Interview with Andrew Braybrook

Steve left his stable job in the middle of 1983 and invited me to work for him in September that year. It was a Friday evening and we usually met at my house before going out for a drink. He knew I was not happy in my job and the "interview" was 3 sentences.

"Do you want to come and work for me?"


"OK, let's go for a beer".

We did indeed occupy Steve's dining room. Having both worked for big companies we continued in that way. I walked to his house and we worked from 9 to 5. We kept to office hours, which was quite different from a lot of programmers at the time, who tend to be nocturnal.

Yes. In the early days there weren't many tools around for programming or graphics, or sound, so anything we needed we wrote. I did manage to buy a multi-pass assembler for the Dragon 32 and that helped tremendously, I could write my conversions of Steve's Spectrum programs very quickly. I wrote a graphics designer in Basic to be able to draw the graphics. I remember writing it secretly as it took me a few days and Steve was expecting me to be writing the game code. I thought it would speed up development and allow me to draw my own graphics easier. Up to then we were drawing on squared paper and converting the images into hexadecimal by hand. It did speed up graphics design no end.

Later on the Commodore 64 I was using the official Commodore assembler and a 5.25 inch floppy disk drive that was painfully slow. Assembling the game would take half an hour so we had a second C64 to be designing the graphics while waiting for the compilation. We used existing tools for the graphics.

After a year or more of that we got early amber-screen PCs and used editors and assemblers on the PCs and then transmitted the code to the C64 or Spectrum via parallel link cables that we made ourselves.

During those times we were doing whatever we wanted to do. We weren't getting advances and the royalties from each game were paying for the next one to be created. Steve and I were quite competitive with each other and also with other games coming out at the time. I always wanted to do something new and better in each new game. I think that the general standard of games at the time was quite variable and it was our experience in professional environments that allowed us to improve the quality.

The games we designed were fitted around the limitations, we were designing for the hardware. Graphically the early machines were frustrating but the C64 was less limiting at the time than some. We developed ways of reusing the 8 hardware sprites to get more objects moving, for example. The machines were what they were. It was good to meet other programmers and see other games to see what tricks they were getting up to. I think mostly it encouraged creativity, it was only when they started converting the newer 3D games coming into the arcades in the late 80's that the requirements outstripped the capabilities of the hardware.

My move to 16-bit was maybe a bit late. It was around 1988 and I was developing a new game on the C64. Graftgold was starting to develop for the Atari ST. The development software was quite similar to the 8-bit systems as we were using PCs to write the code on still. The break came when we got the deal to do the conversion of Taito's Rainbow Islands arcade machine to ST and Amiga. I was keen to get started and my C64 development was shelved. I had been studying the 68000 assembler and we had the Amiga hardware books, but at that time the Atari ST was outselling the Amiga and we were leading the development on the Atari. We all knew the Amiga was better for us but we had to wait.

This was Dominic's 3D game. He didn't quite get it finished before he left. I think my role in that was mainly testing the game, I don't remember writing anything for it. Steve did a lot of work on it at the end.

3D was something Dominic specifically wanted to do. We had simulated 3D in our earliest works but had gone more in the direction of using the hardware after that, which was usually 2D and character-set based. We didn't revisit 3D again until the hardware caught up.

Dominic was keen to try out some of the latest software techniques, and he was doing two jobs, he was writing a 3D game and he was developing a Core system to enable us to write games for both the Atari ST and Amiga, with minimal game code changes. That approach was OK for a 3D game because the scrolling and sprite hardware features on the Amiga were not much help for 3D games, and even for Rainbow Islands with only vertical scrolling it was OK. The system he wrote was Object Oriented, more widely used today it was a bit of a pain then and I'm still convinced it's too idealised and a pain now! Nevertheless all credit to DR, it worked and we could implement an Amiga version of our ST games in about 3 weeks. I was still keen to use Amiga features but as that really would alter the game design beyond what the Atari could do then at this time it was just not commercially viable. We were by this time a company of 6 or 7 and in an office, with publisher's advances, and we were less able to do what we wanted.

