Roland Perry isn’t a game developer, but he’s still helped to influence the games industry. He was critical in the creation of Lord Alan Sugar’s range of Amstrad CPC computers, while his name was given to the many Roland games that appeared on the 8-bit range. Here he talks to us about working with Lord Sugar and how Amstrad brought the CPC 464 to market.
People obviously know Lord Alan Sugar, but your name is integral to the CPC in more ways than one, isn’t it?
Well, I was involved in the technical design of the computers…
But you also had games named after you, the character of Roland, and the CPC’s nickname of Arnold was an anagram of your first name, wasn’t it?
Strangely enough, I don’t think it was. There was a degree of secrecy in the early days of producing the CPC. We felt that if people knew Amstrad was producing a games computer, we may end up with a kind of situation that would be difficult to handle with lots of people turning up and asking you when it’s going to be in the shops, and what’s it going to be like and all the rest of it. So all the absolutely new products were designed pretty much in secrecy. When I was trying to get contractors and whatever on board, it was always felt it was useful to keep a level of secrecy until they were signed. I wasn’t allowed to tell people who the client was so we had to come up with a codename. And that was Arnold.
So it wasn’t at all to do with rearranging the letters of your first name then?
It was more an allusion to Arnold Weinstock, who at the time was running GEC [a British engineering giant]. So there was this kind of forced trail leading back to parts of the GEC empire because people knew it was a UK-based electronics company. Even in those days there weren’t that many to choose from, so it was just a bit of a smokescreen.
Did the contractors not ask many questions?
The individuals I was talking to knew me personally because I was at university with some of them and things like that, but they didn’t know who my client was.
You didn’t get anyone from GEC coming on and saying, ‘Hang on, everyone thinks we’re making a computer here’?
Well we didn’t, but one very strange thing happened. Somewhere along the line, Reader’s Digest got hold of it and published in an article that Amstrad was a subsidiary of GEC, which was very odd. This is after the CPC had all come out and they were publishing either a history of GEC, Amstrad or the electronics industry. But I was completely confused. And that kind of mistake lasted for decades. People would go back and quote that as a kind of authoritative source. It was very strange. These kind of urban myths develop, but there you go.
A picture of Roland during his Amstrad days.
Did Lord Alan ever see that?
He saw everything, really. I’m sure he must have done.
How did Alan Sugar approach the project? Did he say to you ‘Look, we want this machine. It needs to have the keyboard with an integrated cassette drive and one plug’?
He was very much involved himself, yes. He was integral to the concept stage – the fact that the tape drive was included, the fact it has an extended keyboard, the shape of the box, the type of connectors on the back… it had a printer extension port and a CPU bus expansion port on it. All that kind of stuff is what I call the concept; the fact the power supply was in the monitor and so on and so forth. Lord Sugar put that out to an initial designer who came up with a prototype that didn’t work properly, so he transferred the project to ourselves.
What was the CPC like when you first saw it?
The physical design was completely unchanged from the moment we first set eyes on it to when it was in the shops, apart from minor details. All the keys on the keyboard of the first prototype were grey rather than those kind of primary coloured ones you got eventually, and also the plastic of the case was grey rather than the matt black. Apart from that it was absolutely the same – the same moulds and everything. That included inside it the footprint of the circuit board, the position and number of pins and so on of all the connectors. So that was all we had to work with and we basically designed the circuit board around that rather tightly defined footprint.
So why were little quirks like the colours of the keys introduced?
Well, that is another aspect of design. I mean, design is one of these words that has lots of different meanings. I used to get frustrated in later years when people said “What do you do?” and I said “Oh, I design Amstrad computers” and they used to say “Oh, why do you make them that horrid beige colour then?” I said “No, no, that’s not the design that I did. That’s production engineering and artistic design. My design is the electronics and the software and things that go inside them. The concept – so I’m using that as a kind of synonym for design – was that it should look good and therefore have a proper keyboard, a proper numeric pad, and a built-in cassette, and that’s what Amstrad had already achieved by the time we saw the project.
An advert for Amstrad’s range of 8-bit systems.
The CPC was seen as a serious type of computer, compared to say the Spectrum or the Commodore 64…
I think the Commodore 64 was the closest machine in its class at the time and was the reason why, for example, we went for a 64k of memory rather than 32 or 48, which some of the machines at the time had.
Obviously, Lord Sugar wouldn’t have been completely au fait with the electronics side of things, so were you given free reign on what you actually put into the machine?
