There was a problem… on one of the best [comp sci] courses in the world, and applications were dropping like a stone… kids had stopped coding”
What started out as a project to boost the number of applications to Computer Science at Cambridge University by inspiring a generation of children programmers has turned into a worldwide revolution.
November 26th marked the release of Raspberry Pi Zero – a five dollar computer.
Jack Lang, Chairman of the Raspberry Pi Foundation and part of the original Pi Team, took to the Business of Software Europe stage to share the story of Raspberry Pi in 2015: It’s incredible journey from project, to world changer.
Video & Transcript below
Get on the BoS Conference Mailing List
Be first to see new videos, join free speaker hangouts, stay up to date on BoS Conference programmes and a regular dose of thought provoking ideas.
Unsubscribe anytime. We will never sell your email address.
Jack Lang, Raspberry Pi Foundation:
I’m a failed academic. I used to be a respectable academic and then computing got the better of me.
So I live in the computer lab. I also teach in the JBSC school and I start various companies.
Most IoT systems look much the same. You have devices at the bottom that collect data or turn things on and off or do things like that, sensors and actuators. They talk typically to a local hub using a variety of transmission methods, but usually sending packets, using P packets typically or TCIP. That hub talks to the cloud, typically using as a web server, and the cloud talks to either devices or to a server that collects data and does analytics and virtually all IoT systems look like that.
However, they live in silos. This is one company’s IoT system. Doesn’t talk to this company’s IoT system. Doesn’t talk to this company’s IoT system. And if IoT is going to go anywhere, we have to have horizontal standards so that I can buy a temperature controller from this company that talks to a hub from this company that talks to a third party analytics company and so on. And we’re not yet seeing that.
So the common standard as the last speaker said for diverse horizontal integration discovery. So say I am a thermostat reading centigrade in freezer three and I need to talk to a power control center and so on. Hypercat’s are beginning to give us this, is one approach but it’s not necessarily the only right approach.
Ok, this is the academic bit. Adoption curves go typically, they’re not straight line, they’re not even square way, square function, as implies. They typically almost nobody uses it and then everybody gets the idea and everybody uses it and then almost everybody’s bought one, use it whoever will. So you’re getting your product to over this hump or to this curve where it becomes the defacto standard is the important thing to do if you want to win. Anything you can do to get it adopted as the major thing people use, take this TCIP for example, is the way to go. So a closed model where one company wants to control everything, think of Apple, means you have to think of every possible application you’re going to do and make it happen.
That means you have to talk each and every customer or have your distributors talk to every customer and that’s jolly hard work or turn to the open model is that you get people working for you instead. So instead of having a few distributors, you have lots and lots and lots of people who can you access your code and make it better and they work for you for free. You reward them with status instead of with money and that’s a good thing for a small company. So if you’re a small company starting up open source is great stuff because you get the world working for you, and specialising your product into their own particular vertical or their own particular market.
So let me tell you about the Raspberry Pi story. Raspberry Pi is a small form of computer. It’s about the size of a credit card. Who’s got a Raspberry Pi? Well pretty well everyone’s got one. Good. What are using it for?
Audience Member: Media server.
Jack Lang, Raspberry Pi Foundation: Media server. That’s one big use. Teaching. Home control. Can we play the video?
Media server. That’s one big use. Teaching. Home control.
Video: This is a Raspberry Pi. It’s a credit card sized computer that costs around twenty five pounds designed to teach young people to program and is capable of doing all kinds of wonderful things. Back in the eighties, kids had to learn how to code computers to use them and as a result, these kids grew up with an inbuilt understanding of how computers work. Now we need more programmers than ever before, so to deal with this problem, some private people came up with the Raspberry Pi to reignite the spark. It runs Linux a free operating system from an SD card, just like the one in your digital camera and it’s powered by a USB phone charger. You just plug in a mouse and a keyboard, connect to a TV or monitor and you’re ready to go. In schools, not only is Raspberry Pi a great way to learn programming skills as part of ICT. There are also dozens of cross curricula applications like science, and music and all over the world, people are experimenting with Raspberry Pi’s and attending Raspberry Jam events where people of all ages are learning what can be done with a Raspberry Pi. Since the first Raspberry Pi was shipped, we’ve seen examples of people using the Pi in a variety of amazing and interesting projects. Taking advantage of its size, portability, cost, programmability, and connectability. So whether you want to learn to make games, build robots, or even teach a bear to parachute, with Raspberry Pi, the sky is the limit.
