 So we are live on air for a very special episode of the Illinois on Air podcast. It's a podcast on organizational learning and knowledge work and it's a very special episode because it's live streamed from the first center center, which is the podcasting assembly of the remote chaos experience of the Chaos Computer Club. And it's also special because I have two very special guests. One is Claudia or Ching's. Hi, Claudia. Hello. And the other is David from TCM, I think in France. Hi, David. Welcome. Exactly. Hi, everybody. There are a few specialties in this podcast. First is that it's a virtual one. Normally, at the center center, we have a physical stage where we podcast. We would sit there on chairs and have headsets on. But now we do it remotely. This means that you can see the live stream that is brought to you by the very fabulous C3 Walk that's streaming. You also have a chat on the IRC. There's this on hackint.org, a channel called RC3-SendItCentrum. You can ask questions there. We do a Q&A at the end. And as we are a podcasting assembly, you will be also able to lifestyle in with a phone number. We will put the telephone number in the chat later on and then you can come up to the show and ask questions to us to Claudia or to David. Another thing because this podcast is special is because we have this very special guest. We want to talk about today about a tool that enables people to meet in the virtual space. I think due to COVID-19 pandemic, everybody was used to hang in video conferences like Big Blue Button and Zoom and so on. I see Claudia is nodding and David as well. But perhaps you experience that when you're in these video conferences, there's something missing if you compare it to a live conference or live event. People having a coffee together. You just stumble upon each other. You have a chat just by chance, a serendipity. You meet each other. You meet new people. And that's not happening in a video conference, but it's happening in tools that we call spatial chat tools, like tools where you can chat and be in the room somehow. And we have these two people here because they have a lot of experience, a lot of in terms of a few months or weeks since we are in this pandemic situation. And perhaps first I switched to David. I read a little bit in the internet about you. Come from a company called TTM, the coding machine. You do stuff with open source web development. A lot of open repositories on GitHub, to all this web stuff. And you're based in Paris, I think, in France. Yes, exactly. I'm actually the CTO of the coding machine. So we run 90 company. I co-founded the company like 15 years ago. So it's it's been quite a while. And we've been doing a web web development, actually, a lot of PHP. I'm a really big fan of PHP. I've been doing a lot of open source libraries and frameworks around it. I've even participated in a few standards regarding interoperability of PHP frameworks. So this is where I come from. And I'm actually really kind of a nerd. I've been starting coding like when I was seven years old. And I've always had, well, I've been interested into developing video games when I was young. So web video games and you see where this leads us. This is what we will talk about later on. Very good. OK. Yeah, thank you. And then we have Jinx from the Chaos Computer Club. I read also a little bit about you. As you said, you were part of the main organization team of this whole chaos, which normally is a conference last year took place in Leipzig at the Messe. At the fair there with, I think, around 17,000 people. You wrote about yourself that you're an author, a podcast, of course, and also a specialist in data protection. So what else do we need to know about you for this podcast today? Yeah, this year I'm first time in the main organization of this. RC3, remote case experience. We sat in the beginning when we started to develop the ideas for the whole conference that we can't, we just can't take the Congress as we know it and put it just one-to-one into the digital world. This just can't happen. And the thing that we always had in mind when we started writing that concept, which was a colleague, which was Knud and me who wrote the concept for the whole case experience. And we always had in mind something like a four-day LAN party, a local area network party as we did in our students' times, where you get out after four days with a whole pizza overflow and square eyes and total sleep deprivation and come home and say, OK, that was really great. So when we started this concept, we tried to see it from another perspective. So what can we actually make happen and what can't we do? So this was a really interesting thing because what I did in the last years was mostly helping with assemblies, with the assembly team at the Congress or at the camp with the village team or I was with the privacy week, which is also a lot smaller conference, but also a physical one that's over seven days. So I know a bit about all those things, all those conferences and how they're working in the physical space. And yeah, interesting part was how to make something happen in the digital space. That's also