 So we had the Open Mandrava Summit here in Budapest. Hey, with Barrow. What's up? Hey. And this is Laska and Budo, right? Right. They're listening to your song going on right here. So hello, so who are you? Oh, well, I'm Colin Close and I'm the president of Open Mandrava. We're here to have a short conference about where we're going next in our distribution. Cool, here in Budapest. And hello, so who are you? Hello. I'm Christina. I'm making the graphic for the distribution. So right here, just an example of some graphic that's going on. This is the login screen. That's right, yes. And these beautiful flowers and what Ragiada produces. That's Ragiada's Christina's handle. And who are you? Hey. I'm her son. And the driver? No. The driver. Are you taking airplane? He's a tourist guide. Tourist. You're going to Italy or no? You're leaving, right? Yes, we are leaving. Cool, so how many people came to the summit, to the meeting here? I think yesterday we had eight people. So people came from around Europe? Yes. And what did you discuss? We discussed our next release. That's going to be probably in about eight months time. And it's quite a big change for us, because we're going to be moving to a rolling release model, which means we have to make quite a few changes in our building and quality assurance programs. And then we've been discussing this at some length. So that's the new plan, you might say. Let's talk more about that just in a second. But if you could come here and maybe show this fact. What's going on here? We're looking at something special right here. This is a project that we're moving into or trying to anyway. This is a 64-bit ARM laptop that we've created. I'll just power it up for you. It's a prototype, right? This is a prototype, yeah. And you can probably see inside that there's quite a lot of bodging going on. This is based on a laptop kit, in fact. This is the Pytop kit. This is a Pytop kit. And it uses a Snapdragon Qualcomm Snapdragon processor. Is it the Dragonboard 410C? It is indeed, yes. That's the 820 actually. That's the 820 one, sorry. You put the 820 already. This is 820. As you can see, or you will see in a minute, it's running Android. We are waiting, and we're right on the verge now, being able to have a working Adreno driver, which will allow us to run OpenMandriver on this machine. And this is the first step to us creating our own laptop. This one has been created from various modifications, as you can see, that's had an extra bit added on. I can just look around. You can see it's got a new bottom. And now the bottom in here lives the battery, which makes room for the card in here. Nice. So, how significant is this? Maybe you can come up here and introduce what we've been able to do with this summit. Yeah, so, probably given you are running ARM devices.net, people will be familiar with the ARM64 processors. They are really good processors that can almost match the speed of Intel processors at nowhere near the power consumption. They are really interesting for developers right now, because a lot of us will want to target the ARM64 hardware. We don't always want to cross-compile stuff. But at the moment, nobody is building desktops or laptops that have an ARM64 processor in them, and it's high-time that got fixed. So, the PyTOP guys who built this original laptop build kit did a lot of the right things. But they went with the Raspberry Pi, which is a really nice port, but it's not fast enough to do compiling on. It's a quad-core A53, right? Yeah. Yeah, it's like a Dragon 1.4.10c, basically. Yeah, and actually a bit slower, because they've used a slightly different variant. They use relatively slow memory. Like I said, it's a great port for doing embedded work, but it's not that great to compile stuff on. So, we got in this faster port. This is one of the fastest 64-bit ARM boards available at the moment. And the PyTOP allows you to put any board you want in there? How do you do to hack it up to make it work? Probably Colin can tell you more about that one. Well, this was a bit of an adventure, because as you can see, the PyTOP case is not very deep and the Snapdragon board is very long. It comes down to here. In fact, I'll show you inside you can see how we have managed to shoehorn it all in to the box. As you can see, there isn't very much room. The battery had to go in this piece at the bottom, so we had to make a hole right the way through the bottom of the case, so there is no bottom to this case, only this piece here, and the battery sits in this area here. All the connectors that were here on this board all had to be removed, and connectors that were underneath as well had to be removed in order to fit it in the space. But fortunately, there was just enough room, as you can see, just a few millimeters, literally, spare space here to get it in. What is this board? This is from Intrinsic. This is an Intrinsic... On the 820, they have a system like this where you can just plug it in as a little samba, what do you call it? It's SOM, yeah. This is an Intrinsic carrier board, which they created for the 820. Is it full? The batteries are full right now? Yeah, it is running full. We have a mouse and all sorts. You can also connect the USB? Yes, it has USB. We can connect a debugging port, which that's what we've been doing today. It has internet connection. It has GPS. Theoretically, you could even put a mobile phone sim card in it. Cool. And what's the RAM? Two of 4GB? It's a 3GB model, but we hope in the future there will be a version that has a bit more. How much do you want? Most better, obviously. 