I started out as ST Software when Hewson agreed to publish 3D Spacewars written for the Spectrum. This was a Space shoot'em up inspired by Space Invaders. I had the idea of getting 3D into a game, the only game I had seen using this idea were 3D maze type games on the ZX81. I wrote the game code in about three months in the evenings. My first few games were written without an assembler. I used to write in hex code but with labels using my own hex loader. This decoded the labels into addresses. My code looked something like:

rem FE00
rem 28@DEAD

I wrote the code first in assembler on squared paper then coded it by hand to the hex equivalent using a few look up tables. I typed the code in rem statements and my "Hex Autoloader" loaded the code into ram and replaced all the labels with relative or absolute addresses. Grahics were drawn on squared paper and coded into hex for loading in a similar manner. The whole process was error prone and long winded. When I look back its amazing how I ever finished a single game.

I left a steady job as a commercial programmer so I already had a great deal of experience in languages such as COBOL and IBM Assembler. I was determined to keep to a regular pattern of work so used to start religiously at 9.00 and end at 5.00. I worked in the dining room using a Spectrum and a black and white television. Working alone was a lonely experience so I asked Andrew Braybrook, a friend, to join me. This was quite a risk, because I had saved about £10,000 before I left work so I knew I would be able to survive. Taking on an employee meant that we had to get royalties from our games within the first year.

Andrew had a similar history to me, he was a commercial programmer at Marconi's and had programmed several games on an IBM mainframe. He began by converting my first games to the Dragon 64. This turned out to be a disaster as despite being an excellent machine, the Dragon stopped being produced. Sales of Dragon games almost immediately ceased, only a few hundred of each game sold. This taught us an early lesson, it was essential to back the right machine. This was no easy thing to do as new games machines were appearing all the time and it was hard to predict which ones would survive. Andrew moved on the C64 which seemed to be gathering momentum as a games machine despite Commodores best efforts to disguise it as a business machine.

Andrew had the benefit of an assembler program on both Dragon and C64. He wrote us a graphics editor on the Dragon that allowed us to convert graphics into hex code automatically. We still had to key the hex in by hand but it was a huge leap forward. I still could not find a Spectrum assembler that allowed large games to be input so improved my development system by adding a disassembler. Then at least I could get a listing of my program for debug purposes. We each had a decent printer and would print off our entire programs. We would then amend these listings by hand until they were really messy and then print off another. Debugging was done with a clever routine Andrew wrote called Abmon. It allowed us to select any memory address and examine its contents or change them while the game was running. This allowed us to tune up gameplay while we played. Assembling a program could take ages so the ability to change things as the game was running speeded up the process. It also helped in debugging.

On the Spectrum, programs were stored on cassette tape. I used a cycle of five tapes. This was necessary as often tapes would not read in properly, and ensured at least five previous versions were kept. I bought some microdrives when they came out but they were less reliable than tapes. One of the problems with the Spectrum was that it was more like a toy. Peripherals were supplied by many manufacturers and often conflicted. Many had to be slotted onto the Spectrum main board at the back. With a few of these devices piggybacked onto each other the machine was unstable. Hit the keyboard too hard and it would jog the machine and it would crash losing everything inside. I eventually bought a Beta drive. The software for the disassembler, printer and disk drives were all incompatible. I had to reverse engineer the object code and connect the programs together. This system engineering could take as long as writing a game but once done speeded things up.

In the early days we used to visit Hewson a very few months to show Andrew Hewson our latest offerings. He was a small independent publisher that had an edge over other publishers. He had his own tape duplication facility that allowed him to respond rapidly to demand and keep stock low. This allowed him to produce low volume original product. He became well known for the game content of his licensed games. He supported many independent developers like us. We would develop the game, he would market it. This division of responsibility allowed for the optimum in gameplay and worked well until the exponential cost of developing games required a large up front investment by publishers.

Without an up front investment there was no guarantee that a developer would get their game published successfully. With Hewson we never used to ask for advances. This meant we were under no obligation to produce anything and gave us complete creative freedom. As long as Hewson published our games to our satisfaction that worked out fine.