great. And I think we have to talk later a little bit about what this AC3 world is and what the components are. But perhaps first of all, let's talk a little bit about the tool like you called it, David WorkAdventure. WorkAdventure, the description of the GitHub repo is one of more than 300 that you have is it's a collaborative web application in Brackett's virtual office presented as 16-bit RPG video game. As I saw, it was started in late March, has more than 1200 commits at the moment and you released a version one like 10 or 11 days ago with about 12 contributors. So tell us a little bit about the emergence of that tool. Like how did the idea come to build that tool? How did you do it? What are the technical components? What is WorkAdventure? As you said, it started in March and well, I'm in France. And in France, in March, we had the lockdown. And during two months, we could not go out. We could not see our colleagues. It was quite stressful, actually. And so at the coding machine, we started to do a brainstorming about what could we do to improve the situation with our skill set, so as web developers. So we started thinking about could we do a phone application to detect nearby people that have COVID? Well, this has been done by other people. So well, we would certainly not do that. So we were really looking for an idea. And at the same time, I started doing well, some conferences, as I told you, I'm mostly into PHP and I've been developing a library which allows to do some GraphQL in PHP, whatever. And I was doing a live conference about that online. And during this conference, I was presenting. Of course, I was not seeing the audience in front of me. And at the end of the audience, at the end of the conference, well, I could not see anybody. I could not talk to anybody. And it was really like, wait, how many people were there? Was there happy? I did not have any feedback. And I needed to find a way to solve that. So you just looked at the dead eye of the webcam, so to say. And after the talk, it was just over. That's it. Exactly. And basically, I wanted to have a way to have a conversation with a subset of people. Of course, it's impossible to have a conversation with 100 people at the same time. But I wanted to have a conversation with a few people and a few weeks, maybe a few days later, I started to say, okay, this would be really useful also for my office because I want to see my colleagues I'm spending all my days scheduling new meetings. And I'm spending like maybe half my days sending links to speak with people by creating whatever Zoom conferences, Google Meet, Jitsi Meet. And basically, the idea was born this way. I started thinking, okay, what if I could basically walk through my office as if it was a video game? I wanted to come next to somebody and start being able to start speaking to it like that without having to schedule anything, without having to plan a meeting. Just, I'm moving towards someone and I can speak to it in an informal way. I wanted to get rid of the formal I'm sending you an invite, we are going to put something in the calendar, whatever. And this is how WorkAdventure was born. And I started speaking about this idea to colleagues and they were immediately like, of course, it's a good idea. And so we started working on it, well, mostly at night. It was like- So beside your day work, it's not you do it as your day work? Yes, at the very beginning, it was really beside my day work. It was with a few colleagues of mine, we were three or four. And we started hacking a bunch of technologies together and well, basically, I had been playing with all the technologies needed in the past. But, well, it was just a matter of putting them together in the right way. As a child, I had been playing a lot with game engines. So developing a video game was one of my favorite occupation when I was young. So I was really more than happy to start a game. And then we needed the video chat part. And we had been working a few weeks before on a project which was using WebRTC, which is basically the API that is used by browsers to speak together and to allow video communication between browsers. And so, yes, it started like that. Yeah, cool, cool. And Claudette, did you see this effect in the RCC world? Like stumbling across people, meeting people just walking through this world and so on, in contrast to just having video conferences like Zoom or Chitty or something like that? Actually, yes. And this is the reason why we chose to implement it. We had a look at a whole bunch of different solutions, also like Mozilla Hubs and so. And most of them had the problem that they were really, really resource hungry. And we said, okay, we need something that people can also use with a bad bandwidth with an old computer or with not the latest gaming hardware. And so we chose the first thing was, okay, no 3D because when people are sitting at home or probably with their families somewhere in the wild with four embed or something, they just can't participate. We can't do that. We need something that's smaller, that's better to handle also with small bandwidth or all the hardware and so on. So if you're