Jack Lang, Raspberry Pi Foundation: Ok, thank you. So it’s got forty IO pins, ethernet, four USB ports, HDMI, camera, panel interface and so on. It’s a quad core seven hundred processor, so it’s fairly powerful. So you can run control systems and this web server on it. Where did it start?
In Cambridge where I am a member of the faculty of computer science and in Cambridge we had a problem in about 2008. The number of students applying to read computer science in one of the best courses in the world was dropping like a stone.
We normally admit about a hundred students and then we get, chosen from an application of five hundred students and that dropped down to about two hundred students. We weren’t getting enough bright people to fill the course, so we had a problem. We thought one of the reason this might be happening was because when we grew up we had the BBC micro. I helped, I was a part of the design team. I did some of the operating system. People have BBC micros or Sinclair machines that you had to program to make it do anything. These days kids download, they don’t program. You can’t, if you’ve got a cell phone, you can’t easily program it. It’s a sealed unit. You may be able to make an app, but you have to get the approval of the manufacturers to publish that app, so it’s a sealed box. If you have a games console, you can’t make new games for it. You have to get, the games are encrypted so the manufacturer can gain more revenue. So kids don’t program. So we thought that we could give them something that was cheap enough that they could break. They might tinker with it and make some programs and we’d get more applicants. So around 2005, computer firms were expanding but getting fewer programmers, where in that, because we weren’t getting the applicants, we weren’t generating enough graduates so we weren’t getting enough graduate programmers. School GSE computing is terrible. There was something called ICT which was how to use Microsoft products. The kids already knew that. They’d been using Word since the age of three. Various other people had noticed. As Eric Schmidt said you’re throwing away your heritage. University admissions had dropped by 50 percent. Kids stopped programming, they downloaded and various other groups, in particular the Computing at School group which I recommend has formed to try and do something about this. They campaigned and changed the A level syllabus.
So in February 2008, I wrote the paper saying what’s the BBC micro for today and various foreign travelers came along. Eben Upton had been making small computers at Broadcom to see what would happen. David Braben who runs Frontier Technologies needed more games programmers and Alan Mycroft and Rob Mullins from the lab joined us and we built some prototypes and we wanted to get the BBC imprint on them. The BBC had difficulty doing this because they’re now a corporate body and they can’t recommend one manufacturer over another. So we took it down to Rory Cellen-Jones, the BBC correspondent and he said he can’t do anything about it but can I take a picture and put it on my blog? So he put it on his blog and we got eight hundred thousand downloads. So that demonstrated there was a market out there. But then we had a problem. We promised eight hundred thousand people that we were going to make a £25 computer and we didn’t know how. The prototype was hand built and cost several hundred dollars. So we got to work. I missed out one of our founders who is an excellent engineer and he did some product engineering. We learnt about mass production very quickly and things like using the same value of resistor everywhere.
And we raised some funding from Cambridge Angels as donations. We were a charity. We are set up as an academic charity. We got some soft loans and things.
We thought we might sell ten thousand of these units worldwide
We got together a kit for two thousand units. We sent this off to a contract manufacturer Shenzen through a friend of a friend so we sent off a £250,000,000 worth of chips to a flat in Hong Kong, which was the relevant address, and held our breath.
In the meantime, we’d put some software that we had developed for it up on the web and we got sixty thousand downloads of software for a computer that didn’t exist. We’d announced that we were going to launch it in the fourth quarter of 2011. On 17th February which is the middle of the fourth quarter of 2011 we started getting social media messages, this was all a scam and didn’t exist because nobody had seen anything. So we put up some prototypes on eBay and they went for£2,000. This was for a $25 computer. So we realised we had a bit of a problem.
Then a large pallet arrived in my garage and we took one out and tested it and, by god, it worked. We took out another one and tested it and that worked too. So we were lucky with our manufacturer. But we realised that six people part time in a garage couldn’t satisfy the market. So we did deals with Radio Spares and Farnell and changed to an IPR licensing model. So we licensed and designed to them and they used their dollars to manufacture it. They bought our stock. We launched on the 29th February 2012, leap year day and we took both of their websites down. These are multi-billion dollar companies. We took their websites down. We were getting seven hundred orders a second at peak. We sold a hundred thousand on the first day. Don’t forget we only manufactured two thousand.
“So we had a small problem”
So we had a volume manufacturer. We spun up two more CM manufacturers and one in Wales who approached us and they were very good. It was the old Sony factory in Pencoed. And we also we had to get various approvals. If you’re selling a few thousand you can get away with a development board. If you’re selling millions you have to get the right approvals, the CE, FCC, radiation, RoHS and all sorts of other things. So that’s the factory in Wales that’s turning out eighty thousand a day at the moment, about two hundred thousand a week.
