STONEHENGE IN SPACE

With two major hits to his record, Tau Ceti and Academy, PETE COOKE remains an enigmatic figure. RICHARD EDDY went to talk to him on the eve of Micronaut One being released by Nexus, and discovered same facts behind the man.

When Tau Ceti burst onto our screens in December 1985, most people had never heard of its programmer. We all tend to think stars are born with a hit on their hands: and for some strange reason CRASH credited the program to a Steve Cooke! (It was a Smash in the December 1985 issue.)

But in fact Pete Cooke’s pedigree goes way back to 82/83. Hearing that he as about to complete a new game for Nexus called Micronaut One, I decided to visit the man at his home in Leicester and find out more about his work.

Pete Cooke’s house is a modest two up/two down in a long terrace in the Highfields district. His programming area is an upper room, crammed with computers and synthesizers (he once played keyboards in a band). And it was here on a hot sunny day toward the end of April that we sat down to talk, together with Ian Ellery from Nexus, who’d popped over to see how Micronaut One was coming along.

It can be fascinating the way programmers who work on their own arrive at a game design — which arrives first, the chicken or the egg? For instance, where did Tau Ceti come from?

‘It came from a technical idea, I suppose,’ Pete began hesitantly. ‘I’d seen Gyron, and for along while it had puzzled me how on earth they did the spheres. It obviously couldn’t be a sprite because they didn’t have enough memory to have that many sprites.

‘I finally twigged that they must have used a table of line widths, and I thought about it a bit, then realised I could split it three quarters and a quarter, and then it would look like a shadow.

‘I just went away and played with that for quite a long while and got it so that I had the sun in the sky and the shadow in the right place. And it sort of came from that, really.’

Ian, who was Creative Director at CRL in Tau Ceti days, remembers his first sight of the hit: ‘Tau Ceti was this funny little square that arrived one day with a horizon and some objects — that’s all there was — just a little square in the middle of the screen and it had a line in the middle and some little boxes.’

It’s not always a routine that sparks off Pete Cooke’s imagination, though he works from a routine back to a game as often as the other way round.

I generally start by thinking technically what I would like to see on the screen, what would look nice that hasn’t all been done before. You think whether it’s achievable, technically how will you do it and how could you plan it into a game? If you can’t, then obviously it’s a demo and you forget it — or it pops up later on when you’ve worked out how you can use it.’

With Tau Ceti the light-and-shade routine worked out, but obviously there’s a lot more to the game than light and shade. What about the setting, and the little histories?

‘It was all made up as I went along. I know a bit about astronomy, so I looked around for a likely location — because the way the routine works meant it couldn’t be in space; you couldn’t have the sun above because the line-draw routine wouldn’t work, it would have to be on the surface of a planet.

‘And I thought ‘Right, it can’t be Earth because it looks a bit barren for Earth, it’s got to be another planet.’ So I looked round for nearby stars that might be inhabited and Tau Ceti had a nice-sounding name.’

The display looked rather science fictional, with the controls and the flashing lights — how did it come about?

‘An awful lot was determined by technical things, really. I thought ‘How fast am I going to be able to do this?’ Because filling solid areas of the screen takes an enormous amount of time.’

Pete came up with a clever cheat. The problem was sorting lots of objects into order of distance first, updating their positions each frame, and doing it quick enough to give an illusion of movement. Using a full screen, or even half a screen, proved to be too slow and horrible to play. The solution was for the display to take up about a third of the screen.

But, as Pete says, ‘that brings you to the question of what goes elsewhere, and the obvious thing at that point is to say that you need some other information in the game. Then you say ‘Well, the game has got to use these charts’ — but they’ve got to mean something if they’re filling up half the display area, so you have a lot of things going on that you need the charts for... and using the charts doesn’t take anything like as much time as it would to copy the top two thirds of the main display down.’

So technical restraints actually helped Pete design much of the gameplay in Tau Ceti. But with the follow-up, Academy, Pete seemed to have overcome even those restraints.

‘It used a lot of similar routines, but essentially I rewrote it all, because it’s amazing how much better you can do things in a year. For instance, in Tau Ceti you’ve got the list of objects and, because there’s a lot of things they can do, I needed 20 bytes for each one to tell it various flags about whether it was running away, whether it was shot, whether it had a door and so on.