at your parents' house currently with the four embed bandwidth connection, then you can still participate, okay. That was the first thing. And then we said, okay, 2D it is. And then we had a look into several solutions. And our technical guys came down to, okay, we'll look deeper into that work-adventure solution. They were debating on implementing themselves again, but just building from what's already there. And then, yeah, it is 2020 and chaos is chaos. And yeah, building from scratch was not a solution at all at some point. And then we came down to, okay, we just take that work-adventure stuff and build from there because it was also, and this is second part, pretty easy to build the tiles for the 2D world, the graphical pixel tiles. And we had the idea, okay, if we hand out a how-to for those graphical pixel tiles, then all the assemblies, all the people who want to participate will have something to do over the next weeks to get enthusiastic about the event. And people can build something and can participate already in the build-up that took weeks. And it was so cool to see how that actually worked out. Yes, so that's why we chose work-adventure then. Yeah, yeah. Yeah, that's also the interesting thing. As you said, I would like this, a lot of components that came together. I think you used the gaming engine called Phaser Eye. Oh, I've already read it, right? Yes, exactly. Then you used an existing editor tile where you can build these maps and create layers and add smart features like having websites pop up if you go somewhere or starting cheesy conferences if you need to have a touch-head with more than four people and so on. So how came all these things together? Were just technologies that you worked with in the past and just brought that together? Or was there also research necessary of what components to use for the whole system? Actually, I've been testing two technologies for the front-end. I was used to using game engines, but not in the browser. So this was quite new for me. And I started with an open-source library called Pixi, which can be used to render graphics on-screen in a convasse. And it was quite cool, but it lacked a few features, especially the ability to load a map from Tiled, which is the tile editor that you've been using to build the maps. And since there was this super tool called Tiled, and I wanted to be able to load the maps easily, I realized that there was this other game engine named Phaser that can load the tiles, the map quite easily from Tiled. And so I used that. I followed a few tutorials, and it was, I won't say easy, but relatively easy to set up. And so we sticked with Phaser, which is really quite a nice tool, actually. And, yeah. Sorry. One fun thing about Tiled is that John, which is a guy who wrote Tiled, just connected on WorkAdventure like a week ago, and came say, hi, I made Tiled and you're using it. And I was, woo, super. And this was also the bridge, how I came to work adventure. I used it the last six months, a lot of tools like SoCoCo or Rimo, or Gatha or Wanda. And I think Gatha is also a supporter of Tiled because he is also Tiled as an editor. And then I read somewhere there in the forum, there's also this WorkAdventure thing. And what I found interesting that you just put the whole source code on GitHub, and you have a description, like running a Compost.up and have a Docker container and have the whole thing running. I think very interesting because a lot of people can build stuff around, as you said, Clouder, like doing graphics, doing tilesets, there are a lot of repos, links to free repos with tiles and so on, but also build them from scratch. There was, I think, organized by Honky every week on Tuesdays or something like that, workshop on how to use Tiled and how to use Clita, which is a painting program to create own tiles. And so there are a lot of custom tiles in this RC3 world at the moment. It was a very wise decision, I think. I was so much surprised when I learned that you were giving workshop about how to build a map. And I had absolutely no idea. What? Because there's a few... So, thank you very much, actually. At the physical Congress, there are very typical things, like there is Seidenstraße, which is a postal system, and there are specific signs and unicorns and so on, and they were not available in the standard, like, game tile libraries. And I think also that people were really, really creative with the stuff they did. And the virtual Congress looks a lot like the physical one. You have a lot of, like, jokes and Easter eggs and so on. It's very cool. Yeah, so many creative people. And I heard about assemblies who build an underwater world somewhere, and I didn't find it yet. Or a Tetris room. I think it's at the Hexen. Oh, okay. I'll have to watch over there. So the Tetris room is, when you walk through or when you run through it, and you hit the tiles and the place, the Tetris melody. And other really cool stuff. People really got creative, which is the beautiful part of 2020 that people try to make the best of what we can actually have from, yeah, from all the things of sitting at home. Yeah. So, Préves, let's talk