We were very fortunate that we were sent a huge folder of game design documents from Taito. Some of it had been modified slightly by the time they had finished the game, but I was amazed at how much design was done up front. We did less design and developed the game more as we went along, and then documented it at the end. So the conversion job here was to get the most accurate copy we could. We had all the algorithms for the bonuses, the speeds of the meanies, all the behaviour of the main character and the rainbows.

BT sent us a full size arcade machine with the game in it. The first thing we did was play it a lot, all in the course of development. Our best arcade game player by a mile was David O'Connor. I recall that one evening we set up a camera and filmed him playing the game all the way through. There are 3 different endings so he had to play the game to each one. I think it just about fitted on a 3 hour tape. That tape allowed John Cumming to create all of the levels' backgrounds accurately. He wrote an editor and we wrote a compressor to pack the data down. I remember that he found some mistakes in the backgrounds and we had some discussion as to whether to fix them. I reasoned that they would have fixed them in the original if they had noticed or had time, and I wanted our version to be as perfect as possible, and in that respect slightly less authentic, I suppose. The mistakes were mainly missing shadows or slightly badly matching patterns, nothing major. I won.

We were quite lucky that the arcade machine had a 16 colour background, the same as the ST and Amiga. It also had a lot of 16 colour hardware sprites, we had none. So John's biggest triumph was to come up with a palette of 16 colours that would let us copy all of the backgrounds and sprites. He then had to adapt the graphics into his palette. We had some sheets of graphics from Taito but they were not complete, and we didn't have enough knowledge to capture the graphics from the hardware, so it was back to the video tape.

We were concerned that we wouldn't have enough CPU time to handle the number of sprites needed in later levels, so we started off shrinking some of the graphics slightly, mainly to make them no more than 32 pixels wide. That way we could render them more quickly, but later on we got more confident and preserved the sizes of the original graphics.

Memory was starting to get tight and the levels were getting bigger, certainly. We had all the graphics for the 3 bonus islands coming out of the water, and the code to run them. When we started the job we had costed it out for the only 7 islands we knew of. So when we found out about the extra 3 we had a problem. They were big levels, probably double the work, and also the background palette completely changed on island 9. So we decided to go with 7 levels, and organise the 3 endings into 2.

It took a long time for the game to actually come out at all. Probably it wouldn't have come out at all if it hadn't been for the magazines. Gary Penn had seen the game and I think was instrumental in banging a few heads together and getting the game released. We always enjoyed promoting the games and took the magazine reviews very seriously. It was nice to be recognised, and the game was a true team effort.

We were keen to update some of the older successful games so I set about building a demo. Horizontal scrolling on the Atari ST was going to be difficult. It can be done, but requires limiting the background colours, or using hideous amounts of memory to store pre-scrolled data, or by reducing the size of the scrolling screen. We implemented the backgrounds in less colours and had a demo running with all direction scrolling, which the Amiga could have done with ease, but we needed a solution for both formats. The limited number of colours were also not the look I wanted.

The only way to get all the colours involved was to use our Rainbow Islands vertical-only scrolling system. I worked out that the maps in the C64 version were not especially big, and by rotating them 90 degrees I didn't need to scroll horizontally and the game would use almost the full screen. It was a compromise but I thought the benefits far outweighed the loss. Amiga fans may have been less forgiving, so we all lost out.

By the time the Atari ST version was ready we may have had an opportunity to consider adding in horizontal scrolling for the Amiga, but the maps would have needed re-doing. These took a considerable amount of time each to create and there was no realistic way we were going to redo 7 ships. The conversion didn't take very long because of the Core system. I used the hardware sprites for the score overlays but that was about it.

I did consider rotating the maps back 90 degrees to make them wider and have to scroll, but it should be noted that if you have to scroll then you can't therefore see all of the map, and not having to scroll horizontally is an advantage, not a disadvantage.