8 would be great, 16 would be better, 32 would be even better, but 3 is workable. Did you show this to Intrinsic? Not yet. There's a couple of things we want to sort out. Primarily, we want to get OpenManDriver on it, and we want to make it a little more reproducible. Obviously, it is hard to build more of those the way they are right now. Is this 30.3 inch? Yes, 30.3. So, there's free Drino available for the Snapdragon, right? So, potentially, you could use that to make OpenManDriver work? We will certainly go with free Drino on this one, but the problem is free Drino currently doesn't work on this particular revision of the chipset. But that's only a question of time. I know that Rob Clark is already working on it, so that shouldn't take much longer. And he said that he was showing off 820. He was able to do his free Drino stuff on the 820. Yes, there's still a couple of glitches, but it's beginning to work, and we'll soon have something really good running on here. A matter of days, no? A matter of days? Possibly, yes. When Barrow gets back to Switzerland, he's going to fix it up. How good is Barrow? He's brilliant. He is just so clever. I mean, to give you an idea, he has eyesight like Batman. If you come here, and you look here at his terminal screen, you can see I can't even make out the individual letters. Is that a 4K display? Yes. So you've got a 4K laptop right here to compile really fast to it? Yes. This is the first laptop I'm really happy with. On most other laptops, I've had the CPUs would always go into 800 MHz from constant overheating. This is meant as a gaming laptop, so it has a couple of things like advanced GPUs that I don't actually need, but it's big enough to allow some air circulation. It's got some big vents right here so you can get your compilation and clang, all these GCC stuff you're doing, right? Yes, sure. Are you using GCC anymore? No, we've now moved entirely to clang, and during our meeting today and yesterday, we decided that we would move towards LLD, which is the clang linker, so we will now be fully using clang all the way through our distribution. The switch was flipped today, shall we say? Why did you decide this? Well, that's a good question. It does open up the use of our software to do more commercial application, and that's one of the main reasons, but the clang compiler is much faster. It takes less space, it generates better binaries, which means we can build things quicker. Generally speaking, it's a big improvement. I'm sure Barry could give you more technical detail than I could. What can you say? One of the big reasons that Colin hasn't mentioned yet is the code of the compiler itself. So GCC is a great compiler, and that's great optimizations, but its code base is almost 30 years old, and that shows when you want to do something new, implement it back in for a new architecture, so it gets a little complicated because you have to deal with a lot of craft. And LLVM and clang are based on a much newer code base. They are much nicer to read, C++ code, and this sort of thing just enables new features like support for ARX64, support for newer revisions of a chip or so to go in much faster. That's the main reason why I think even though right now GCC and clang generate both pretty much equivalently good code, in the future probably LLVM and clang will be much better and it's obviously good to get the head start on the other distributions moving towards it. We are I think the only, the first and only distribution using clang to compile their distribution at the moment, so we are quite unique in that sense. And does that put you in an advantageous place? Well, we feel it does because when chips change, especially when this ARM revolution takes off as we feel it will, because low power is going to become the real need. Data centers and so on and so forth are getting overloaded now. They've reached the thermal limits of their building, so ARM servers are going to start coming along and so on. Anything ARM based has an edge when it's using clang because as Bear explained it can respond quicker to new chips to changes in the architecture. And ARM is far more flexible, it's not just limited to one manufacturer, you've got the core of the processor if you like and then the architectural aspect is down to the individual fabs to do different things with. So this requires a lot more flexibility from your compiler and from your build tools and the like. And at Linaro, the Linaro Connect that just happened around the corner here in Budapest. Can