a little bit about the components of the R3-3 now. Like, you'd also already talked a little bit about the concept, like you wrote it. You had an idea of how many people might join such a thing. Typically, I think, David, you haven't been at a Congress so far. You have, last year, we were at the Leipzig fair. We had, I think, almost all of the space there. There were rooms for giving talks, like lecture halls, but also assemblies where the Hexenters can have tables and everything was with LEDs and stuff. A lot of projects, self-organized sessions and so on. So, Claudio, perhaps you talk a little bit about how this evolved from your first idea of a concept, like how you imagine, you can imagine how this might look like to what it is now that we have since yesterday. Yeah, as you just said, we have the physical Congress with about 17,000 people. And we came to the conclusion in about May, I think, we had monthly mumble meetings with a bigger auger group with some people from assemblies and from teams and the auger and so on. Lots of people coming together to geek ends what to do. And in about May, it was like, okay, no, we can't. We just can't do a physical event. This is never gonna happen. Okay, so what are we going to do? And in August, we said, okay, let's just do it completely online. And then we thought, okay, how many people could actually come? And there were so many people, actually, so many people also in the community who said, no, that can never, that can never happen. How are you, how can you bring the Congress to the internet? And as I said before, we never tried to bring the Congress to the internet, but we tried to do something like Congress-y feeling-y, but different. And we said, okay, so many people say, no, that can't work, okay, how much people will we probably have? Well, probably we will handle like 1500 tickets and 500 people will use it. And we will totally be happy. That escalated quickly. And our first batch of tickets ran out. Took a while because there was one of our mistakes in the build-up, let's say, or in the development of the whole event or world, the whole experience, that we built while the ideas were still spreading. So it was always a parallel thing. And it was not like, okay, this is the plan, then we build, then we do the, and no, but everything was- What do you call it, Agile? Yes. For some values of Agile. But yes. So we didn't communicate so good what people would need that ticket for. And we said, okay, but we would need a login area. So we have some private messages and something how people can enter that 2D world. And we still have a bit of an idea of what's actually happening. Yeah, so our tickets ran out and people were, oh my goodness, how can that happen? Yeah, well, as we learned- It's always enough room in the internet, like why are there tickets at all? Why are they limited and so on? Yeah, but this thing like the internet is scaling limitless. The same presumption like, yeah, my electricity comes out of the plug. Not working. And we had to rethink how we can scale the whole event. And people ran out into their teams and into the areas, okay, what can we do? What can we fix? Where can we, yeah, how can we scale it? And we had a certain round of tickets giving out. And yeah, now we're here with a whole bunch more than 15,000 users. With a lot of people. Yes, with really, really, really many people. I think it's an interesting aspect like scaling, David. I think I read somewhere when I first used WorkAdventure, you say, on the servers you use, you suggest to not be more than 100 people because then the frequency of updating where everybody is on the map gets changed and people run really fast and things like that. Yes, exactly. And this was the main issue for this Congress to scale that. So perhaps talk a little bit about your experiences, like what kind of events did you do so far? Is it mainly around 100 people or 500 people? What are the dimensions that you worked with so far? Yes, so far we've been, as I say, I'm a PHP developer. So I know pretty well the French PHP community. So when I started communicating around WorkAdventure and saying, hey, look at what we've been doing, I was noticed by people doing PHP in France. And so basically a number from the French PHP community PHP user group told me that they were having a problem organizing their big annual meetup, which is about 600 people. And they wanted to put it online. They did not know how to do it. And so I proposed to organize a part of it on WorkAdventure. And basically, well, I worked like during one month, just on this, we were a team of three people and WorkAdventure was like a prototype. And we had to make it something solid enough to host an event with about 600 people. Actually, we had about 100 simultaneous users, user at the same time during the conference. But so I was pretty confident that about, we could go a bit above this limit and like 300 or 400 people would be okay. Now the main problem is that this is developed using Node.js, which is basically in JavaScript, JavaScript, everybody will tell you, hey, it's super fast and so on because whatever. It's actually single