We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

Generating virtual 3D exhibitions from Wikipedia

00:00

Formal Metadata

Title
Generating virtual 3D exhibitions from Wikipedia
Subtitle
Let's dive into some rabbit holes with The Everything Exhibition!
Title of Series
Number of Parts
287
Author
License
CC Attribution 2.0 Belgium:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
Abstract
Wikipedia and Wikimedia Commons are a treasure trove of free information about a lot of topics! Wikipedia alone contains over 50 million articles in almost 300 languages. We're building a new, fun way to explore this content: "The Everything Exhibition" is an open-source generator that creates interactive, virtual 3D exhibitions on arbitrary topics, which you can explore in the browser! It has a multiplayer mode, so you can visit the exhibitions in groups. And, other than in a real museum, you can scribble on everything! In this talk, we want to introduce you to how the generator works, and show you our current progress. We're using the 3D rendering engine three.js, and defined a simple datastructure for describing the exhibition content, to make the generator arbitrarily extendable. We'll go into which algorithms we tried for creating the room layouts, share what we learned in our playtest sessions, and talk about what's next.
Projective planeFreewareFile formatOpen setWebsiteDemo (music)Interactive televisionForm (programming)Source codeNetwork topologyState of matterInternetworkingDiagramMeeting/Interview
Demo (music)File formatComputer animation
Digital photographyMereologyWeb browserFormal languageTouchscreenDescriptive statisticsGraph coloringDrop (liquid)Content (media)Logic synthesisDifferent (Kate Ryan album)Video gameOcean currentComputer animation
First-person shooterSource codeOrganic computingGroup action2 (number)Data structureDifferent (Kate Ryan album)Graph coloringMereologyInformationCircleLink (knot theory)Execution unitConstructor (object-oriented programming)BitoutputCASE <Informatik>Presentation of a groupArrow of timeMedical imagingPoint (geometry)HierarchyPerfect groupVirtual realityComputer animation
Open setSelf-organizationMultiplication signAudio file formatProjective planeQuicksortOpen sourcePrototypeSource codeSign (mathematics)Product (business)Position operatorContent (media)SpacetimeHypermediaMereologyArithmetic progressionMedical imagingPlanningBitSoftware developerData structureKey (cryptography)AreaFood energyWeb pageLibrary (computing)Open setWikiPlug-in (computing)UsabilityWeb browserOcean currentComputer animation
Integrated development environmentData managementIntegrated development environmentPlug-in (computing)WebsiteProjective planeDifferent (Kate Ryan album)Content (media)Type theoryHypermediaBlogDirection (geometry)SpacetimeObject (grammar)Ocean currentAsynchronous Transfer ModeGraph coloringUniverse (mathematics)Connected spaceWikiMultiplication signGroup actionTwitterBitWordRecursionMereologyCASE <Informatik>Square numberRevision controlSocial classSource codePlanningInformationMusical ensembleNumberLink (knot theory)DemosceneQuicksortPhysical systemSound effectFingerprintTrailNormal (geometry)State of matterKey (cryptography)Point (geometry)Computer animation
Content (media)MereologyBitWeb pageProjective planeKey (cryptography)MathematicsEndliche ModelltheorieText editorComputer animationMeeting/Interview
Endliche ModelltheorieSource codePublic domainHypermediaBitSheaf (mathematics)Multiplication signMeeting/Interview
BitContent (media)Endliche ModelltheorieTerm (mathematics)Right angleWeb pageTouchscreenClassical physicsMusical ensembleWebsiteType theoryMeeting/Interview
Type theoryAbsolute valueContent (media)Web pageCompact CassetteMeeting/Interview
Type theoryGame controllerView (database)Three-dimensional spaceLibrary (computing)FreewareSet (mathematics)Web pageBeat (acoustics)QuicksortSpacetimeMeeting/Interview
Event horizonMereologyWeb pageQuicksortBitSource codeMeeting/Interview
Right angleFrequencyQuicksortMultiplication signMeeting/Interview
Peer-to-peerWebsiteServer (computing)Client (computing)Library (computing)Moment (mathematics)TelecommunicationService (economics)Meeting/Interview
HypermediaGroup actionContext awarenessComplex (psychology)Point (geometry)Coma BerenicesQuicksortMeeting/Interview
Video gameGame theoryWeightMeeting/Interview
Meeting/InterviewComputer animation
Transcript: English(auto-generated)
Hello everyone, we are Bleeptrack and BINUI and we want to show you a project we are currently working on. We call it the Everything Exhibition and it's motivated by the following idea.