Graftgold Ltd.
I formed Graftgold Ltd when royalties started to rise. The idea was to keep the royalties in the limited company to finance future games and pay us a steady wage. Graftgold was immediately successful with its first products of Avalon and Paradroid. At last Andrew Braybrook's games were selling in quantity. Conversion of the games to other platforms was not done by Graftgold as it was decided that it was best to keep generating new products. Hewson was to find other developers to provide conversions. This was slow to get underway and it was not until Dragontorc and Uridium that successes were exploited on other platforms. Looking back this was a mistake. The games market was showing a distinct change towards being a commercial industry. Heavily marketed titles that had a licensed name were regularly charting higher than original product despite very often being poorer games.

The Commodore 64 and Spectrum dominated the sales charts but new 16-bit Amiga and Atari machines were launched. Sales were poor but magazines gave more and more space to the new 16-bit graphics. Developers such as Graftgold were keen to move onto the new machines but smaller publishers such as Hewson were reluctant to move on until the market was established. Graftgold were asked to produce budget games by Hewson. These were a new thing vastly undercutting the price of standard games. To make them pay, we would have had to sell them by the 100,000 or be able to produce a title every four weeks. The formula just didn't work.

On the Spectrum I was still enjoying success but every title seemed to sell less than the previous one, an ominous sign as they were taking longer to produce. The public expected the machine to be taken to the limits and beyond. Graphics routines were constantly rewritten to squeeze extra performance from the machine. After the success of Dragontorc I decided to leave the final part of the trilogy to later as sales were much less than Avalon. I opted for a space adventure. I conside red Astroclone to be one of my finest designs, integrating adventure strategy and shoot'em up elements.

The product was misunderstood, it did not fit into a neat category. It had some adventure sequences I was really proud of. I came back with Quazatron with borrowed gameplay from Paradroid and married this with a graphic engine I had no game for. Ranarama followed which again used ideas from Paradroid but in a magical context. The game did well in Spain but was kept back awaiting an Amstrad conversion that was late. I ended up doing the conversion myself. Unfortunalely by this time, there were about four top down dungeon type games on the market. Despite being regarded as the best in a comparative review of all these games, the game was not successful in the UK. Hewson however agreed to convert the game to the Atari ST as an experiment. The conversion was a copy of the Amstrad version and did not exploit the features of the Atari very well. John Cumming, a Hewson in house programmer, gave the conversion a face lift by adding som 16-bit graphics. The four colour ones were just not acceptable. The conversion taught us that you cannot just port a game to another machine. You have to consider the new platform and exploit its features. When Graftgold undertook its own conversions, we did not make the same mistake.

About this time I bought a couple of 8086 machines and across assemblers to Z80 and 6502. I built an interface to the Spectrum using my Kempston printer interface, rewiring it as an input and output interface. The C64 was easy to interface as it had a parallel port. We used the PC parallel printer output to "print" the code to the Spectrum and C64. This was an enormous step forward and was our first really professional quality set up.

I was working on Magnetron, a follow up to my succesful Quazatron, and Andrew was putting the finishing touches when we heard Hewson was in trouble. Two of his programmers, John Cumming and Dominic Robinson had decided to leave thinking the company was finished. They rang me up and asked if I was interested in employing them. I agreed and Graftgold thus doubled its size and moved to its first offices above a greengrocers in Witham. Debbie Silletoe, Hewson's second in command, left Hewson to join British Telecomsoft. She warned us that Hewson were falling apart fast and introduced us to British Telecomsoft.

We decided to let Telecomsoft publish our games. Hewson managed to survive and challenged us for the right to publish the two existing games. As Hewson had not paid us any advances or given us a contract, we felt we had the right to change publishers. A court case followed which for a time prevented publication, but generally did noone any good. Eventually, Telecomsoft and Hewson agreed to settle out of court, and Telecomsoft were allowed to publish the games. Later they contracted Graftgold to convert the arcade game Flying Shark to the Spectrum and Amstrad. Dominic Robinson did a superb job in six weeks with John Cumming doing the graphics.

