Article - Issue 57, December 2013
David Braben FREng
An Elite Entrepreneur
Bell and Braben’s Elite broke new ground when it was published in 1984. It was published by Acornsoft for the BBC Micro and the Acorn Electron computers
A multi-million pound business that develops computer games had its origins in David Braben’s schoolboy hobby of programming a simple PC. Three decades later, his company, Frontier Developments plc, has successfully floated on the London Stock Exchange. Braben has been instrumental as a co-founder of the Raspberry Pi Foundation, helping to bring to thousands of schools a modern equivalent of the hobbyist computer that got him started. He talked to Michael Kenward at Frontier’s HQ in Cambridge.
As a schoolboy, David Braben loved writing programs on an Acorn Atom. This was a time before personal computers provided sophisticated programming packages to manage all aspects of life. It was certainly before computer games rivalled feature films in their visual effects and complexity.
The Acorn’s attraction for this youngster with an interest in science and engineering was that he could easily teach himself how to write programs as all the tools and languages were provided with the machine, from BASIC to full assembly language programming. “I found the whole learning process extremely compelling,” he says. These days, Braben runs a software business based in the Cambridge Science Park with a game development studio in Halifax, Nova Scotia.
The company, Frontier Developments plc, employs more than 250 people, with around 100 programmers working alongside 100 designers. The business develops computer games that sell in their millions.
These games have large and devoted followings. Last year, nearly 27,000 fans opened their wallets on Kickstarter, the crowdsourcing site, to raise £1.25 million for the development of the latest version of the Elite game, now one of Frontier’s flagship products. It will cost a lot more than that to develop the game, says Braben, but crowdsourcing also offers “a fantastic, novel opportunity for bringing a game to market without using a publisher and building a community around the game.”
David Braben FREng
This is a far cry from the development cost of the first version, which grew out of Braben’s undergraduate days. He wrote the game with a fellow undergraduate, Ian Bell, during his degree course in natural sciences at the University of Cambridge.
It was parental advice that had kept Braben from taking a degree in computing. His physicist father had advised him that he would be better off studying physics to add to his knowledge of computer science. “He was absolutely right,” says Braben. The wisdom of that decision became clearer after he started his degree course. Cambridge didn’t have a course in computing back then, just a ‘computer science conversion’ that allowed anyone studying natural science to switch to computer science towards the end of their degree.
That was Braben’s plan, until he saw what Cambridge taught as computing. He was not impressed. “However, this was down to the state of the subject at the time. The field was so immature that you could already pick up most of what was being taught.” In the event, Braben switched to electrical sciences for the final stages of his degree course.
Braben continued to develop his computing skills during his undergraduate years. He and Bell managed to complete work on the first version of the Elite game during this time, but did not know what they could do with it. Who would sell it for them?
Computer games were primitive in the early 1980s, as was the business around them. “It was very amateurish,” says Braben. ”Most games makers set out to mimic arcade games. It was focused on what the industry called the coin drop, with the goal of extracting another coin from the player’s pocket.”
Braben saw little point in forcing the same four or five minute experience on games for the home market. “You have got no coin drop, so there is no point. With a leap, we said, ‘Why do we have that?’” Braben and Bell wanted to create a game that you could play for hours. This did not go down well with the major entertainment companies, such as Thorn-EMI. When Braben talked to the company about selling computer games, “They said, ‘Oh no, total playtime of 10 minutes should be the maximum.’”
One company took a different view: Acorn. “It couldn’t have been a bigger contrast. They were proper enthusiasts, whereas I would describe Thorn EMI as being businessmen.” So when the two undergraduates were looking for a home for the result of their ‘hobby’, they went with Acorn.
In 1984, the duo’s open-ended Elite computer game – originally written for the BBC Micro – didn’t just break the four-minute record, it set new standards for programming with its 3D graphics. Elite ended up on most consumer computer platforms, becoming a classic, which helps to explain why an army of fans is prepared to pay up front for a new version.
Elite’ s arrival on the market coincided with the end of Braben’s undergraduate days. “My university work became my hobby. Essentially, by the time I left university I was running a business.”