So regulatory issues, we had to go through the development process and get through that. It turned out people are using this thing for all sorts of things other than educational use so that had implications for our charity status. So we had to set up a separate trading company, selling for non-academic use and start tackling worldwide problems like import tariffs. In Brazil, if you import it as a finished unit, you pay a hundred percent import duty. If you import is as components, you pay three percent. That’s worldwide distribution. That’s people who register it. This was an app put up by one our users, a sixteen year old called Ryan and he put it up and people started registering it and that gives you an idea of the worldwide distribution of Pis.
We’ve sold about six million. Typical uses are scratch, that’s the MIT educational language, Minecraft, there’s a Minecraft extension with an API to Python and Java so you can programmatically build walls. XBMC media center. More than five million, that’s out of date now. It’s more like six million. And the applicants for Cambridge went back up.
It’s very popular in developing countries where low cost computing is important and the models that use it for example, Rachel Pi from World Possible, which puts onto an SD card, the Kahn Academy with the whole of Wikipedia, Gutenberg texts for schools, medical advice and so on and drops this in as a knowledge server into a school or into a community.
Lot’s of people doing IAT things. That’s a beer fridge that you can monitor the temperature and you can also monitor the consumption. It’s got it sitting on a weight sensor, so you can monitor who’s taking the beer out when. If your eighteen-year-old son is stealing the beer and not having the sense to refill it, to put it back with a can full of water. That’s an engine testing rig run by Raspberry Pi. That’s a hollow lens. We run Windows 10 IoT and that is a Raspberry Pi controlled robot with a virtual robot superimposed on it, hollow lens. That’s a garden control system which automatically waters your plants for you.
So the market will surprise you if you’re a startup.
Fail early, fail often
because you’ll only get to find out what the markets are like when you get out there. We discovered this market accidentally. Go for the low hanging fruit and get it out there. Get other people using it and testing it. The best is the enemy of the good. If you wait till it’s perfect you’ll never ship something. New skills are needed because we’re living in an age of global information and global accessibility. What happens when every light bulb and every light switch has got an IP address and possibly a camera? How’s the world going to look?
Computer Science is not an optional extra, it’s part of the fabric of the way of the world is happening. The way we’re teaching is changing. Instead of having a boring lecturer like me sit up and talk to you, you do a bulk transfer online and the you spend the face to face time doing a tutorial. So it’s not about the hour a week in school that matters; it’s the three hours a night in your bedroom. It’s like learning a musical instrument, learning to code. You see, then thousand hours the kids have to put in. Ok, that’s the official end of the official part.
Audience Question: Firstly thank you for the BBC because that’s how I started learning to program in basic. I think Raspberry Pi is an amazing project. I’m involved in teaching primary school children to code. What do you see as future products that we could do as an industry to help create a new generation of software engineers after Raspberry Pi?
Jack Lang, Raspberry Pi Foundation: Well I’m not going to announce our future projects because remember the Osborne company? They announced these future projects before they finished selling the previous project and then people stopped buying them and they went bust. I think the thing is to make it accessible and make it fun. Physical computing works quite well to get the kids involved. When they see something move or they can actually make it useful for them. Take robots, for example, they get involved and get a thrill. Let them make mistakes because that’s the way they learn. Also, let the kids have ownership rather than the school have ownership and keep it locked up. I’d say you probably have to talk to your IT people about allowing access to the internet.
Audience Question: You talked about the need for change in schools, curriculum change and that kind of thing and you mentioned an organisation that was set up to do that. Are you seeing progress in that area?
Jack Lang, Raspberry Pi Foundation: Very much so. The GCSE and A level, the syllabus has changed to a real computer science syllabus from what was a computer user syllabus. It’s a question of what you think the balance between computer users and computer designers should be. I believe everybody should have at least basic knowledge of what’s inside the box otherwise it’s magic. Advanced technology looks like magic. You need to learn the spells. The notion is you can make new spells if you know how the spells work.
Jack Lang, Raspberry Pi Foundation: Jack, thanks for your work. I got an SMS a couple of hours ago letting me know my washing machine finished running back in the states [laughter] thanks to Raspberry Pi there washing it. So, to the interest in computer science and that kind of reversal of trend, what were some of the other factors there that were driving that?
Jack Lang, Raspberry Pi Foundation: Hard to say. The economy possibly. A change of emphasis in the media from computer programming being seen as a rubbish likely to be exported to China to seeing it as a real career move and possibly a future career. It used to be that being a bank manager was thought to be a safe job. Being a journalist or being a programmer, not safe. Now it’s the opposite way around. So I think it’s probably, I think it’s probably the change of direction so your grandmother doesn’t think programming is serious, it’s anything to spend your life doing.