you hold this one thing? This is a big deal, getting all this ARM support smoothed out and optimized and accelerated and everything. How soon and how much is it going to cost to get this OpenMindriva Snapdragon 820 laptop mesh produced? That's a good question. So the first problem we have to sort out is to carry a board we'd like to replace it with a smaller board that actually has equivalent functionality. You want to optimize a little bit the whole thing, the size and everything. This is built on purpose to swap kind of like the board? Right. Next step will be actually building a case for which we don't have to do because this amount of manual work to get the new board in. We also want a better display. This is good enough for a Pi because that actually doesn't have the processing power to handle high resolution display. But given we have a nice board with a decent GPU we want at least an HD panel in there. Do you want a 4-key? That's too much. One of the big problems with the development of this is getting the correct display because these processors are typically dedicated towards mobile phones. They use the MIPI standards for the graphics connections. And unfortunately display panels currently are not manufactured very much bigger than tablet sizes. So it's very hard to find something to interface with the card. Did you have to do something to make this one work? Currently this one is using the HDMI board. So that issue doesn't come in. But when we want to go bigger. Somewhere down there. It's here. And the Pi top board has an HDMI to EDP in this case. Conversion process. So this is one of the sticking points at the moment in order to try and get a bigger display without having to carry all the extra current consumption that conversion involves. So we need a large display that uses the MIPI. I'm just checking there. Let's check if there's the brightness setting. So let's say OpenMendriva arrives on this. Let's talk about OpenMendriva. The advantages. Can you grab that laptop there? Are you able to log in and show some stuff? Yeah, here we go. Currently we are running KDE Plasma here. Can you carry this up and hold it up over here? So what is that that you're showing here? This is KDE5. The Plasma desktop. Fully optimised. And I currently use this wonderful full screen launcher which I find a really, really useful device. We provide all a wide range of software here. I don't know whether the easiest things. We provide the whole office suite, for example. Numerous tools for sorting out various issues. A good development process. We have our own build farm which we offer people use of. Build farm? Yes, it's called ABF. This is a distributed server build farm where users can join and provide build nodes to the build farm. It's quite a community-based process. That's a typical application, maybe K-Mail here, which I probably won't start because I haven't set it up on this box, but there we are. This is a very old machine, so it's not exactly the fastest thing on Earth, but... So do you think this meant to open and reverse the best Linux distribution? I do. Why do you think that? In many ways, most of the distributions are similar, and I definitely think there are other good distributions out there. I'd use any of the distributions over Windows or MacOS, but OpenMandriva does a couple of things that nobody else does, like we are the first to switch to Clang as the compiler. We always have the most current kernels available. Like right now, we have 4.10.2 and 4.11. Rc1 in our repositories. We always have the latest G-Lib C. We always try to optimize everything to take the last possible bit of performance and power savings out of the computers. I think one important thing is that, even though that might change at some point in the future, we are a small enough community to still be able to make quick decisions, like that switch over to LLD took us a couple of minutes to discuss about, and then a couple of minutes to flip the switch, while in most of the bigger distributions, that would take discussions taking at least a year. And it was decided right here, in this room? Yes. Yesterday afternoon. Over many biscuits and lots of tea. And some dog food. These guys. Guido is always ready to be on video, right? Don't you tell. So what else makes it the best? Can you talk about, can you show some stuff? What's going on over here? One of the really good things is that the user interface is so simple that I can give open and driven installation to my parents also who are not overly technical without having to worry about them not being able to use it. Did you do that already? Yes, of course. That's what they use. And they have no viruses, right? Right. No CIA spyware. But you might have installed a key log here, but that's just to make sure that they are okay, right? No, no, we just use secure shell, you know, log in remotely and help them out that way. So it's easy to use. It's using the latest stuff. How come is it using the latest stuff? We are just really active in looking at when new stuff is coming out