‘At that point it started to become fixed — I had all the routines to step through it and sort it, and at the end I found I’d left a byte unused. Well, you just say ‘Sod it’, it’s just not worth it; to times it by 20 is not too difficult (times by four and add one and then times by four again, which is all quite easy SHIFTs and ANDs), but to times it by 19 is actually slower, so it’s better to be 20 long.

‘I wanted it to be 16, just four SHIFTs, but given that it wouldn’t fit, it was worth throwing that space away. When you start you don’t know what it’s going to need at the end, you’re always guesstimating along the way — ‘How much space is that going to take?’ You think about how much memory there is and think ‘Well, the code is going to be roughly this much and the data has got to fit in that much’.

‘Continually, as you’re working along, you’re thinking ‘Maybe I can compress a bit of it somewhere’, or ‘I’m going to have to rewrite the code to save some space’. Second time round, of course, you’ve got the hindsight of having done it and knowing that it fitted in this much space. It saves a lot of fiddling, so that speeds the game up.

Was that why Academy turned out much more of an arcade game than an adventure?

‘Yes. I mean, the code’s more efficient. There’s a couple of basic things I did that speeded it up an awful lot that I should have done in Tau Ceti. I spent a lot of time optimising Ceti routines, but it still scanned every object, every frame, to see whether to draw it.

‘But I flagged each one based on a field of view, so if you flag each one in that area, once each pass-round; then you know it couldn’t possibly get from there to there in the 15 times round, and you’ve eliminated most of the processing straight away. It’s immediately 60 per cent faster for that.’

With the experience of Tau Ceti and Academy behind him, Pete might have drawn heavily on the same routines for his next game, Micronaut One. But...

‘No, basically I ditched the lot. It works in a totally different way. It got to the point where I thought I could do another ‘safe’ game, but it gets less challenging after a certain point.

‘And there was the move to Nexus — I thought it would be better, having moved, to do something different. It wouldn’t annoy CRL as much, for certain; and it would look a lot better, and yeah, I was bored with doing another Ceti game.

‘In fact, I’d got another one sketched out which was going to be a bit like Psytron, in that you were defending something. But I didn’t think I could make it different enough using the same sort of display, and I’d long wanted to do something with tunnels, cos there’s been a lot of things attempted but they haven’t quite worked, right back from New Gen’s 3D Tunnel.

‘There’s I, Of The Mask, which again looks stunning, but actually it was just a set of frames — it must have been because you couldn’t actually do anything lathe tunnels but go forward and back.’

Did Realtime’s Starstrike II help?

‘It gave me a few ideas, but it suffered awfully from being a subsection in a larger game. I looked at it and thought ‘The big problem is junctions’, and that’s what none of them had, cos you can’t do them that way. Drawing a tunnel is a doddle — it’s four lines converging to a vanishing point, and you just move it around. But branching tunnels is totally different, and I spent ages thinking about it.’

Needless to say, Pete cracked it. Micronaut One’s routine has a list for the tunnels; it starts in the section you’re in and says ‘that’s the visible window at that tunnel’. Then, using what he describes as a ‘recursive descent’, it stacks that as the visible window of the screen and looks to the next two sections, gets two visible windows, and down it goes. By the time it gets to the far section it’s got a tiny window and then it clips all the lines to that window.

‘Effectively what happens,’ says Pete, ‘is that the program is doing hidden lines, but not having to actually draw solid surfaces — so it’s just lines, which means it can be very fast. I thought Starstrike II was too slow and too simple, and stuff like 3D Tunnel with attributes just didn’t work, so it had to be fast, and the way to do that isn’t to draw solid shapes all over the screen, but to find some way of working out which areas you can see and which you can’t, and just clip the line to it.

‘It looks about four sections ahead and then it simply stops. After about two sections it stops drawing the floor because that looks cluttered and horrible, you just get a black mess.’

At this point we paused, partly to look at Micronaut One (for last month’s CRASH preview), and partly to go for lunch at the local Indian. Over the poppadoms I asked Pete about his background, and how he got into computing.

Is there a dreaded ZX81 skeleton in his cupboard? Laughing, Pete admits it all.

‘I used to be a maths teacher you see. The whole thing started when I was teaching — I got hold of a ZX81 in 1981 and then a Spectrum and took it up as a hobby. I ran a little club for various of the kids sort of 14, 15, who were dead keen, and I’d bring stuff along I’d done and say ‘Have a look’. Eventually some of them turned round and told me I ought to sell them. They just hassled me till I sent one off, and it got snapped up and from then on...’

