 Well, yeah, my name is Eben Upton. I run a thing called the Raspberry Pi Foundation, based out of the UK. Today I thought I'd talk a little bit about some of the things that people have been doing with the Raspberry Pi in the 18 months since we launched it, and about some of the things that we've been doing, in particular some of the things we've been doing with the free and open source software community to try and make the Pi a better platform, and as a side effect, I guess, try and make ARM Linux a better platform. I think it's worth maybe briefly recapping how we got here, or at least how I got here. When I was a child, I had a machine called a BBC Micro-Computer. This was a machine which was largely confined to the United Kingdom, but I think many of us in this room grew up in that era of having machines in our bedrooms which were reasonably open, machines which maybe we didn't buy them in order to become computer programmers. Maybe we bought them to play games on. Maybe we had to make that kind of drive. I was never forced to make that kind of tough decision, but maybe we bought them to play games on. Maybe we bought them to do our schoolwork on, or maybe our parents bought them for us to do our schoolwork on. But they all had one thing in common, which is that you turn them on, and they go beep, and they give you a programming prompt as the very, very first thing out of the box. And so to some extent, if you wanted to do anything with them other than be a programmer, the first thing you had to do was choose not to be a programmer. Now, I got my BBC Micro when I was 11 years old, and it's amazing. I guess that's 1989. And it's amazing how different that era was. I remember as a child, Acorn who made the BBC Micro computer obviously went on to have an interesting history. They made a machine called the Archimedes, which had a processor in it called an ARM processor. So they've done all right for themselves. And I remember the magazines that talked about the BBC Micro would also talk about this Archimedes machine, which was an order of magnitude more expensive than I could ever hope to afford. And one of the wonderful things about the Archimedes was it had had what they call the WIMP system. It had a GUI. It had a mouse, and it had a mouse pointer, and it had windows. And it looked absolutely amazing. And I remember thinking as a child, it would be fantastic if my BBC there was no hope of me ever buying one of these machines. But I remember thinking as a child, it would be lovely if I could have that same sort of experience on my BBC Micro computer. And so I saved up, and I think I had a birthday. And the main thing that I was missing, I felt for my BBC Micro was a mouse. I just didn't physically have a mouse. So I went and looked in my computer magazine, and I went out and bought a mouse from a company called Watford Electronics, who were fairly infamous in the United Kingdom for their lousy customer service. And this thing turned up. I waited a week, and this thing turned up. And it was a mouse in a box. And that was it. It had a piece of wire that you could plug into your BBC Micro. And I had absolutely no idea as a child what to do with it. And I was a bit disgruntled. And I asked my father, who'd put in the order for me, could you phone customer service and find out what I was supposed to do? And he phoned up customer service and said, you sent my son this mouse. He didn't send any driver software. And the guy on the end of the phone said, if your son can't write his own mouse drivers, he doesn't deserve a mouse, which was fairly typical of Watford Electronics customer service for anyone from the UK. And that's how I learned assembly language. And somehow we've gone in a matter of 20, maybe the 15 years maybe. After that, we've gone from an environment in which that was a feasible answer from a nominally customer-focused organization on the phone to a world in which a world dominated by a different breed of hardware, a world dominated by a breed of hardware, where rather than the first decision you have to make being not to program, it becomes increasingly hard, or in the case of many of the platforms that we have in our houses today, which are, of course, much, much more powerful than the machines that I grew up with. These platforms are, in many cases, completely impossible to program. They're completely closed systems. They're games consoles, or they're tablets, or they're set up on boxes, or they're mobile phones. And these machines often completely unpreambles. So a group of us, this doesn't seem like such a bad thing. In many ways, of course, this is a great thing. This has expanded the pool of people who can use computers as productivity devices. But it's had this negative side effect that we've lost this. We used to have a positive externality that people buying computers in order to be users had a good chance of becoming developers. So for the software engineering community, there was a wonderful externality from these machines that we used to have. Cambridge, in the UK, where we're based, we have a university. We have a number of technology companies. And those organizations used to benefit from this stream of people who would arrive at university, as I did at the age of 18, having already been computer programmers for eight or 10 years. In the mid-1990s, we had something like six or seven applicants for every place on computer science at the University of Cambridge. We could rely on virtually everybody who came in the door having a good working understanding of how at least one computer worked. The vast majority of us coming in the door had done at least some assembly language programming. The first thing that the course had to do for many of us was to convince us that there was stuff left to learn. The first term would consist largely of beating you over the head with functional programming, which