and testing it and making sure that once it works it gets to our users quickly. So we have two repositories at the moment. One is Cooker, which is where all the experimental stuff goes, where you will sometimes get an update that breaks things, but that is expected because that's a developer tree. So sometimes we have to break things in order to fix them in the longer run. And the other is our 3.x repository, which is stable where only tested stuff goes. And that's what regular people would be using. So what's the history of OpenMendriva? Where does it come from? Oh gosh, that's a long, long story. Is it not going back to 1998 or so? It was founded by a guy called Gail Duval in France, and it was a fork of red hat, I'm sure. It became known as Mandrake Linux. And then after a while it was quoted on the Paris Stock Exchange and they actually issued shares, and some users bought shares in the company and it was listed on the boss. And it went well for some years. And then as things go, sometimes there is consolidation in the open software. Realm and Mandrava was, I think, suffered a lot from Microsoft. They lost several large contracts in Africa, for example, which set them back. Did Microsoft attack their markets or something? Yes, I think that would be fair to say that that was the case. Did they undercut them? They undercut them, yes. Like illegal rebates, like sometimes big companies do. They were undercut and they had a very large contract in Africa, in the education system, and they were effectively undermined, shall we say. Not by Intel, but by Microsoft. This is the same thing, Intel did to the one laptop per child. They just went around to all the countries and basically gave away a few hundred thousand laptops so they would sign with Intel and not take an OOPC. So that happened? Who were the people that started it and stuff? Well, Gail Duvall was one of the great founders and then there were several directors after that. Gail Duvall stayed on the team for quite some time until eventually he was moved on. There was a split in the management, the exact details that I don't know. You were not there? I wasn't there. You didn't instigate the split? I certainly didn't instigate the split. None of us had much to do with the company. I was actually working there between 1998 and 2000 or so, but there was also some disagreement with management, which is why I left the company. Then later on, let's jump directly ten years into the future when the company was going out of business, the distribution was returned to the community and then we came together and said, okay, we are going to maintain it and we are going to bring it back to the great distribution that it once was. How do Linux distributions exist and stuff? What's the business model? Good question. Everything is free and open source, right? Yeah, but there is still a business model somehow. We don't really have a business model. We are a non-profit and we are all volunteers, but of course we are really happy to get some donations so we can keep our server infrastructure running and if we were to find some really good source of money we could actually hire a developer to do some of the work full-time that would allow us to move even faster, but at the moment we don't really have any income and we are still getting by just fine, so we don't really need a business model in the traditional sense. We do have one sponsor. OpenMandrover was very successful in Brazil. Brazil had laws that forced open source software to be used there and Mandrover bought a company called Connectiva who were quite big in Brazil and although that company is now gone there are still people very interested in OpenMandrover there and we have one sponsor from there who sends us a small amount of money each year to help us along and they use our distribution commercially for government and business use. So it's sound enough for that. How many users are there, do you know? Good question. Well, in the last three months we have had 10,000 downloads of our ISO from SourceWorge so that will give you some idea. There are a lot of people who still think Mandrover is quite a good distribution because it started out as being very user friendly and it got a lot of people into Linux. I mean, the main reason I started using it was because it allowed me to access to Linux without too much pain. Obviously over the years I've learned how to deal with it in other ways but at the outset I would have been a lot more trouble if it hadn't been for something like OpenMandrover. It's easy to go on and install apps? It's much, much easier, yes. Do we install with that application? There is a simple interface here. Is it called Play Store? No, it's not called Play Store. Is it called App Store? No, it's not even called that. It's called RPM Drake at the moment although this is about to change and there's part of the meeting here that we had. Yeah, I'm not filming your password, don't worry. There's a possible test on 235 anyway. So here, this is our current interface but we will be updating this to a