people is that Wikibase is built with the idea in mind that the world is very complex and that we want to capture that complexity in data. And that is very hard. But I think we've gotten quite far in making that possible. So the world is weird. For example, at some point, Sweden decided, it would be really nice to have a February 3rd, 30s in just one year. Because why not? So the rule that February only ever has 29 or less days, nope, the Swedes broke that rule. Or there was this woman who decided to marry the Eiffel Tower. The world is complex. And these complexities are exactly what Wikipedians want to write about, of course, right? Because other people want to read about it. So Wikibase was built to deal with that and to be able to express that flexibility and complexity that the world has. And the last thing I want to mention is, this is the connection to Wikidata. If you are running a Wikibase instance, then it's very easy to connect that to Wikidata and benefit from a lot of the work that has been done there already. In the future, we want to make it very easy, for example, to take over all the decisions that the Wikidata community has already made about how to model the world. Like the fact that we want a human to have a date of birth and a place of birth, and that is something we want to express, that should be very easy to take over to your own Wikibase instance. And the other thing is that if organizations, companies and so on are using data from Wikidata, then we want to make it very easy for them to also jump to your Wikibase instance in the Wikibase ecosystem. And with that, I hope I gave you a glimpse into the Wikibase ecosystem and made you excited to try it and join us. Thank you. Can you decide, or not you, but how someone can decide whether their data or the idea of what they want to put in the open belongs to Wikidata or needs its own Wikibase? So Wikidata has notability guidelines, like most Wikipedias have. They are very relaxed, but they are there. And they basically say anything that has a Wikipedia article can have an entry in Wikidata. Anything that is described by a reliable source out there, a library system, for example, can probably have an entry in Wikidata. And anything that is needed to fulfill a structural need. So let's say you have a famous person. Their child was not famous, but the child of that child was famous. Then you can create an entry for the person in the middle to complete the chain, basically. That is a rough guideline. And then there's a question of how much information do we collect about each of those things? And for example, as I was saying earlier, the Marmalade Museum is probably too much for Wikidata itself and probably deserves its own Wikibase instance. Also, because you can build up a community of specialists who really care about this particular topic and have a lot of experience about it and know how to model the difference between a chutney and a Marmalade and a jam, which I do not. Yeah, so, and if you're really unsure on Wikidata itself, there's a discussion page where you can bring it up and get some input from the editors. How the collaboration between Wikimedia and the libraries works. So they have, the ones we're working with, they have dedicated people inside the library who are tasked with evaluating Wikibase for their particular use case. And then we regularly exchange with them, we have meetings with them to talk about what they want from Wikibase, if that's already possible with it, if we need to make changes or if Wikibase for that particular case just isn't the right thing to use. And then over several months, we go through that evaluation phase with them and support them in that. Interesting how these links between different Wikibases are, is it just web links or is it some more advanced kind of federation where one would pull information and for another one this kind of thing? Right, so right now it's not just web links, but it's links and links in RDF, for example, that you can follow as you do in the semantic web. We are working on more advanced federation features. We're starting now in two weeks or so to start building more on top of that so that you can take more advantage than you currently can of this distributedness of the whole ecosystem. Yes. So exactly because of things like that, we have built Wikibase, not around the idea of there's one truth out there that we have to model. But you can have, for example, conflicting data. You can have things like China claims that Taiwan belongs to China and is a part of it while Taiwan itself says no. And you can have those statements next to each other and then you have something called qualifiers where you can qualify this. For example, saying that this particular statement is supported by the United Nations or you can say that this is based on a particular treaty, for example, or that something was true only from one point in time to another in the past. So there's a lot of flexibility in that. It's of course not able to express all the complexity in the world that you can think of, but it gets quite a lot farther than typical databases. Right, so what we're starting to work on now is the ability to use Wikidata's properties and items in your own Wikibase instance so that everyone would be able to use the same vocabulary ideally. How that plays out, we have to see, but I have high hopes, yeah. So that is the idea for now. And of course there's a lot of links and same S links as understood in the Semantic Web so that you can say, okay, if they are talking about this concept, that's the same thing they are talking about and then you can follow those. Yes, so media. Wikipedia has that too, yeah. Maybe Faden can help you understand more of that but there is a change of stream that you can follow both looking at a website but also via APIs. Sorry, I have a hard time understanding you. So the initial storing and editing of the data happens in MediaWiki and then for querying and answering more complex questions on that data, we have a Sparkle endpoint and we use PlaceGraph for that. PlaceGraph, yes. Is there a roadmap for those new features about the... Yes, so we want to start in the next two weeks and then roll out the first version of that, hopefully in very few months, depending on how it goes and then iterate on that based on what people tell us they need in addition. So the first version will be very simple. You will be able to use only the properties that Wikipedia provides like date of birth, place of birth, links to other databases and so on. And then we'll build on that. The way of dealing with data often comes with some kind of licensing. I guess the most of Wikipedia's base is Creative Commons. So Wikidata itself. It's a different licensing, how do you deal with that? Right, so Wikidata itself is CC0. So only data that is licensed to CC0 is supposed to go into Wikidata, which is one of the reasons why organizations might decide to run their own Wikibasins because they can't release their data under CC0. And then it would be up to the consumer to decide because they see where the data is coming from and that Wikidata has a license declaration and they would then need to see, depending on how they use the data, comply with the license. More questions. Thank you very much. Thank you.