is a great way of convincing an assembly language programmer that maybe there are some things he doesn't understand. Now, I made it through the functional programming. I finished my time at Cambridge as a student. And I became what we call a director of studies. And the job of a director of studies is to go out and recruit new computer science students to come to the university. By 2005, when I started doing that, so 10 years after I'd arrived at Cambridge, the number of people applying to study computer science was halved. And the typical skill set of an 18-year-old arriving at the university to study computer science had gone from one or two different sorts of assembly language to maybe having done a little bit of web programming if you were really, really lucky. These were still massively bright kids. But we were having to spend an increasing amount of our precious. We have about 60 weeks of contact time. Cambridge is famous for its short terms. In the UK, we have about 20 weeks of contact time per year. So we have 60 weeks to turn somebody from a high school student into somebody who can start a UK three-year PhD program. And we were having to spend an increasing amount of those precious 60 weeks bringing people up to the sort of level that we'd originally been able to rely on. So a group of us at the university came up with this idea. We came up with the idea that we have a hypothesis. And of course, at this point, we're only 18 months into shipping Raspberry Pi's. So it still is a hypothesis. I don't think we've got any proof that we're right yet. The disappearance of those programmable pieces of hardware from the 1980s was directly responsible for the decline in the number of people applying to study computer science and the decline in their typical skill set. So a group of us at the university came up with the idea that we should build something that maybe there was a niche that we could fit something into. This was in 2006. We spent about five years up until about May of 2011 just trying to figure out what this thing would be. But we knew that it had to really be four things. We knew it had to be programmable, obviously. We knew it had to be fun. We knew that this machine had to get into children's lives for some reason other than just as a worthy programming toy. We're wondering if there's any questions. Let's see. Programmer? Fun. Robust. We wanted a thing that would fit into people's school bags and taken out many times without breaking. And we wanted it to be cheap. And we settled on the idea of building a thing which costs the same as a school textbook. We didn't go and check what textbooks cost. We thought they cost $25. So yeah, so first do your research. Yeah, we would have had a much easier time on the engineering front if we'd done a little bit of research first. So to cut a long story short, in May of 2011, we told everyone we were going to do it. By February of 2012, we actually knew how we were going to do it. And we put them on the market. We sold 100,000 Raspberry Pis on our first day. We took down the websites of all of our distributors, which made us very unpopular because our distributors are electronic competitor distributors. We had a 12-hour period where no one in the United Kingdom could buy a resistor. And in the 18 months since then, we've moved to our 1.6, 1.7 million Raspberry Pis. And so what I'm going to share with you today, I'm going to share with you a couple of cool projects that people have done with this. One of the wonderful things for us is we have a website. And we spent a lot of time early in the project talking about what we were doing to try and deliver this $25 computer to people. The wonderful thing for us these days is that we can feed our website almost entirely with, I think this speaks to the comment about how your user community will dominate the amount of content generated for a platform. We could feed our website entirely with things 10 times over with the things that other people are doing with Raspberry Pi. So I'm going to share a couple of those with you. I'm going to talk about a few trends that I think are hopeful for the Raspberry Pi. I have a technology demo, which may work even. And now I'm going to try and draw a couple of lessons, things that I've learned. Because I've really come to Raspberry Pi. I am a real neophyte in the world of free and open source software. I come from, I guess, a fairly traditional commercial software engineering background. I had a lot of misconceptions about what it was going to be like shipping a Linux-based computer. And maybe just share a couple of those with you. So first of all, some projects. A personal favorite. I don't know how many of you have seen this. This chap here is called Babbage. He's a stuffed bear. That below that you can see in the background is Oxfordshire in the United Kingdom. It's about 40 kilometers away. It turns out the big surprise for us. We were very, very focused as an organization on education. What we hadn't realized was that there were a large number of adults who would find interesting stuff to do with the Pi. So this is a chap called Dave Ackerman in the United Kingdom. He likes to stick Raspberry Pies under weather balloons and send them up in the air. He's been doing that a lot over the last year. There's an Austrian chap called Felix Baumgartner who set an altitude record for skydiving recently. And there was a desire maybe to have Babbage the Bear beat that record. Felix got to 39,969 meters. 