much, much newer KDE interface very soon. Sometimes the internet is really, really fast here in Budapest but sometimes it's not, so it's a little mixed. There's a little bit of a mix. Some guy here, right here. There we go. That was out of focus. So here we can pick applications. So how many apps? Oh gosh. I think there are around 15,000 packages under repository sweat now. That's just the main repository. We have a contrary repository which we're finding it hard to maintain which has got about 16,000 or so in it. Most of them, well, a good number of them still work but we could use some help with our contrary repository. So if there's anybody out there who wants to learn how to use RPM and get a few lessons in compiling and building software, then there's a big job there that we'd be more than happy to help somebody do. So you have the Chromebook R13, right? This is a Mediatek. Quadcore Mediatek big little. How many days before you have all the imagery we're running on it? That won't take much longer. As usual with ARM hardware there's a problem with graphics chip. In this particular case, we have an imagination power VR which doesn't have an open driver but the guys at Mediatek are actually really cooperative and will be sending me at least a binary driver so we can get that going quickly. They're actually part of the narrow, right? So you have some contacts? Yes. So you just send an email and they respond within three hours, right? Sometimes, yes. All right. So that's coming for sure. Why don't we have it work on the Samsung Chromebook Plus that I have with the ROG chip RK3309 OP1? I can't guarantee it because right now I don't have the hardware but in theory I don't think there's anything overly complicated in it. It's a relatively generic ARC64 and it's a Mali GPU which is problematic but which also has a driver that should at least be available so I don't see why it wouldn't work. Samsung is probably watching this video so they can just ship you one directly to Switzerland, right? That would be nice. So you're going to do the Chromebook Plus very soon and what else could be in the horizon? What else could be in the horizon? What could happen? We're going to be doing some major revisions and the big thing that's going to be happening is that we're moving to a new version of RPM and we're moving to the DNF package manager. This is probably one of the biggest changes that will go on with the distro in the next year. This is a big thing. For many, many years we've used a program called URPMI which is a Pearl-based package manager. It's done good service. It's been really useful but it's very, very difficult to maintain. It's very old code and Pearl and GTK don't go terribly well together. So we're looking at using the DNF packaging manager and probably what we will do is provide a wrapper script so that our users will not notice too huge a difference between the usage of URPMI and DNF. They'll still be able to type very similar commands to do the job. In order to do this, there are some very big changes that have to go on in our infrastructure, our ABF automated build farm which is an online build farm which helps us create the distribution. One of the big things is that we will have to change the structure of our repositories. So this is going to be quite a painful time for us. Probably we'll have about a month or so of really struggling a bit until we get this sorted out but once it's done with a bit of luck, packaging will be much more straightforward and the repositories will be much cleaner and things will work a lot better. So that's one of the biggest changes that we're going to go through. The other change is that we're going to try and implement a built-in QA tool into our ABF or our automated build farm. Currently we use an external tool but really it's not up to scratch. It takes our QA team a long time to test a very few packages. So we're hoping to improve that and so that we can be more responsive with updates. The distribution will get new software much more frequently because we'll be able to test more efficiently. So that's one of the biggest changes that's coming up and that will be in our 3.1 release which we think best guest estimate is six to eight months is what we're looking at for our next release although we will be putting out sub-releases in between on the current build which is LX3. So there will be an LX3.01 which is already out and an LX3.02 and maybe even a 3.03 to the current release. Could you describe some of the team that's working here because you said there was eight guys here yesterday so where do you come from and who are they? I'm the current president. I've not been always the president but I'm the current president. I'm based in London, my name's Colin Close. This is Bero, Bernard Rosengranzer who lives in the Alps. He's our most skilled developer. We have two other developers. One is TPG or Tomek. He's from Poland and we have another developer who's not very far from me in London called Crispin Boylan and those are our three main developers. I do a small