This did not stop Braben from toying with the idea of staying on to do a PhD. He had already applied his computing skills to astronomy – looking at gravity and planets around binary stars. He thought about accepting an invitation to turn his work into a PhD project, but, with a business to run, Braben went on to become a freelance game designer. This was not, though, the end of his interest in astronomy. As a games designer, he would indulge his interest and design his own universes, galaxies and planets as central elements in a series of successful games.
Braben worked as a freelance designer for several more years, during which time the industry began to mature. He set up Frontier Developments in 1994. Since then, the series of RollerCoaster Tycoon simulation and strategy games has sold more than five million copies, spent time at the top of the sales charts and even found a niche in management training.
Frontier continues to travel in space with the Elite series. The original game sold more than a million copies and still has its fans. Does it, though, have any relationship with real astronomy? Don’t the games break all the laws of physics? Braben takes onboard such ill-informed questions and patiently explains that, yes, games may mess around with physics, but they still have to follow rules or the players would be able to demolish everything. One gets the sense that messing around too much with the laws of physics would offend Braben.
He responds with the same forbearance when asked if writing games really is ‘engineering’. “Most of what we do here is engineering, software engineering.” He compares what games writers do with how architects work. Both work mostly on a computer, creating buildings that someone else will realise. “I see engineering in the virtual world as no different from engineering in a real world.”
Games software also requires heavyweight software engineering, with unique requirements for reliability and the ability to work on different hardware.
Braben does see signs of growing appreciation of the sophistication of gaming software. This is helped by the fact that many of today’s engineers are, as the jargon has it, ‘digital natives’, people who have been surrounded by PCs, games and other software all their lives. Then there is the rise of simulation in many other areas of engineering. Simulation, says Braben, is at the heart of gaming. Games designers, for example, simulate the appearance of objects, such as a car’s surface. “When we render it in a game, you can’t just render one surface, you have to render all of the subsurface layers. If we didn’t capture and model all the subsurface ‘scattering’ that naturally occurs, then it would just look plastic.”
Games can also be important drivers of computer sales, perhaps increasingly so. The power of most hardware is now such that games are just about the only software that puts any strain on the system. “If you look back 10 years, people were getting faster computers because they really did want them so that they could do things like video. Ten years ago, a computer couldn’t really do decent video. Nowadays, they can do it without breaking a sweat. The real issue for the big computer companies”, says Braben, “is why do people need to replace their computers? What is the driving force? Currently, it is games.”
One of Braben’s recent ventures is about as far removed from the cutting edge of computer hardware as it is possible to get. He was one of the six founders who devised the Raspberry Pi, a mini computer that harks back to the Acorn Atom and BBC Micro – the computers that first appealed to Braben and thousands of young people like him and helped to create a generation of computer enthusiasts. He says: “In the early 2000s, I noticed that at Frontier the number of people getting through to the interview stage for our jobs was decreasing. That was quite worrying. We wanted more people, not fewer.” At first, he wondered if it was the company’s image, but he soon found that it was an industry-wide issue.
This didn’t make sense to someone who could see the huge sales of computer games, many of them to just the sort of people who might be tempted to study computing at university. Braben turned to the keen young players who tested the company’s games prior to release. “I just asked the kids, ‘What is your favourite subject at school?’ And then, ‘What is your worst subject at school?’” The same answer to the second question came back loud and clear: it was Information and communications technology (ICT) as the lessons were usually dull and limited.
There seemed to be two reasons for this, the pupils often knew more about the subject than their teachers, and the curriculum contained next to nothing about computing. It was all about using spreadsheets and word processors. “Essentially, it was office skills. That is what ICT had become.” On top of that, many schools had dropped computer science as a subject.
When Braben raised the problem with government a decade ago, he was brushed aside and accused of special pleading, despite his protestations that a skills shortage would bite every industry that needed good graduates. Braben carried on investigating what was going on. “What horrified me was that the problem got worse and worse and worse.”
Along with other concerned individuals in the sector, Braben set about finding a solution. After looking at various ideas, in 2008, a group of six people from the IT industry and from academia set up the Raspberry Pi Foundation. The charity set out to make a tiny and cheap computer that enables people to teach themselves the basics of computing.
The Raspberry Pi is a credit-card sized computer board that plugs into a TV and a keyboard. It was an instant success when it went on sale in February 2012, with prices starting at not much more than £20. A round of media appearances by Braben provoked a rush that crashed the websites of the companies selling the Raspberry Pi. The first batch sold out “within seconds”, with over a million units sold in the first year. Manufacture of the Raspberry Pi is now based in the UK, with production at Sony’s Pencoed factory in South Wales passing the one million mark in the UK in October 2013, and also, at 1.75 million units sold, outselling the BBC Micro.