threaded. So basically you can use only one core of your CPU and there is only one core that is doing all the computation. And when you run out of power, well, everything starts to get sluggish, people start to lose connections and everything goes bad. So I started working on stress tests to see how we can scale things up. And basically when we started talking together or when you came to us and asked us, how many people can we have? Yes, the limit was about 200, 300 people at the same time. Since last week it's much, much higher. And at the beginning of XTR, we're going to start really a new version that will really scale much more. But for RC3, you had guys to deal with a version that could not scale really well. And you did an excellent job at making this scale amazingly. So wow. I think it were a lot of pizzas and nights of coding and a lot of gray hairs and so on. But what is the main thing that drives the resource hungryness? The more people you have on the map, the more updates you have to send, where is everybody? Yes, actually, the interesting part is that it is only about the position of people moving. It's not about the video, because the video is going via web RTC or via GTC, so you can scale GTC, that's not part of work adventure. Or if you are speaking to someone, this is web RTC, which is actually doing a peer-to-peer connection. So the video is not going through the work adventure server at any point in time, so we don't have any problems scaling the video. However, sharing the position of players moving on the map is really challenging. And so we made a bunch of optimization regarding this problem. We started by trying to send the position only of people that are close to you. So if you've got a huge map, we're not sending the position of everybody on the map. And the problem is that if you've got 600 people on the map, when someone is moving or 600 people close together, when one people is moving, you have to send the signal 600 times to everybody. And if everybody is moving at the same time, a lot of messages. And they must be shared on a single server because there is a single server that must have the position of everybody to be able to know who is speaking to who and who can speak to who. Yeah, true, I understand. Yeah, so basically the challenge is here. So I think this will be interesting when I think the changes that CCC made to the source code will be published on a public repository as well to sort of learn from each other what the tips are. Yes, I'm dying to see what you've guys have been doing. One of the things we were talking about in the preparations was also that how can we actually manage if we have 1,000 people or more joining the world? So they drop into that lobby area. How can we manage to have 1,000 avatars in one room? We can't, that's not possible. And we said, okay, let's divide them and say we can have a lobby room and this can have a host like, I don't know, 50 avatars. And then everyone who's joining afterwards will come to another lobby room parallel to the first one. And then until the next lobby room is full with 50 and so on and so on. I think it's a bit more complex what our guys brought up, but this is the basic principle because we could not either technically, nor from a user perspective put, like, 1,000 avatars in one room. Even if we would simulate the glass wall or something, it doesn't work, it just doesn't work out. You don't see anything anymore. It's just taking too much of a space and so on. And that was also a pretty, yeah, interesting thing how to work all these things out. Also how to get so many people on so many maps because we have nearly 300 assemblies and lots of them were building maps. So actually you will find, I don't know the exact number, but if we save 300 assemblies, lots of them building maps, like 200 maps in the whole world. How do people get there? How can they interact? How can they probably link together? So you come from one assembly to the next and so on. This was a very interesting experience yesterday because I think the first two hours or so, a lot of people tried to get to the Send It, Send On podcasting assembly. And on the chat and on Twitter, you had all this discussion like from the lobby, you have to go up, up, up, then left, then there is a little thingy, then you have to go down. It ended up at someone like doing a video on how to get from point A to point B. You put it on Twitter. And stick that to our profile. So everybody was able to like get there. Until somebody found out that in the assembly list, you find the work adventure map as a room and you can directly click on it and you come to the map of the assembly sort of. But things that were, I won't say totally easy in the physical space. It was still tricky in this really large, large space in Leipzig to find a certain spot. But there was things like C3Nav navigation system and a map of the whole area. It was easier, I think. And people have to learn in this new normal to adopt to that. Like how do I navigate in this virtual space in the same way? You know, the first time I saw the map layout when the work adventure guys from RC3 team showed me the