38,969 meters. Babbage did just over 39. So he's beaten Felix by 31 meters. And there he goes. There he goes. Lovely. I mean, this is really neat stuff. And I think it's a testament to the power of what you get if you put powerful tools out in the environment and let cool people do cool stuff with it. I should probably try and link it up to education. The lovely thing about this is that the amount of money required to do this is like $200 of equipment to do this. When you take pictures from the capsule, there were also pictures from the capsule that he's in, you might as well be in space. You can see the curve of the earth and the blackness of space. This puts the space program within the reach of every primary school in the developed world. I think a big trend for us over the next year with Raspberry Pi is going to be people using it, not just to get people excited, not just to get children excited about computing, but to get children excited about the whole broad range of what we call STEM subjects, the science, technology, engineering, mathematics. Another one I really like, you can probably tell I'm a bit of a space cadet. I was sitting there actually itching to go and see how SpaceX got on if they launched yesterday. This is a picture of the moon. This was taken using Raspberry Pi. So these are actually both camera projects. We have a little camera peripheral for the Raspberry Pi. And this is someone who's taken his Raspberry Pi and attached his camera module to the eyepiece of a thing of four-inch telescope. Taking just these really amazing photos again gives schools, and there's some sort of stacking process he does where he takes many, many exposures. So again, another thing that would be lovely to see this sort of thing going on in education. So some trends. We really didn't think we were going to sell very many Raspberry Pi's. And so we didn't bother to do things like make a case. We thought it was not going to be economic to make an injection mold for a case. We didn't build any add-on boards for it. And what this has done, this has created an enormous ecosystem around the pie. We have in the UK two million pound, two companies, the two startups, unfunded startups that turned over a million pounds each in their first year making cases and add-on boards to the Raspberry Pi. So one of the wonderful things we discovered, again, I guess an example of our user community kind of out executing us, is that by leaving all of that value out on the table, we've created this big kind of ecosystem around the pie. We've created manufacturing. We've been able to create manufacturing jobs in bits of the United Kingdom that traditionally have some manufacturing jobs and don't really anymore. And it's just really wonderful to see this. And these are often young people. One of the guys, a guy called Jake Marsh, was I think a 20-year-old student, scraped together enough money to cut an injection mold for a pie case, turned over a million pounds in a year, kind of fun. Education, it is really early days. I don't think we can claim, there was a bump in the number of applicants to computer science at Cambridge last year, but I don't think that we can claim we were responsible for it, except maybe by making a lot of noise about right now how easy it is to get into Cambridge to read computer science. So, but I think we're starting to see just the first little whispers of an indication that we really are starting to have an impact. We are seeing children using the Raspberry Pi to do stuff. Tom and Alyssa here, the thing they're holding up is a Raspberry Pi with a shield on top called an Air Pi, which is an air quality monitoring system. These guys are at Westminster School. Anyone from the UK will know Westminster School is not exactly your average underfunded in a city-state school. But these guys have admittedly from a really advantaged background have been able to build this piece of hardware, open source all the designs. They ran a Kickstarter that got them enough money to deploy these in 100, 200 cities around the world. I think a big challenge for us at Raspberry Pi is to take these sorts of things that are currently happening with children at schools like Westminster and are currently happening with children whose parents are engineers. Right now, the Raspberry Pi is a great platform to work with your, to use to explain to you, if you're an engineer, to explain to your child what it is you do and share some of your enthusiasm for engineering with them. I think the big challenge versus the foundation over the next year is going to be to start to expand that out, to start to make that same sort of quality of experience available to children who are at less advantage schools, available to children who don't have an adult in their immediate environment who is an engineer. But we're getting there. Developing World, this is something that we, this is a Pi-based computer lab in, I think, in Ghana. Not something that had occurred to us at all, but of course, there is enormous interest in a low-cost Linux-based computer in the developing world. I don't think there is any question at all that Linux, and in particular Linux on ARM, are going to dominate Africa. I don't think there's any question that we're gonna roll out the same sort of utility computing experience that we have in the West in the developing world. One of the wonderful things about our little upside surprise in terms of sales, we do make a little bit of money every time we sell a Pi, and this has generated a surplus that's allowed us to engage with a number of free and open source software projects. These are all projects that we've directly contributed either engineering effort to or funding to. Very quickly, certain amount of small talk. It turns out if you wanna do education on computers, small talks, the language is a very relevant language to optimize. We've been funding work to accelerate the squeak, small talk, open small talk implementation on ARM, and then we've been funding work to improve one particular, very popular small talk application, Scratch out of MIT, which is kind of a de facto standard for teaching young kids to program. We've done a large amount of work, and I hope to show a demo in a moment, of on Wayland and Western getting, both getting Pi-specific