I then had 14 days to convert the Spectrum version to the Amstrad. This was an interesting experience and our first foray into commercial conversion. Telecomsoft wanted to finance Graftgold's growth and specifically wanted Graftgold to start work on the 16-bit machines. However after a few months they expressed a need for Spectrum and C64 games. Programmers Gary Foreman, David O'Connor were taken on. Jason Page was taken on as a trainee and soon produced all Graftgold's sound and music. Dominic started work on the Atari Jaguar 64 and the Amiga producing some stunning demos. These eventually evolved into Graftgold's 16-bit operating system kernel and the 3D game Simulcra.

By the time the first round of 8-bit games Intensity and Soldier of Fortune were finished, Telecomsoft did not seem to have any interest in 8-bit original product. Dismal sales followed and Telecomsoft was bought by Microprose. This was a good move as far as Graftgold was concerned as the people who originally been behind Telecomsoft sponsored Graftgold's entry into 16-bit product, Paul Hibbard and Pete Morley, had moved to Microprose. Thus Simulcra was finished, but by this time Dominic had become disillusioned and burnt out being Graftgold's only 3D programmer. I finished Simulcra off redesigning the gameplay to fully utilise Dominics excellent 3D engine. The game got excellent reviews and could have been a massive hit if properly funded in the early days. At this stage Graftgold were increasingly reliant on publisher advances and had to program the games the publisher thought they wanted.

Graftgold's best success with Telecomsoft was the conversion Rainbow Island's. Using Dominic's 16-bit engine Andrew masterminded one of the best conversions ever made. David programmed the Spectrum version and Gary the C64 version establishing Graftgold as a master conversion house.

The trouble was that Microprose bought the game before it was published and fell into dispute with Taito the owner of the game. Thus it was Ocean that finally published the game. John Cumming did a tremendous job converting all the in game graphics as did Jason on the sound. Although everyone could program it was clear that specialists were needed to achieve the best products. Developing a game had become a team effort. Graftgold took on Michael, another artist to help John who excelled in pixel art. He had tremendous patience and would painstakingly build up detailed bit maps pixel by pixel. He worked with Andrew on Paradroid 90 a long awaited 16-bit version of the old favourite. Graftgold had gradually reestablished relations with Hewson and had provided music and graphics for various products. Paradroid 90 was also the companies first venture into the ever growing console market. Gary Foreman developed a PC Engine version. Unfortunately it was not to be. Hewson foundered again and this time went into liquidation selling Paradroid 90 publication rights to Activision. Graftgold were already developing "Realms" a 3D strategy game for activision. It was a bleak time when Graftgold heard that Activision was closing down its operation. Graftgold did not receive any royalties for Paradroid 90, had its first console game canned and a major product with a defunct publisher.

Those were desperate times but such was the reputation of Graftgold that help was already at hand.

Graftgold had developed both 8-bit and 16-bit conversions of Off Road Racer for Virgin. The conversion business was providing bread and butter funding for Graftgold. Original product was becoming increasingly expensive and risky to develop. Publishers were not keen on investing in original product unless it was finished. It took huge subsidies of our own money to pay for the experimental work that original product needs. The problem was for a run of products not one of these had produced any royalties. Virgin changed this with Realms. We bought back the product from Activision and licensed it to Virgin. They trusted us and gave us a realistic managed advance. The game worked well but was only moderately successful. Graftgold came back from the brink of extinction but at the price of being wholly reliant on publisher funding.

The console era took the market by storm. Graftgold completely missed the Nintendo 8-bit boom period due to its other commitments and a licensing policy that meant that few could develop for the machine. Virgin had to get another company to do the 8-bit conversion of Off Road. They made a fortune using Graftgold code developed for the Spectrum. Graftold with Virgin's sponsorship became a Sega developer and converted a series of games to the Master System and Gamegear including Offroad Racer and Superman.