whole layout, which was, it's just big, it's just big. I was like, okay, you really, really made it to simulate those times you need from A to B in that big area of Leipzig. This is pretty cool. And what I heard today is that there's already a guy who's walking on his love band. Trainer. So like conveyor belt, training, cross-trainer. Yeah, training belt, whatever. And he's running on that or walking on that and he built or wanted to build a Raspberry Pi so that he can navigate his avatar by walking on his belt. I was like, okay, achievement unlocked. Good. So those times you need to come from A to B are already in the digital space. Yeah, what I find also interesting, you mentioned this thing of technical scaling of the system, but also scaling of social processes that happen inside. So when I came to the lobby for the first time, there were a lot of people around me and normally if you stand close to somebody, you see if he wants to have a conversation or not. Perhaps he's looking at a smartphone or writing a message or something. In this system, as soon as you get close to somebody, you're in a video conference and if there are a lot of people, then you don't have a chance to get out. So my pattern was to run up, like run and then you have this ping, ping, ping, ping when crossing all these video conferences until I got to a space where not so many people were and where I could organize myself and see where I have to go. But you don't have that if you don't allow your camera in the start. So you do not automatically join video chat with other people, so you can learn to navigate through without getting into conversations in the second one or second two. And it's, yeah, it's a bit of adopting to it. And I have the same problem. Actually, we've been walking, it's not really a fix that on the next version, when you are running or when you are walking, you won't be entering into any bubble. So it won't do any ping while you are running. Very good. Which is pretty cool, actually. Because you're not disturbing anyone if you're walking through them without stopping. And if you are stopping, well, probably you want to have a chat. They are cool. So as before we go in the last round, we will talk a little bit about the future of WorkAdventure and also the RC3 world. Just as an announcement for the people who watch the live stream, if you have any questions to us, to the two of them, put them in the IRC or dial in the phone number and then we do a little Q&A if there are questions left. So as a last question to both of you, perhaps David first, what are your future plans with WorkAdventure? I had a look at the GitHub issues, what are ideas in what direction the project might go. So give us a little insight, what do you think are the top one, two, three things that will happen to WorkAdventure in 2021? Okay. Well, we have a few ideas. Well, first of all, we would like a compatibility with mobile phones. Very cool. That would be cool because today you need a keyboard to play WorkAdventure. I read on Twitter, there were people connecting a Bluetooth keyboard to Android phones or so and played it on the phone. Oh, okay. They were like this one. Excellent. Yeah. And then we're going to focus maybe on... Sorry. Letting people... Today, you did a good job at having people subscribe to be able to connect to WorkAdventure. This is not part actually of the core solution. So we're going to probably build something around that maybe with a kind of virtual phones that enables you to connect to colleagues or to friends. So this is a feature we're going to add maybe more features regarding organizing events and the ability for instance to host big events regarding scaling, regarding the ability to directly from WorkAdventure play a video to 100, 200 people. And we're probably going to build a SaaS solution out of it. We want to keep it open source. The core of WorkAdventure will always stay open source. This is really important for us. This is one of the core value I believe in. At some point, we will certainly have to make money. But that would be probably by allowing people to create spaces easily without having to install their own WorkAdventure server. So we're going to work on this in 2021. So it might be a similar business model to GTSI or something like that, where you have the core software open source and have a... Yes, exactly. I was thinking about GitLab, but yes, GTSI would be exactly the same. And Claudia would... Yeah, right. What do you think? If you look to 2021 and I think of Gulasch Programmier Nacht and other events like that, perhaps at least at the first half year and next year there won't be the possibility to meet with thousands of people in person. So will the RC3 World be sort of a prototype for something that will be used next year as well or are there no plans or thoughts about that at the moment? Well, my glass bowl is broken, but... Yeah, let's... There are many people talking about what could happen next year already. There's speculation. I know I can tell for 2020 that on day four there will be a takedown party or a takedown as always after the closing. So we said, okay, we will tear down