In the internet, there's like a huge amount of cool open data and free licenses. Wikipedia comes to mind, Wikimedia Commons comes to mind and a lot of other projects. And we were thinking that maybe for some people like, for example, younger people like kids,
it's sometimes not as accessible or as interesting to look at these sources, right? They might not be motivated to read through a long Wikipedia article, for example. And we were thinking maybe there would be another format they would be more excited about to look at and study and learn from.
So what we're working on is a project where you have a website where you enter a topic you're interested in learning about and then it generates a 3D exhibition around this topic, sourcing from open data sources. And then you can walk inside, look around
and interact with it, which might be a format which younger people are more used to these days. And we will just give you a demo of how that currently looks like and what you can do with it. All right, so I'm now hopping on into the browser so we can do a little live thing.
So the first thing you can see is this screen, you will end up on something like that. You can enter your nickname on the top and choose a color and now to the most important part, you can choose what topic you want to browse on. And there's this little dropdown and we'll also give you some suggestions when you start typing
and you can search in different departments basically. So currently we are connected to Wikipedia and all the different languages of Wikipedia. So it's already multilingual. We're currently in the English Wikipedia here. And when you press generate, you will end up in this cool looking room or more or less outside of the room. So we are in photosynthesis
and let's take our first few steps in here. I think you will end up in a quite big room. You will always end up in an entrance room at the beginning. This is all procedurally generated from the content you can find in Wikipedia currently. So yeah, maybe peek into the first room.
We have the introduction here and you can see it looks like a little museum. You have pictures on the wall with descriptions of the pictures, whoops, little graphics and yeah, quite some text and information. You can just scroll through and find the stuff that you are most interested in.
Yeah, the setup is hierarchical. So we are aligning as to the structure in the Wikipedia article and the thing that I already showed you a bit is that you can also fly. So let's take a bit of a better look on this whole construct that is generated. You can see all these different rooms
and everything of the hierarchical structures. So for example, let's hop in here. You can see that there are like sub rooms about water, photolysis for example. Yeah, let's take a peek in here. All the information and cool images.
All right, there's more features and I guess I already peeked a bit. So if someone is hiding in here, that's Minnie and we implemented a little multiplayer part. Maybe you already guessed that because you can choose your nickname and the color and also the input on the left side of your nickname
is a little ASCII face you can set for yourself if you want to have a bit of a different expression and yeah, the idea is that you can also, hello, that you can also explore in different groups and what we found out is when you hang out in groups,
you sometimes like want to point somewhere or you want to show something to another person and then you can do what Blue is doing. You can draw like little arrows or point these and I can also do this with my mouse and also for example, draw a little arrow here. That's a very fun thing you can't do in a museum.
Like you can like draw on all the paintings on all the walls and like circle in all the important information you want to highlight. And another thing you also can't do in a museum is you can hop from museum to museum, of course, because we are here in a virtual environment. I can just grab, yeah, an interesting text part
that I want to explore more. For example, I want to learn more of LGs so I can press the link and it will need a few seconds and I will hop into the next cool exhibition where I can start exploring more information
and finding more interesting stuff. And yeah, sometimes maybe that was more of a dead end and I just want to go back where I came from and in this case, there's always presented this back to link where I can just hop back. Is there a thing that I miss Blinry that I might should explain maybe?