Audience Question: So are you generally seeing more optimism from your students? More optimism, they feel better about the future?
Jack Lang, Raspberry Pi Foundation: I think more optimism and the gender divide is getting better as well. It used to be we had no girls. Now we have about twenty percent girls. That’s still not enough
Mark Littlewood: What do you think about the difference between, there’s a big move towards code in a day, or code in an hour, or learn to code and problem solving as a discipline in school.
Jack Lang, Raspberry Pi Foundation: I think learning to program which is where you do something to actually test it is the new Latin. It’s a way of thinking, a compositional thinking is a new way about solving problems.
Mark Littlewood: My son is eight. He is taught some computer science at school, but he doesn’t like thinking of himself as a coder or a programmer, but he spends sixteen hours a day on Minecraft if he can, which, and he’s a red stone expert. He doesn’t think that is coding. Are there different ways of teaching people and getting young people involved in coding?
Jack Lang, Raspberry Pi Foundation: I think you have to show it does something useful so if he discovers the program I think, input to Minecraft so he can automate some of his, some of his moves, then he’ll get interested and that will extend to other things
Audience Question: Do you think there’s an analogy in how you teach people who are established in and running businesses a bit more about technology and engineering? It’s very easy in an environment like this to expect somebody or more than one person to be technically capable given a tech related company, but I deal with many companies, some of them governmental, some of them not, and there still isn’t a pervasive understanding that you need to know about technology in the way that you wouldn’t employ someone who didn’t know about finance or didn’t know about time management, and these sort of initiatives to get children drawn in are understandable and important, but they don’t deal with the next ten of fifteen years where we still have companies being steered by people who kind of still think it’s funny to laugh and say ha ha I’m not technical but, you know, if you went to a hospital and the consultant laughed and said ha ha I’m not medical you probably wouldn’t accept that.
Jack Lang, Raspberry Pi Foundation: Yes, I think everybody should learn to program. But you have to make it useful to them. Particularly, for example, in the urban countries. Imagine giving a farmer something to program when what he really needs is simple water. If they can’t type, it’s not much use. A smartphone would be much better for them where they can access market data, for example. So its horses for courses.
Mark Littlewood: Let’s all say thank you very much to Jack Lang and for being an inspiration to a new generation and of people that are going to solve problems and change the world
Jack Lang, Raspberry Pi Foundation: I should ask what do you want? You’re the guys who are going to change the world. I’ve had most of my fun. What do you want in the world?
Mark Littlewood: Violet, what do you want?
Violet Littlewood: Equal opportunities.
Mark Littlewood: To program?
Violet Littlewood: Yeah?
Mark Littlewood: Do you get equal opportunity at the moment?
Violet Littlewood: Well, yeah and I’m very lucky but I want equal opportunities for not just myself but for everybody.
Jack Lang, Raspberry Pi Foundation: Maybe that’s equal access to data.
Audience Question: I would love it if my twelve year old daughter could not feel intimidated in a computer science class because it’s filled, because it’s a boys thing. I hope she’s not the only one.
Jack Lang, Raspberry Pi Foundation: Something strange happens between primary school and secondary school. Primary school is about fifty percent male/female. Secondary school is male dominated.
Mark Littlewood: I think you’re doing a little bit to sort that Violet. I’m very afraid of my daughter
Jack Lang, Raspberry Pi Foundation: Buy her Raspberry Pi.
Audience Question: I’m not really a techie per se, however, I have an eleven year old daughter who I just signed up to hack lab for the summer and she’s artistic and that telling girls that they need to have to coding as part of their lives that they can’t really opt out of it even if it is a boy thing. Do you think that the parent, to a certain extent it isn’t just down to schools, it’s down to parents supporting their girls in doing, in pursuing that sort of thing and then even within secondary school making sure that, for example, I volunteer at my primary school in the tech club and I’m the only woman who does it and I will endeavor to do that in her secondary school so that the girls within the secondary school, you know, I’ve learned to code through tech club at school and at night she takes adults, showing the way to certain extent for girls as much as hoping the teachers will do it, just to comment, really maybe that more than a question
Jack Lang, Raspberry Pi Foundation: Exactly. Professor Mitros said what you need is grandmothers to say yes that’s very good, what else could it do?
Mark Littlewood: Fantastic. On that note, it’s lunch time. Thank you very much indeed, Jack.