This was a second boom time for Graftgold and profits were ploughed back into state of the art 486 pc's and new development systems. The conversions allowed Andrew to undertake an original product, in part financed for Mirrorsoft. Graftgold had learnt the hard way to survive it was necessary to split funding across several publishers. When Mirrorsoft went down Graftgold had two products with them, Fire and Ice and a console conversion of Total Recall that was canned. This double blow could have hit Graftgold hard but Renegade stepped in to publish Fire And Ice. This began a long relationship with Renegade and the Bitmap Brothers who in part owned and ran Renegade. Renegade was very much a developers publisher and was set up to give the developer a fairer share of revenue. Gary Foreman converted the Bitmap's Gods to the Megadrive before leaving to join Sega.

Graftgold was signed by Sega to develop Ottifant, its own product. Ottifant was a cartoon program featuring an elephant family that aimed at a similar market, but never achieved the success of the Simpsons. Most TV networks did not buy the tv series so the game only had a minimal publication rather than being a major licensed product. John Kershaw organised the background graphics. Colin Seaman joined the company as an animator and did most of the animations. Eldon Lewis programmed the Megadrive version and Kevin Holloway programmed the Master system version. Many a late night was spent trying to keep to an impossible schedule. We were initially given three months to complete the game to a standard rivalling Sonic the Hedgehog. Half of this was wasted waiting for a signed contract. Experience had shown it was folly to start without the contract being completed. New starter Emma Cubberly was the companies only female employee, apart from my wife who eventually did the payroll. Emma provided some of the background graphics and was one of several trainees taken on by Graftgold. We often trained staff from scratch unable to afford experienced staff.

Fire And Ice was a big seller on Amiga and plans were made to convert it to PC, Megadrive and Gamegear/Master System. Unfortunately Virgin wanted to publish the Sega versions and beat Renegade in approaching Sega. Eventually Sega Virgin and Renegade agreed to publish the Master system/Gameboy version but the Megadrive version despite being completed was never published.

All of a sudden 8-bit console versions were not required anymore and conversion work dried up.

Graftgold made a late entry into the Nintendo 16-bit market programming a football game for Empire intended for the 1994 World cup. We also converted the game to Amiga and PC and Acorn. The Nintendo version was never published despite being finished in time thus failing to recoup advances.

Graftgold started a series of original games for Renegade. Andrew developed a CD32 version of Fire and Ice that was not published. He went on to develop Uridium 2 an Amiga version of his hit C64 game. Iain Wallington programmed Virus Alert borrowing KLP2 as the hero from Quazatron. It was the end for Commodore and Amiga sales plummeted despite constant interest from the public. A Motocross game was canned on the Amiga and a PC version started. During this period game content and development costs went through the roof with the advent of the CD. Instead of 2 megabytes around 20 megabytes at least were expected increasing costs by 10 times. The quality of graphics steadily improved as advanced graphics workstations became the norm. The Playstation was in its early days. Graftgold was signed as a developer but the costs of Playstation development kit were prohibitive.

As time went by all resources were concentrated on Motox, a dangerous but necessary tactic. The company just could not afford more than 1 product without major financing. Renegade increasingly extended advances for Motox to enable the program to converted to a CD quality product. A playstation product was signed funded by Coconuts of Japan. This allowed Graftgold to buy Playstation developer kit. Many staff moved on to other companies in this period. Experienced programmers and artists were being offered big money to join companies such as Psygnosis. Many small developers were disappearing or being bought up. Most were dependant on publishing funding at this time. If a publisher chose to can a product it could mean the death of a company. Renegade was a good partner during these years of increasing commercial pressure. However in the end Renegade decided it could not compete with the giant publishers dominating the market with multi million pound budgets. Renegade was bought by Warner, a move welcomed at the time by Graftgold looking for investment to finance a new round of state of the art games.