I think that's the features we have. I think you got it. All right, perfect. Then let's hop back to the presentation. This project is supported by a really cool thing called the Prototype Fund, which is an organization funding open source projects
in Germany supported by both the Federal Ministry of Education Research in Germany and the Open Knowledge Foundation. And you can apply to them. And if they pick you, they will basically give you money to develop open source projects for half a year. And we are currently like two thirds of the way through the funding.
We have some more time to go, but like this is very much an in progress thing for us. We wanted to talk a little bit about the technologies we use. It was really important for us that this project runs in the browser. So it's as easily accessible to people as possible.
So we picked a library called three.js for making the 3D graphics happen. And yeah, that's really easy to get started with. You have like really short loading times for the page. That's awesome. We are currently using the Media Wiki API
to get the content, which is like, it really surprised us how many projects we can access now. It's not only Wikipedia, but it's also like Wikimedia Commons. You have a lot of like images, there are image galleries and audio files.
And we can use other projects like, WikiCrote, for example, or Wiktionary. And then all of these sources you can like enter in this 3D space. If you have any like Media Wiki lying around and you want to explore it, it should be possible to do that as well. And our idea is also that there could be more data sources
next to Media Wiki APIs. So like our current plan, that's something we're currently building, right, Bina? But maybe it should be possible that you just generate any markdown content in the end, and then our exhibition generator understands that
and you can explore it that way. So if you have any like unused sources of data you would like to see in the space, you can write a plugin. For the multiplayer feature, we're using a library called Yjs, which makes it really easy to have a common data structure,
which is shared among all players. So for example, the player positions and the things you draw on the walls, for example, are shared via this library. Yeah, we were thinking quite a lot about the layouting part of the exhibition. We're just scrolling through all the notes we made
in the past months. And we looked at like real museums and what their floor plans are. We started with these recursive algorithms, which like assemble little squares together and then started refining that over time and trying other things. We had this, it was also a bit of a dead end, I think,
where we tried to like make a path through all the rooms, which like gives you a little tour of all the content basically, inspired by Hilbert curve, as you can see here. But when we implemented that and tried that, it didn't really feel so,
you didn't have as much overview over the content, I would say. So we mostly stick to the recursive structure for now. Next steps in this project, probably like, yeah, making the environment a bit prettier. Currently it's like all the walls have the same color
or you don't have as many like objects as you would have in real museums. Maybe we will add little plans into the rooms or some benches or some other environment art around the scene. And yeah, we would also really like to have more sources for the information which we can put in.
And here are some use cases which are in our minds. In schools, teachers could use this project like to give their class a more fun introduction to a topic. They could all like go into this group, use the multiplayer mode to explore some topic together,
maybe walk into different directions of a room and then come together in the end and talk about what they found. Yeah, also ourselves, we really like to explore rabbit holes you probably have been, you fought up to this trap to like go into some new Wikipedia topic
and then just get lost in it forever. And in this 3D space, you can like literally get lost. That's a fun thing for us. And when we tried this like a month ago with a group, someone came up with the idea to do speed runs, like going from one topic to another
with a fewest number of hops. And that might be a fun thing you can play in the space. And yeah, also just in general groups who would like to explore a topic together, they can just meet in the space and look at all the stuff that's there.
Yeah, now also you can take action if you want. So if you like our project, please spread the word. So maybe you have connections to schools or universities or you are still at a university, yeah. Establish connections or tell about the project and let them try it. Also, if you are trying it or others are trying it,
please give us feedback. This would be extremely helpful for us to improve on our project. And if you have your own interesting content sources, maybe it's your blog or your website or your hack space has a media wiki or has a very different type of content management system, you can write your own plugin. That would be super awesome for us
to see what cool spaces we can explore. If you want to try the Everything exhibition yourself, there's a link below, blingree.org slash everything minus exhibition. So yeah, that's a deployed version. Just go ahead and try it, have fun. And if you want to reach out to us, you can find us on, yeah, definitely Twitter
and also different social media websites under our nicknames, BLEEPTRACK and blingree or email us at bb at bleeptrack.org. That would be awesome. And now I think there's a nice amount of time for the Q&A session. We'd love to see you there. Thanks for watching. Thanks, bye.