bit of low level development but largely I help with the QA and then we have Christina which you've already met who is in charge of our graphics and she helps out with our forums. We have Raphael who is and JC, Jean-Claude they manage our website services and all the nuts and bolts that are required to keep the front end of Mandravelle advertising our websites, our forums, our mailing lists they keep those going and then we have two other developers in Russia Fedya and his shadow, they're good friends and they are largely responsible for maintaining and building new additions to our ABF build farm and they also work on the ARM processors. Fedya is very skilled in ARM processors and that's an area where he works that's some of the team. Who have I forgotten, Bear? I'm sure I've forgotten some of it. These guys? Well, of course they're from Sweden, Alaska. Sweden calls our mascot. Community evangelist. He appears on some of our logos and graphics and although he's much bigger than when he was first appeared on there. Are there enterprises or education who's using? Well, there's one other person I've forgotten. I knew there was somebody, our secretary of course, Kate Lebedev, who couldn't be here this time unfortunately but she's got a new job and she's really, really busy but Kate lives in Germany now in Berlin and she acts as secretary and is a great organiser and makes sure that we get things done basically. She's very good for that. And now would you like to encourage more people to join, right? Yeah, of course. Ideally I'd want a community that is slightly bigger than what we have right now but it is not super large like the Debian or Fedora's communities because when you start getting that big it starts becoming less flexible. It becomes more corporate though, right? The thing is it gets harder to get decisions done right now. It's easy for us to get even unanimous decisions to switch to some new technology even if people don't have as much experience with it as they used to have the older technologies. If you have hundreds or thousands of people working on it it's hard to find the consensus. And then people are kind of fighting for different roles and stuff? That could also happen. I want to be the maintainer. Then they fight to the death or something? How does it work? Fortunately so far we haven't seen too much of that but that was actually to some extent going on while Mandriva the company was in charge. We certainly won't tolerate that sort of behavior in our new community. How does it work to choose who's the maintainer? It's just the guy that has the most skills that has submitted most patches and he's the maintainer automatically. How does it work? Yeah, right now we also don't have super strict roles like I know I have some packages that I need to look after but if someone else has something to contribute to those packages I certainly won't eat them if they just go ahead and commit something. In the worst case we have to revert a commit that was broken but... You won't eat them for sure. I may feed them to the dogs but... Because you have vegetarians, right? Your dogs they eat anything, right? These guys they eat anything. Well we use GitHub for our RPM spec repositories so reversion is quite straightforward and we can obviously work with pull requests which makes it quite easy because the maintainer can review the commit before it's actually committed so it's quite easy to avoid conflict so we don't have a larger problem. We do try and be a very friendly community but that's the thing we don't tolerate. Bullies like Linus Torrells. You know, sort of bad behavior shall we say. I mean we set out not to have this rather adversarial aspect that was sometimes found in Linux distributions. Do you filter out or do you, what's called, censor bad words in the discussion? We just don't do it. There's no automatic kick-ban. No, I mean we try and be polite and civil and fun as well obviously. I mean I think working on the distribution it can become very dry and that's what we try and avoid. We try and always have a little bit of fun as well as do some serious work and it seems to work, doesn't it? I think we're all good friends and enjoy a company. If we're on IRC we quite often chat about other things apart from what we're actually doing. Like Trump maybe. Well, no, no, we don't. No, it's just not like that. So you're the friendly distribution, you are the latest tech, the best. And now there's a whole opportunity right here with the ARM laptops that are coming out, the ARM Chromebooks with Delta per mode, what do you call it? Of course and I'm also working on an ARM desktop box which should be ready pretty soon. Currently waiting for a mainboard to be delivered that is getting a little bit delayed by its manufacturer but afterwards we should also have a pretty nice desktop based on ARC64. And so the guys and girls that want to use an ARM laptop, ARM desktop, that might think that Chrome OS is not a real OS like you say kind of. They would like to crouton OS a call to get different Linux and so there's