the RC3 2D world to keep the event character or to keep it special. And it's a bit... Sad. But also the thing is the Congress, as it's in a physical space we also tear down every time and we come back and build it anew the next year. And if, big if, not when, but if there would be a second RC3 necessary or if we do the next impossible thing, a hybrid event which is more impossible than doing an online only event for several reasons. It's a possibility, a good possibility to build up from what we created this year. But as I said, glass bowl is broken. So we'll see. What I found interesting today in our assembly was a discussion during lunch that at a normal Congress you take a hacker space or community that meets in the physical space and you have a representation of an assembly at the Congress. And we were talking about using our map that we have now after the Congress as well only as a map for the assembly as a virtual meeting space of this community afterwards. So this is I think also interesting that these two dimensions of physical and virtual are sort of swept like the Congress took part in a virtual way and perhaps the assemblies take their map and use that afterwards as well. So the Congress as a whole is teared down but the maps might survive. Yes, I heard from several assemblies that they would use their map afterwards for their own purposes which is pretty cool and probably if we use it or use work adventure and develop it on for some future events or whatever we think of next, I see good probability that there will be some pretty creative more developments also with the maps and whatever happens to the source code and so on. Pretty amazing to see what the community is making from it to make the best of 2020. I have two questions in the IRC. If somebody wants to call 8001 then you're connected to us or the long numbers in the chat. Someone's asking if I'm sitting in a cheesy room and see that someone else on the map is passing by, I want to be able to call the person and invite them in the cheesy map. What do you think of a feature like that? If you're sitting on a table in a physical space and say, Claudia, come here. I'm sitting here with two other people. Come join us. That would be really interesting. It's true that today you have to go out of the gist meeting room and run towards the person and say, hey, come in. Yes, exactly. It might be a good idea. Actually, I've not been talking about the limitations we have. Basically, when you're going to speak to someone you can be up to four people. This is four people because you're in a web RTC and when you speak to three other people you must send your video stream three times. Usually, you're quite limited regarding upload bandwidth. At some point it does not work and you need a GC server to do something. What I would like, ideally, is to have something that is basically between GC and the bubbles we have today where basically when we are five people we start using the GC server but keeping the UI of work-adventure. Basically, the bubble would go bigger and bigger as many people are coming in the bubble and so you could maybe work more easily towards someone and say, hey. Well, okay. Similar to the real world, the bubble of people talking getting bigger and then it splits into subgroups perhaps and so on. Yes, and if you want to go out of the bubble you can at any time and with small... There are actually big challenges regarding the UX, the user experience and how you do this and we have been quite lucky because when we started working on work-adventure during the first few months I think we got it right and it was really completely by chance. We got the size of the circle that was close to what we... Well, yeah. It worked at the first time. We did not have to make any trial and error. It was just pretty good. But improving this is quite hard. I believe that. Okay, so at the end of the time of the podcast thank you both very much that you showed up and shared your experience and your knowledge. I think Jinx, everybody at the CCC knows how to contact you perhaps. But if you want to, you can say in URL or your Twitter link, Nick, how people can come in touch with you. I'm not so much on Twitter anymore. But you can find me on Masterlon at vianarite.literatur.social. Very well. And David, if somebody wants to share experience with you or drop some lines of code or just say thank you, how can people get in touch with you? Yes. Well, I'm on Twitter, David underscore negrier, n-g-r-i-e-r. And especially you can actually connect to the GitHub account, the GitHub repository of WorkAdventure. It's the coding machine slash WorkAdventure. And if you have any questions, just fill an issue. Or comment on Twitter. If you want to start the project, do not hesitate. We'll do that. Yeah, thank you. And we'll give you a hint as soon as the repos are published so you might have a look at it as well. Yes, definitely. So thank you very much, both of you. Thanks everybody in the livestream. Thank you very much. Thanks to the Center-Centrum team for running the technology behind the podcast. Have a nice evening. Goodbye. Thank you very much. It was a pleasure.