Pete knew very little about programming — just what he’d picked up doing his maths degree. He played around with BASIC, and started learning machine-code routines. But what was that first game he sent off?

Invincible Island.’

Invincible Island? The Richard Shepherd adventure? So long ago! And indeed CRASH Issue One’s LIVING GUIDE does credit Pete Cooke with the game’s authorship.

He followed it up with another adventure for Shepherd, the well-loved Urban Upstart.

Was he well paid for those early efforts?

‘Yes, I think I was. Comparatively, it hasn’t gone up a lot since then! I got a lot from Urban Upstart — it sold a lot and it kept selling. But after that the market bottomed out for adventures, though I did a couple more for Shepherd... one was called Inferno, based on Dante’s Inferno.

‘It was interesting, but it was a bad project to get into, it didn’t work as I had wanted it to. The constraints of the book were too much in the end. For an adventure it became very linear.

‘The other I wrote on the 64, Up A Gum Tree. It was the best adventure I did.’

He’d packed up teaching to become a full-time programmer — but, having earned lots, he spent six months on Up A Gum Tree and got just £500 for it.

‘I thought I’d better look round for a job. I gave up work at exactly the wrong point — when the real drop-out came.’

But he did another project with Richard Shepherd — Ski Star 2000.

Fighter Pilot had been out long before then, but Ski Star was the first 3-D vector-graphics game that wasn’t a Battlezone or a flight simulator. You could design the courses and pick things up and drag them with this cursor stuff — I think it was one of the first games to use icons as well. I scrapped icons after that — they’re terrible.

‘I’ve played too many games where it’s ‘What’s this icon doing then?’ In the end language is the best way of communicating we’ve got, so why don’t we use it? I’d like to write a game with no words in it, something that’s so transparent you don’t have to use language at all.

‘But I don’t think that the in-between ground works — you have a weird mishmash of these wiggly, squiggly symbols. They’ve put me of the Atari ST a bit. I keep thinking I’d like to get rid of this and get a good on-line interpreter back where you can do things quickly rather than having to wander round endlessly.

After Ski Star 2000, Richard Shepherd went under and disappeared from the software world, leaving Pete as a struggling freelance without a software house to market his games. Enter Ian Ellery and the company he then worked for — CRL. Ian takes up the story...

‘I got this little cassette,’ he recalls, ‘through the post with this little letter (adopting kiddie’s reading voice) — ‘Dear CRL, I’ve written this game called Jackknife, which is a lorry-driving simulation. From Pete Cooke’ — and I expected ‘aged 13 3/4’ after that. The first one didn’t load. He sent me another and I still thought it was a kid sending me a program. I loaded it up and I saw what was a very original game — I wasn’t sure how commercial it was, but it was still very playable.’

The game was, of course, renamed Juggernaut — presumably a more commercial title.

How did Pete come up with the weird idea?

‘I had a mate,’ he says, ‘a mechanic-cum-car buff, who wanted to be a long-distance lorry driver. He saved up for this £100-a-week course, went away to Birmingham on it, and came back having failed it. The problem was reversing — it’s impossible.

‘I thought about it and played with little models, and it is a bit strange — when you push it one way it goes the other and it just buckles up.

‘There seemed to be an idea in that and it gradually metamorphosed into a game all because he said how difficult it was to reverse an articulated lorry. I put loads of stuff in about ferrying freight around so you didn’t have to do too much reversing. But that was essentially it.’

Juggernaut wasn’t a commercial hit, but it covered its costs sufficiently to encourage an Amstrad conversion which Pete did himself, and fire CRL’s enthusiasm for Tau Ceti.

In between Tau Ceti and Academy, Pete did Room 10, a Ping-Pong game, on the Amstrad.

‘I wanted CRL to put it out as a budget game and they wouldn’t. I said it ought to be around £4 cos that’s what I thought it was worth. I did it on the Amstrad because it was a chance to use the colour display — you can have a coloured floor and a coloured ball over it without attribute problems. It’s the simplest game there is.’

SOFTOGRAPHY

INVINCIBLE ISLAND 1983 Richard Shepherd

A pre-CRASH text adventure set amongst a primitive tribe which may be hiding treasure.

URBAN UPSTART 1984 Richard Shepherd

Reviewed CRASH Issue 3 (April 84)
Street cred text adventure with graphics — get out of horrid Scarthorpe and live.