an opportunity right there for OpenManDriver to be the one people choose because you have the latest tech, right? Right. So it's potentially going to be a big deal. Possibly. That's why we're doing it. It's been jolly hard work. I mean we've got, we've now at a point with this where we know roughly how much it's going to cost to produce a prototype. We have approached, I've approached a number of companies and they have given me a price for creating what we need. And how much? The... Is that a secret? Well, it's less than 100,000 pounds. Okay, that sounds good. It sounds good. Then you can afford it. We can afford it, right? Yeah. It's painful but I work at Linaro so I'm not exactly poor. I'm going to put some of my personal resources into this so we will get it done. And then people will be able to buy it somewhere. They'll be suing some news about this. We hope so. That's not so wrong. We're still working away at it. It's going to happen. The current estimate for timing for a first prototype is about eight months. That's what I've come up with so far. It may be possible to shorten that if we can deal with the display issues. One of the problems, of course, as Barrow explained, is we don't have open source drivers currently for the SOM that we wish to use. But that's almost solved now so we're hoping that that will help in the timing aspect of things. But currently we're having to look at a design that covers a lot of different bases just in case this won't work then we can use something else. And that's not really the best way of going about things. Hopefully that issue is now resolved. It's going to be hard to get it ready for all the engineers in the next linear connect. That would be nice. That would be a pretty cool group of four or five hundred test users, right? Right. That would certainly be users interested in it. It's basically a laptop you put in in 96 boards. That could also be that, right? That would be another interesting variant that might also happen pretty soon. There's quite an issue with 96 boards though because many of the earlier ones don't have any interface for SSD drives or anything like that. They all operate off some relatively slow memory devices so they're not really suitable for this board. Well, there are some that are coming out. There's also two consumer edition boards coming out soon that will be fast enough to reveal any details on those. You know lots of secrets because you have access to everything. The whole world. You have an overview of what's going on, right? I know what's happening inside Linaro but I don't know what's happening in the rest of the world. Okay, cool. It doesn't get told. Cool. All right. Thanks a lot for showing this off. World exclusive first view on the ARM 64-powered laptop soon to be running OpenMendriva. Here at the OpenMendriva Summit in Budapest. When is the next meeting going to happen? The next meeting will be next year, sometime around the same time. We haven't decided on the venue yet but I daresay it will be in Eastern Europe because it's convenient and not too expensive so that makes it life easier for all of us. There seems to be a huge train station just up there. That's right. That's the Western train station and there's another one just the same a few miles away which is the Eastern one. You're going to take the train in a few hours? Yes, I'm leaving here at 8.40. Taking the train was certainly the most convenient way to get the dogs here. So these guys, they don't want to go in what's called a luggage compartment. Right, and I don't want to torture them so... But this guy, you could carry him up as a hand luggage, right? Probably. I don't want to separate them. He wants to be an easy jet, I think. He's posing, isn't he? Look at him. Because this is a very cheap wizard. Easy jet to direct Geneva, that's very near your home, right? Yeah, that's about a two-hour drive from home. And there's also other airports like Basel or... Yeah, Basel and Zurich and Milano are all within driving range from me. Do you have some people come to your home to work on some projects? You haven't hosted any summits there yet? Not yet, but it would certainly be possible. Yeah, you should do a two-week festival. Yeah, yeah, I'll count for that. Two weeks in the house? Yeah, this summer you can compete with a paleo festival in Switzerland or some other, and do a whole bunch of campers outside your home. Is there a big space, right? It's up in the top of the mountain, basically. Not exactly on the top, but quite a bit above the village. I'm at 1,300 meters altitude. Can you make sure there's going to be a gigabit ethernet connection? Not quite yet, but we have pretty good DSL connectivity and fiber will probably come not too far away. Swisscom is pretty amazing. They want to give a gigabit to every Swiss by the end of next year or something. It's coming, so for sure your festival is going to be a pretty fast connection and people can just bring their own tents, right? And stay with all the mountain dogs that save people. Thanks a lot. Thank you, Nikas.