At this time Graftgold were wholly reliant on monthly advances received from Renegade. Profits had been hit by the failure of Empire to publish their Nintendo soccer game, and the non publication of Sega Fire and Ice and CD32 Fire and Ice and Uridium 2. Motox was being funded by Renegade but was costing far more to develop and needed to sell well. Ominously Bitmaps had chosen not to let Warner publish their own games. Graftgold showed Warner a new major game prototype based on the Avalon gameplay. The demo stopped work in the Warner Office as all the staff came to see the demo. The game was immediately signed up and work started. The project would have been Graftgolds largest and staff were recruited and the first few months milestones worked and paid. Then Warner changed their mind about their whole position in the industry. Our project was one of the many that were canned as budgets were pruned. The decision almost killed Graftgold, indeed it seriously wounded it. Later a friend at Warner told me they did not expect Graftgold to survive. We managed to sign a conversion of Rainbow Islands to the Playstation and Saturn. A second conversion fell through after a few months work due to problems of the publisher obtaining the licence. Graftgold was surviving but only just. Monthly income paid the bills just but the company had no finance for working on new product and still had Motox to complete. Motox had been "finished" for about 6 months but Warner kept asking for upgrades and tweaks. Such was the market, only a triple A product was required. Finally another advance was negotiated for extra work on Motox. We signed a game with Coconuts called MBT, a 3D tank combat game. By this time Warner were being sold and it seemed that Motox may not be published. Motox was finally accepted by Warner and published. I do not think it had any marketing spend at all, an astounding decision in a market dominated by big spending publishers. We could not see any prospect of Motox getting any royalties above th e advances already spent. It eventually sold very well recuperating all advances and became Graftgolds biggest profit making product. I can only wonder what the result would have been with a marketing spend in scale with the development spend. Delays in the production of MBT prompted Coconuts to attempt to reduce our monthly advances to one third. As the company was absolutely dependant on this income there was no way this could be accepted so I decided I would have to lay all but a skeleton staff off. I was sitting ready to tell everyone when the phone rang, it was the MD of Perfect.

Perfect had enjoyed considerable success with their Discworld products and had the finance to assist Graftgold. I sold Perfect the majority of the company, and they started negotiating with Coconuts. We had to lay off our musician and all but one of our graphics artists as Perfect were to provide these functions. Over the next two years the MBT project was remodelled by Perfect designers to become Hardcorps. The idea was to add loads of high quality movie graphics to make the game blockbuster. Perfect installed a network at Graftgold but the company never regained its momentum. Perfect funded the day to day activities to keep Graftgold going. The idea was for Perfect to get extra funding from a publisher. This was eventually done but delays in the game development caused funding to cease before the game could be completed.

So 15 years of developing games came to an end. In all I believe Graftgold achieved my aims of producing high quality entertaining software. We made a lot of money on the way despite the ending and most of the time had a great deal of fun. We became known and respected in the industry. One day who knows I may rejoin the industry. For now only having to work 37.5 rather than 80 hours a week and having a cheque at the end of the month rather than worrying how to pay the staff, is a nice change. I would recommend the games industry to anyone wanting an exciting career buts its certainly not an easy ride. Most publishers we worked with either went bust, sold out or simply did not publish the game to our expections despite tight contracts. The trouble is the developer does their bit first then the publisher can choose the level to do their bit. Unless you can get real commitment by way of big advances you cannot rely on a publisher. The problem is finding a publisher who is willing to commit. There are more games than publishing opportunities so publishers can cherry pick. I have heard of publishers over subscribing product by a factor of many times. For every game they publish many more are deliberately canned sometimes killing the developer. Which game will they can? Unfortunately they tend to can the game they have invested the least in, not the one that has the least prospect of making money.

Publishers do take the lion share of the risk when they put up the finance, but they take the lion share of the profit. What I found was a reluctance to reinvest this in the fabric of development. The developer is expected to do this out of their share of the profits. The economics of the industry means that the publishers income is 4 or 5 times as big as the developers. Out of the developers 10-20% not only the successful games have to be developed but the investment in new ideas made. Publishing is expensive but not five times as expensive as games development. One of our publishers paid more for each of their desks than they would loan us for a network server.

Still all in all I have fond memories of all the people, staff and publishers alike I have worked with over the years. I would like to say thank you to all those people who believed in me and made the products what they were. I still dream of games as yet not written. Maybe one day I will again realise one of those dreams.

An old employee?
If anyone reading this ever worked for Graftgold, or indeed produced any games for them, feel free to let us know your experiences of working there.

» Go back to the articles

» F.A.Q. - look here before you send off an email.

» Credits - the list of people who made all this possible.

» Scene interviews - C64 sceners answer 20 questions about their time in the scene.