work, getting Western developing a Pi-specific back-end for Western, but also general platform work improving the quality of ex-Wailand, improving some elements and adding protocol features to Wayland itself. I've been funding some work on Pi-Pi. We funded a little bit of the work that led to the recent release quality version of Pi-Pi for ARM, and then also bits and pieces of Libre Office, largely startup time optimization, and then Libre AV and XBMC. We have an honest-to-god consumer application for the Pi, which is, we reckon, of the 1.6 million we've sold, about 300,000 are being used as media centers are being used in the consumer application. We've, I think we've overtaken cracked Apple TV2s as the largest non-PC XBMC platform, so we've been spending money both on making the underlying codecs in Libre AV run better, and on making the XBMC try to hunt down some performance issues in XBMC. I mentioned desktop. Can we try for that demo? Ooh, there we are. So this is Western running on Raspberry Pi. I guess three things to demo. The first one, I mentioned we've spent a significant amount of work on Blimey, really pushing our luck, launching three copies of that. This is Scratch that I mentioned, and of course, demoing some interesting transparency issues with implementation at the moment. Okay, so this is demoing three things. One, the little animated widget up there demonstrating native Wayland applications running on the Pi. The copy of Scratch there, eventually, we would like to have a native version of that running under Wayland, but for now, that's the demonstration of X Wayland. That's the demonstration of the work that we've been doing on getting X Wayland to work performantly on the Pi and elsewhere. And then these little spinning triangles here are the first kind of proof of life from our EGL integration. So we're very hopeful we're gonna ship the technology preview of this in the next couple of weeks and open some bug tracking on it, and we very much hope by the end of the year we'll have a quick, back to my slides before anything. Perhaps we won't turn the bug tracking on on the first day. So I guess that's gonna be really exciting though. And this speaks, I think, to our trying to move the Pi consciously in the direction of making it a usable utility computer for the developing world and also trying to make it a machine which is more relevant for children. I said earlier that I really, really wanted this to be a fun thing for kids and I think making it a fun thing for kids means it has to run games, it has to play videos. You have to be able to go on Facebook on it. There's nothing wrong with those things. Those things can be hooks that will get this machine into a child's life. So, three lessons. If you build it, they will come. This was a real surprise to me. Like I say, I come from a very conventional commercial software engineering background. I have been constantly amazed by the extent to which with the Pi when we started doing a cool thing, it accreted people very fast. It accreted very, very high quality volunteers into the project extremely quickly. You know, guys like Dom Cobbly at Broadcom, guys like Alex Bradbury at the University and then a whole halo. So people who are fairly geographically close to us and then a whole halo of people who have been contributing really, really high quality, really, really high quality patches to our software. But there is a flip side to that. Like George R. R. Martin, the free and open source community is not your bitch. I think I maybe had a little bit of a vision that we could put something out there and the free and open source software community would fix all of the problems with it. I think what we found is that we've had amazing input from the free and open source software community, but you don't get what you want. You know, you'll get what the community wants to give you and there will always be a gap between what the community wants to give you and the things that you think you need as a commercial organization to try and to deliver your project. And so this I think is what's driven, you know, after the first six months, once we actually have some money, this is what has driven us to do those engagement projects that I mentioned earlier. And finally, low hanging fruit is plentiful and juicy. We're living in a really interesting time, right? Moore's Law just ran out. Economics just killed Moore's Law. We are not going to be able to rely any longer on this free doubling of the performance of any random piece of software every 18 months to two years. On the one hand, this is bad. I mean, we've done extremely well out of this as an engineering community for the past, you know, for my entire life. But the flip side of that is, there's never been a better time to be a software engineer. A lot of the engagement that we've been doing, I'd characterize as being attention to detail engagement, you know, finding things which are deeply, deeply unsexy and which work fine on an XA6, on a Core i7, and taking those things and just doing the kinds of attention to detail software engineering that can give you another factor of 100. There's an awful lot of factors of 100 out there available in the piece of the software. So it's actually quite a, you know, for me as a software engineer, I take a little detour into being an RMSOC designer. Following some recent posts, I've been checking my break cables and what's been put in my coffee. But yeah, I've taken a brief detour, but I'm a software engineer. I'm a software engineer at heart. I always have been, and there is no better time to be a software engineer. There's no better time to be working on platforms like Linux. So that's all I've got. We've had a hell of a time. We're really hoping that what we do over the next year is going to be as much fun as what we've done over the last year. So thanks very much. Thank you.