Having discarded the 4 colour version of the original graphics I was free to start again from scratch. I wanted to add in some depth so I had bits of the scenery above the robots. We wanted real-looking robots rather than a computer display. We had a lot of difficulty designing the space ship walls. I ended up doing them myself. Probably the colour palette limited things a bit as we have 16 colours only and we wanted metallic graphics, which requires a lot of shading.

The ships were getting bigger as the game progressed, and I was quite conscious of the time taken to play through the game. I originally opted out the transfer game as a way of speeding up the testing. There was also a feature to try any of the ships on its own. Removing the transfer game made the game quicker to play and more arcadey, I also had made the transfer game a bit bigger and it took a bit longer to take that all in. Probably it wasn't my original suggestion but I am happy with that one.

Well observed!

There is a Graftgold key on each ship, some hidden, and they are collectible. Just like Rainbow Islands there is a secret level. If you collect all of the keys on the 6 ships it will reveal a 7th ship. It's a toughie though as it is the pirate ship. As you may have noticed, a feature of '90 is that pirates start beaming in after a time. This was originally a game mechanism to liven up the latter stages of a ship and indicate that you are taking too long. The last ship is full of pirates.

I think that by this time Dominic might have moved on so I had taken over looking after his OOPS Kernel. We had all the source code for the original arcade machine, in Z80, but I think we ended up modifying the original algorithms. Gary Foreman was the main man on the Amiga version.

We were all working in the same office so we all learned from each other. All the games we did were team efforts so I can only hope that the guys found me useful.

I don't think I had much involvement In Realms at all. Steve and I rarely worked on each other's 16-bit projects. Even in the 8-bit days we didn't do much on each other's games, other than Steve doing the music and sound effects for my games. We used to playtest each other's games towards their completion.

Steve does like playing strategy games and then dissecting them, especially if they cheat. So he was keen to try and build a game that was more immediate and accessible, and that doesn't cheat. Being fans of arcade-style games we wanted our games to be immediately playable.

I tend to prefer faster arcade games, I'm too impatient.

Fire and Ice started off as a son of Gribbly's Day Out. I liked the idea of a bouncy character. Initial demos to Renegade and their demos to others indicated that the bouncy character was too unconventional. We were persuaded to change him to a walking and sliding character. I wonder if that demo still exists?

I was really impressed by Turrican and in fact we had some technical discussions with the authors. They even got us invited to a show in Germany. We implemented their scrolling system, which was very clever and economical, but still couldn't run as many objects as we needed. The eventual game was more platformer than shoot-em-up. We actually had a lot of changes along the way with this one as my original design used the fire monsters from the last area in all the levels. It also transpired that the slippery ice level was one of the hardest to control and yet needed to be at the beginning.

Fire and Ice was the first Amiga-led game we did. There was never going to be an Atari ST version. A PC conversion did follow.

Some of the game philosophies and mechanisms from Rainbow Islands survived to become part of Fire and Ice. We never start from scratch. Every game we finished tended to get stripped down and all the good bits get saved and reused, or improved. Rainbow Islands had a very wacky Japanese feel to it which would have taken over the style of Fire & Ice had we used too much.

The puppies were interesting. Getting something behaving sensibly is always tricky. I'd have to say it wasn't as difficult as getting balls rolling around on the contoured backgrounds realistically. I think I should get the game out and have a play.

Our artists did a great job of adding some 16 colour backdrops for the CD32. These also worked on the Amiga 1200. The A500 didn't have the chipset to support the two layers in that many colours. The backdrops and extra objects all helped the look of the game. I did have a nicer sunset algorithm which would have worked nicely with the extra colour definitions on the A1200 and CD32, but the algorithm had been lost by then. It simulated a proper sunset rather than a coloured roller blind. So it would have been great to include the backdrops, but not technically possible, alas.