Well, we had some rules. We also effectively had a budget for the number of components that we could use, and one of the rules was it had to have a 6502 processor in it. That was a bit unfortunate because the hardware and software designers that I found, who were streets ahead of any others that I found, both said, “Well, we’ll do this job but only if you can change it to a Z80.” So we had to go back and say, “Is it all right if we change the processor to a Z80? And the reason we’re doing it is because the designer is more familiar with it.
There’s just as many particularly indigenous UK games been written for the Z80 and the Spectrum as there has 6502 and the Apple/Commodore.” And so he says, “All right then, change that.” And then it was simply a case of building the design around 64k of memory and a Z80 processor. It had to have sound, obviously. And for that class of computer, it’s relatively straightforward what electronics it needs to have in it. It needs to have a CRT controller, it needs to have a joystick parallel port interface chip, it has to have a memory controller, it has to have a keyboard controller, and they’re all fairly well-known design elements. We ended up putting quite a lot of them into a custom chip rather than having separate components. I’ve got a picture framed on my wall of the circuit board. It’s not a very densely packed circuit board.
So was the idea to beat the 64, the C64 and the Spectrum, or to be on a par with it?
I think we were trying to achieve something that was pretty equivalent really in terms of features and processing power, because again they’ve all effectively got the same CPU in it. The fact that a Z80 has got a 4mb processor and 6502 has a 1mb processor is a bit of an illusion, because the Z80 just divided that down for most purposes. So you’ve got the same amount of horsepower and you’ve got the same amount of memory; a printer port is a printer port and a keyboard is a keyboard. There’s a limited amount you can do to change it. But the things we did put some effort into were arranging for it to have both a high-resolution screen mode and a multicolour screen mode. We had 27 colours rather than 16, which turned out to be trivially easy to do. Some designs coming out then only had eight colours, which is no good at all.
One of the big differences between computers now and back then was the prompt you got when you turned the machine on. The CPC just declared “Ready”. And there was no help for the first-timer at all…
There wasn’t, but the point was that people had to be able to switch the machine on and for it to say something on the screen and not be just a cursor blinking away. It had to say something recognisable and encouraging, something warm on the screen. It had to have a cursor that moved left and right, and up and down when you pressed those cursor keys, which again was unusual for a machine of that time because most of them would only let you go left and right and press enter. But why would you have cursor keys on the keyboard if they didn’t move the cursor around the screen?
Did you design the machine for games?
I think it was aimed at allowing people pretty much the same sort of range of capabilities as something like BBC Micro. In other words, you could do serious programming on it if you wanted to. You could dabble in programming – maybe play a tune, then wiggle some lines on the screen and flash some colours. Or a professional games writer could write a reasonable game that was as good as the games on any other machines around at the time. It did just a bit of everything.
But games were certainly important. It’s why Amstrad launched Amsoft and looked for games to publish, isn’t it?
We were asked to have a complete range of games ready on day one. That was my job and some colleagues’. What we did was trawl around existing software houses for Spectrum and BBC Micro and whatever, and we said, ‘We’ve got this new product coming out. Would you like to convert your games to it?’ because in those days people weren’t writing games from scratch for the 464. They’d always be converting an existing game.
Was it an easy task?
We did have to convince developers. Some had been bitten in the past by other firms who had asked them into their offices to talk to them about their computers, convinced them that they really were going to make the machine and then nothing would come out the other end into the shops.
We knew that some developers had wasted their time with other firms and that when we arrived they would have heard it all ten times before from budding computer manufacturers. So we had to give them a pitch that we really were going to produce this and we really did need the games, and so on and so forth. It helped enormously that a well-established company was doing it and not a start-up.
Did you hand out prototypes?
Our first job when we designed the motherboard was to hand-build around about 40 prototypes, and then we gave one of these to each of the software writers so that they could write the software for it. This was something that nobody else had ever done before. They’d either just give people a paper specification to write to or they’d lend them the prototype for a couple of days or something. It was completely unheard of for somebody to turn up with one of these things under their arm and say ‘There you go, you can keep it for a couple of months while you finish your game off.’ The logistics of manufacturing and then distributing those prototypes was probably the biggest part of the project as far as I was concerned.
Notable Roland Perry Hardware
You can read the rest of our interview with Roland Perry in issue 78. Buy it now from GreatDigitalMags.com
Retro Gamer magazine and bookazines are available in print from ImagineShop