Has Raspberry Pi really done anything for computer education? “It is getting better,” says Braben. “I like to think that Raspberry Pi has created visibility and has made the subject a bit more fashionable.”
The rendering to get the lighting effects and material textures on the spaceship ‘Anaconda’ will establish a new level of realism for the Elite series of video games. Elite: Dangerous is due out in 2014 © Frontier Developments
Braben accepts that it will take time to reshape the teaching of computer science in schools but is optimistic that this will happen. For a start, the government is listening at last.
David Braben was involved with the Royal Academy of Engineering’s informal Fellows’ computing group which supported the joint proposal with BCS to create a new Programme of Study for Computing among 5-16 year olds as part of the National Curriculum Review. They presented their findings to the Department for Education in 2012, and most of their recommendations will commence for the next school year, starting in September 2014. The national curriculum for computing aims to ensure that all pupils can understand and apply the fundamental principles and concepts of computer science, including abstraction, logic, algorithms and data representation. It will also ensure that students have repeated practical experience of writing computer programs in order to solve computational problems. The Academy Fellows’ computing group has also led to the development of a low-cost programmable robot for teaching computing in primary schools.
Braben suggests that we think of education as a staircase. “What has happened in the last 10 or 15 years is that the first three or four steps had been demolished, with the steps higher up, at the universities, still there,” he adds. Raspberry Pi is just one the bits needed to rebuild those first steps.
He is also on the campaign trail for computer education and training further up the staircase. He is on the games board at BAFTA, Creative Skillset and various university advisory boards. He is particularly keen to see that, when universities offer degree courses in ‘computer games’, they teach something that would have the potential to qualify graduates for a job in the business.
Instead, too many degrees teach basic programming that would not prepare students for demanding tasks in one of the more complex sectors of IT. Braben says: “I don’t like the games industry being used by academia as a way of selling easier-to-teach degree courses.” There has been success here too. “We have now got quite a few courses that are approved and that are reasonable, but they are not easy courses.” Look for Creative Skillset degree accreditation, is Braben’s advice.
David Braben holding the Raspberry Pi, a single-board computer roughly the size of a smartphone. It is equipped with an ARM processor that runs Linux, and USB ports that let users attach a keyboard and mouse. Video is also supported via HDMI and RCA with storage handled by SD cards
This year, Braben participated in the flotation of Frontier Developments, where he is still the majority shareholder, on the AIM market of the London Stock Exchange. This raised nearly £4 million and valued the company at around £40 million.
Then there is the work on the next version of the company’s games, with Elite: Dangerous due out in 2014. With its emphasis on space travel, the game lets Braben continue to develop his interests in astronomy and physics, and to explore elements of space, time and gravity.
“I love the effects of low gravity and what that means for architecture. For example, most structures in space are actually spun, so they still obey the rules of physics,” Braben said. It’s a theme he returned to in a talk earlier this year at the TEDxAlbertopolis event in London, where he explained procedural generation, the sets of algorithmic rules using randomised strings of numbers to create realistic and convincing game effects, or computer-based graphics tools. “There’s a lot of physics in an airbrush, the viscosity of the paint, the speed of the motor, where the artist holds it away from the paper and it creates a beautiful pattern within an artist’s control...the designers of that art package will have mimicked as best they can what an airbrush does,” he said. “Inside the computer it simulates that science...now imagine doing that with clouds, or a bit of hillside.”
For Elite, Braben and Bell had to build layers and layers of rules to create galaxies, governments and economies. They also needed to create a set of rules for naming new constellations, so had to look to linguistics. “I remember looking at it and trying to get names that were pronounceable, so you wouldn’t get phlegm on the screen because that’s what came out – essentially things called phonemes.” Procedural generation was useful, too, for saving scarce memory in early computers.
The Raspberry Pi has several times the processing power of the BBC Micro or the Acorn Electron, so memory limits are less of an engineering challenge for today’s young programmers. Given time, some of them may learn to code the kinds of complex rules-based structures that will lead to Frontier Developments’ next blockbuster game.