This was mostly Jason Page's work. I think I worked on some of the gangster cat algorithms, not much really. We did this for HMV, and had already done a similar little game for Powergen, involving a coal mine.

The game code was a stripped out version of Fire & Ice. The movement algorithms and background definitions were mostly there. We had a very tight time budget to get the job done, it would not have been possible to start from scratch. The guys managed all new graphics, movements and sounds. I think they had to get that done in less than 10 weeks.

The requirement was for a 1-2 minute game that could be played easily by lots of people. Probably that differs from both an arcade game and a computer game. I was amazed at the scores people were getting. We designed it and tested it a lot and we couldn't manage the scores that won the competition.

I think they had planned to do an Amiga version. They ported the C64 code over and tried to get the Atari ST to just run the game as it was, with no attempt to think how to get the Atari to do it. I think they did have an Amiga version but they hadn't cracked the smooth scrolling at that time. It certainly wasn't released in the UK.

There's no reason why the Amiga shouldn't have been able to take any 8-bit game and have a better 16-bit version, provided that the programmers think about the backgrounds. The 8-bit computers used 8 pixel wide character graphics and the 16-bit computers used 16 pixel wide blocks for the background, so we had to redesign the graphics to the strengths of the machine. It took us a while to get to grips with smooth scrolling a large bitmap without taking most of the CPU time. It was so rewarding when we got that working. That was all down to the generosity of the Turrican guys, they were really helpful.

The disappointment of the Mindscape conversions was definitely a motivation. Once we had the working scrolling system from Fire & Ice we wanted to push it further. We started off with a dual playfield version but found 7 colours too limiting, and with no pretty sky to add colour we reverted to a more conventional 32 colour screen, which ultimately expanded to 64 colours on the A1200. We had a CD32 version all ready to go as well.

I wanted to use the exact same control mode parameters, so that no-one could say one any version was better or worse. Then we just thought about what we could improve on. We added more weapons, dual player/ship modes, vertical scrolling was added. Anything the team could think if was considered. We also had a completely new set of graphics for each set of 4 ships. We had a fair amount of trouble getting a look that we could live with. Steve Rushbrook did a lot of the art work as he seemed more in tune with the original game.

The sub-game on the C64 seemed much too simplistic for the Amiga. I worked for some time on a top-down Paradroid-style system. The pilots would have to destroy the reactor. Unfortunately we ran short of graphics space and needed to change it, so we went for something a bit more immediate. Pity really, I wanted to show what Paradroid might have looked like on the Amiga. It did rather slow the game down though. To answer the question though, the C64 original was quick to play and right for the game at the time. I don't think it was enough for the Amiga.

From a publisher's point of view they wanted one product to market, it cut down on production costs. We did try to put in features that used the extra CPU speed available. The CD32 productions did allow us to produce proper A1200 versions, but games were taking 18 months to do, and with two or three graphics artists on the team that's a lot of work to do again with more colours. We did Uridium with proper shadows but again we had only a few weeks to get the A1200 and CD32 product ready. The market didn't support completely separate works. I think by the end of a development we would want to get on with something else.

Multi-player modes were something we had steered clear of. Running 2 players at once requires good design up-front. Uridium 2 supported alternating player turns as well as simultaneous. We did find it hard to play well in that mode, but never underestimate the game players. I did like the two-ships-at-once mode though. Fire & Ice wasn't really a 2 players at once game, nor was Paradroid. We did try the latter though, as we were asked to do a two-player version for the PC Engine. It did cause us to think a lot about how it was going to work, what with the possibility that two players could try to transfer to the same target.

Jose Doran was the lead programmer on the Amiga version. This game had what must be the maddest copper list ever. I was on hand offering moral support, I think I got involved in the goalkeeper algorithms. Jose did some of the graphics too. Lee Banyard was doing the sound by then, I reckon. Jose wanted to credit everyone in the company.

All Jose I reckon. He wanted to do a fun arcade game, steer away from the style of other games at the time, and really show off what the Amiga could do.