Questions from the chat and would be happy to answer. Feel free to add more as we go.
For example, people asked about the, like what's next for the project and we already talked about several of the things we have in mind for this project. I think one of the parts we still want to add is a bit more interactivity, I guess, maybe like being able to equip different tools
you can use to interact with the content. Like we were sometimes thinking about having this edit gun, right? Which you can point at pieces of text and as you click on it, you will get transferred to the edit page in the corresponding wiki, which would help like to actually make the content that's available better as you go and as you maybe discover some errors,
things you want to change. Yeah, I think a question that fits great to this topic is about displaying 3D models from comments, right? This is something that we are actually currently experimenting a bit with, not only taking them from comments,
but also we are searching a bit from other sources. Like what did we try last time, Sketchfab, I guess? There's also a lot of public domain 3D models. So this would be absolutely interesting to us to show more 3D stuff, of course, yeah. The media comments, I think in the very beginning when we were surveying what's actually there
in terms of content, we didn't find as many models, I think. Yeah, right. And it was a bit hard to associate them to like specific topics. What we did was enabling audio content. So if you visit like an exhibition around a topic of classical music or something,
sometimes you will hear little snippets of stuff going on. When we were developing this, we tried with the page for the help screen. Yeah, right. We've got very annoyed by those sounds. But yeah, that works quite well. I think, yeah, sorry. Absolutely, if anyone has like ideas
for other types of content we could embed into the exhibition, definitely let us know. Yeah, absolutely. On the page we linked and on the talk site, you will find ways to contact us. One question was about we are accessing the museums in VR and I guess maybe that's something you can also talk about a bit, right? I have a friend who owns Oculus headset
and I very briefly when I visited tried to enable that because we are using this free JS library. It seemed surprisingly easy to do that. You especially like give embed a new type of camera into the 3D space and then this gets attached to where your headset is basically
and where your handheld controllers are if you have them. So that shouldn't be a huge step. We haven't prioritized that because we think not so many people have access to those VR headsets, I guess. And it should be able to just do it in the browser or in a like 2D view kind of setting.
But thanks for the request again. That would be a worthwhile thing to do. Yeah. I think another question was about, yeah, giving pages some sort of chronological order in a sense that you can like skip through historical events.
I think what also would fit into that topic is about like overall like sorting or prioritizing certain parts of the museum. Maybe a bit difficult because you would need the data from some sources, right? You would need to like collect behavioral data
when people explore the museum maybe to know which topics they like. Yeah, having a chronological timeline is already kind of happening if you visit like historical themes and you will have like certain sub rooms explaining different centuries of some certain period of time, I guess.
Yeah, right. Someone else asked about like the multiplayer technology whether it's a server or whether it's peer-to-peer. We are super happy that this does not require a server. It's like a peer-to-peer communication thing going on with the wide JS library.
And it just needs a tiny signaling server, which like there are public servers for that which allows clients to find each other. And after that they do everything peer-to-peer. And that's really convenient for us because then we don't need to host anything. Except probably the website, the setting website. I guess, yeah. Absolutely, yeah.
Yeah, I'm not sure if you talk about that in the talk so much, but also we currently are trying to do some group feature basically so that if you are, for example, in a teaching context in school that you can create like closed groups so that you can like walk through an exhibition together but don't have like random people joining in and out.
And I also don't know if we mentioned that in the talk, but at this point, the generator supports any arbitrary media vikis. So if you want to go to like a fandom viki or want a certain topic, I think it's not called vikia.com. You can absolutely do that. You can visit like a NetHack exhibition
or a Game of Life exhibition if there are vikis around that topic. Right, and I guess the Q&A is coming to an end now. Yeah, please contact us and if you have questions and have a great Fosten. Enjoy, bye. Bye.