INFERNO 1984 Richard Shepherd

Reviewed CRASH Issue 8 (September 84)
Adventure based on the book by Dante. Every location had a simple graphic.

UP A GUM TREE 1984 Richard Shepherd

Commodore 64 adventure

SKI STAR 2000 1985 Richard Shepherd

Reviewed CRASH Issue 14 (March 85)
Debut appearance of icons for the piste designer in this first non-Battlezone vector graphics simulation.

JUGGERNAUT 1985 CRL

Reviewed CRASH Issue 19 (August 85)
Almost a simulation, as you truck heavy goods on long articulated lorries through narrow town streets.

TAU CETI 1985 CRL

Reviewed CRASH Issue 23 (December 85)
A Smash for this complex 3D arcade-strategy game set on a hostile planet.

ROOM 10 1986 CRL

3D Ping Pong game, which only came out for the Amstrad.

ACADEMY 1987 CRL

Reviewed CRASH Issue 36 (Christmas Special 86)
Another Smash with the follow-up to Tau Ceti

MICRONAUT ONE 1987 Nexus

Reviewed this issue.

Finally, the time had come to ask the Big Question — why did Pete move from CRL to Nexus?

‘I like Clement Chambers, the boss of CRL,’ says Pete carefully. ‘He’s a nice bloke and I get on well with him, but the reason I joined CRL was an advert that Ian Ellery put out, and the thing that worked at CRL was me working with Ian.

‘Essentially it wasn’t much to do with Clement. Clement would come in and ask how things were, and he’d be pleased when the product turned up. But it wasn’t Clement who’d seen the demos and seen what the potential was.

‘When Ian left CRL I thought, ‘What I’ve got is a fairly good relationship with Clement, but it isn’t a working relationship’. Obviously Ian, who had gone to Nexus, was interested in working with me again as things had worked out alright before.’

Perhaps Ian enjoys working with Pete because he delivers programs on time...

‘I’m always late, but never terribly late,’ Pete replies, laughing.

‘If there was ever a wait it was worth it,’ adds Ian agreeably.

‘No, I’ve never been six months late but then you can’t when you’re freelance because there’s no wage, you know — if it’s late I don’t eat!’

All Pete Cooke’s games have been original, developed from ideas of his own. Has he ever considered doing an arcade licensed conversion, or working within a team?

‘No, I’m not interested. Well, I can’t say that about any arcade licence, but I can’t think of any offhand. I don’t like the philosophy behind them — the philosophy of the quick buck — and games I play are not like that. My favourite game to play is Revs, and it took me six months to crack it. Now that isn’t going to appear on an arcade machine, is it?

‘As for programming teams on such projects, I think they’re actually graphic designers, and looking at the games I don’t see any programming in them. When I looked at Jack The Nipper I thought ‘This looks really nice, very pretty’, and then I saw it moving and they all move in diagonals and I thought, ‘Oh no, why?’ It’s a trivial thing to have a look-up table and a shallow diagonal, so it looks realistic rather than bouncing off walls.’

‘I wouldn’t work with another programmer, but a graphics designer — I can’t draw so somebody else has to do it, and that’s Ian.’

And what’s next after Micronaut One?

‘There are three ideas waiting in the queue. One’s about nine months long. I’ve got a routine that’ll do solid 3-D. If you say ‘Let’s remove the need to do updates ten times a second’ — say something like Lords of Midnight, where the frame rate is slow but very effective — then you could do an incredibly detailed 3-D display with solid surfaces, and you could move around it. I suppose it could be very interesting as a sort of graphic adventure. Again in first-person perspective.

‘Second on my list, I suppose, is converting Micronaut One for the Atari ST, which Nexus really want me to do. I’m looking forward to learning the 68000.

‘And the third one is quite a simple idea — I don’t know whether I should give it away. It’s a shoot-’em-up. There’s a lot of parallax scrolling games — this is the ultimate — perspective scrolling. You can have shadows — you’ve seen pictures of Stonehenge at sunset — imagine shadows of rocks and boulders actually changing like that as you go by, and just a spaceship over the top with a shadow, zapping things. It’s something simple like that.’

‘To be written in coffee breaks!’ Ian retorts.

‘It baffles me that there’s so much stuff around that you’ve seen before,’ muses Pete. ‘I’ve got millions and millions of game ideas floating around waiting for somebody to land on. I’ve never found it difficult thinking of ideas — it’s more difficult limiting them. Picking and choosing, I suppose.’

Search for more information