Yes it was. Jose also liked to do things that we all said were impossible. He would think out all these wacky effects and things to do, mostly by torturing the copper chip. Very original.

This was Iain Wallington's game. I was technical support. I did the virus algorithms and helped out with the end of level scenes.

I think we were thinking about both formats at the same time.

I don't remember much detail, sorry. I suspect we wanted to give the A1200 owners something extra. With more memory available we might have been able to go a bit madder.

It's nice to put a nod to other games in. A bit of Gribbly's Day Out is in Fire & Ice, for example. Graphics artists quite often come up with things like that. We usually go through a number of rounds of development before the game is finalised. Often things don't make it in because they don't work in the game, or they don't look right, so the graphics artists used to work hard and were quite competitive.

The fall of Commodore was very sad as it meant no new hardware was going to be produced, so you can only expect the market to shrink. I was getting involved in the Playstation version of Rainbow Islands and was using the Amiga version to prototype changes. So I had a gentle transition into the new toys.

You never get statistics on how many copies there are, we only know how many people bought the proper version of the games, and at the end it wasn't that many. Certainly not enough to justify the lengthy development time. We had advances to cover development costs, but we needed royalties beyond the advances to make a profit, and that wasn't really happening.

We had a constant turnover of programmers and new guys started on new platforms. We had good experience of working on consoles having done work on Megadrive and Master System, so getting used to new hardware wasn't too bad. We had to change languages from Assembler to C, that was probably the biggest transition. Some of the guys had already done C at university so it wasn't so bad.

It was rather difficult. We knew it was coming but it was tough to let go. In fact I signed on at the Job Centre virtually next door and went back to the office! Steve owned the building so there was clearing up to do so that he could let it out to someone else. I applied for some jobs and made contact with some agencies and they found my current employer, at which point I just sort of drifted off to a new job, which it turns out I got because games programmers do have a reputation for being quite bright.

8-bit was certainly suited to individuals, the machines weren't so big that you needed a whole team. I knew my graphics limitations and by the time we got to 16-bit I needed graphics artists. I could still manage the majority of the coding, only the sound and music was done by Jason. It was an enjoyable creative time, certainly. I can't see those sort of times coming round again, unfortunately, though maybe there are some smaller companies still doing creative work on the PC. I still see some good old-style games and I wish we had been able to do graphics like that 25 years ago.

Yes we did, for good or bad. By the time you finish a game you've pretty much exhausted your ideas in one particular genre and are keen to try something different. Sometimes you'll get ideas that need to be filed away for a different time and a different style.

Difficult one. I'm not unhappy with anything as there were reasons why each game is like it is. I was never pushed to let something go before I felt it was done. I am unhappy that we fixed the last bug in the Playstation version of Rainbow Islands at about 11pm on the last day after desperately searching for it only to find 3 months later that they released the previous version with the bug still in it. The end sequence crashes. Fire & Ice on the CD32 was particularly pretty.

My ode to Gribbly's Day Out became Fire & Ice. The CD32 version of Uridium 2 never saw the light of day. There was an Archimedes version of Gribbly's, BTW. The lead version of Empire Soccer was on the Nintendo, that was never submitted, so we lost out there. There was a Megadrive version of Fire & Ice, I'm fairly sure that never got released. The final Graftgold game was a 3D tank game on PC and Playstation, that never got finished. We tried.

I have a PC and a Nintendo Wii, so yes, I still play some games. Super-Swing Golf amuses me. Mostly I still prefer the 2D retro games though. 3D games are OK for some things but they aren't the be-all and end-all. If we had the computing power in 1985 that we have today you'd see some fabulous things.

I'm still programming, mostly in C, on a PC. I get to check other programmer's code and deploy our software systems. I've been to Ireland, Kuwait and Singapore on business so I get to travel a bit.

I might get round to writing something games-like one day. I've installed a development system on my PC at home. It's a long old day though in business so I don't know if I'll get round to writing anything meaningful. Never say never.