Fire and Rain

15 May


While the Apple Macintosh, Atari ST, and Commodore Amiga weren’t exactly flying off American store shelves in 1985 and 1986, they at least had the virtue of existing. The British computer industry, by contrast, proved peculiarly unable to produce 16-bit follow-ups to their 8-bit models that had made Britain, measured on a per-capita basis, the most computer-mad nation in the world.

Of the big three in Britain — Sinclair, Acorn, and Amstrad — only Sinclair really even tried to embrace the 16-bit era on a timely basis, announcing the QL the same month of January 1984 that the Mac made its debut. They would have been better off to wait a while: the QL was unreliable, ill-thought-out, buggy, and, far from being the “Quantum Leap” of its name, was still mired in the old ways of thinking despite its shiny 68008 processor, a cost-reduced variant of the one used by the Apple Macintosh. It turned into a commercial fiasco, and Sinclair never got the chance to try again. Torpedoed partly by the QL’s failure but more so by a slowdown in Spectrum sales and Sir Clive’s decision to pull millions of pounds out of the company to fund his ridiculous miniature-television and electric-car projects, Sinclair came within a whisker of bankruptcy before selling themselves to Amstrad in 1986.

Acorn, meanwhile, gave their tendency to overengineer free rein, producing a baroque range of new models and add-ons for their 8-bit BBC Micro line while its hugely ambitious 32-bit successor, the Acorn Archimedes, languished in development hell. Undone by the same slowing market that devastated Sinclair as well as by an ill-advised grab for the low-end in the form of the Acorn Electron, Acorn was also forced to sell themselves, to the Italian company Olivetti.

That left only Amstrad still standing in an industry that had been just a year or two before the Great White Hope of a nation, symbol and proof of concept of Margaret Thatcher’s vision of a new, more entrepreneurial and innovative British economy. Unfortunately, Amstrad’s founder Alan Sugar just wasn’t interested in the kind of original research and development that would have been required to launch a brand new machine based on the 68000 or a similar advanced chip. His computers, like the stereos he had been selling for many years before entering the computer market, were all about packaging proven technology into inexpensive, practical products for the masses. There’s something to be said for that sort of innovation, but it wasn’t likely to yield a Macintosh, an Amiga, or even an Atari ST anytime soon.

This collective failure of the domestic makers meant that British punters eager to experience the wonders of 16 bits were forced to look overseas for their new toys. Yet that was a fairly fraught proposition in itself. The Macintosh was practically a machine of myth in Britain for years after its American debut, absurdly expensive and available only through a handful of specialized shops. Only wealthy gentlefolk of leisure like noted Mac fanatic Douglas Adams could contemplate actually owning one. And the Amiga, not even available in Britain until June of 1986, also suffered even thereafter from an expensive price tag and poor distribution.

That left the Atari ST as the only really practical choice. The situation was a surprising one in that Atari had not traditionally been a big player in Britain. The Atari VCS game console that had left its mark on the childhood of an entire generation in North America was virtually unknown in Britain, and, while Atari’s line of 8-bit computers had been nominally available, they had been an expensive, somewhat off-kilter choice in contrast to the Sinclair Spectrums and Commodore 64s that outsold them by an order of magnitude. But Jack Tramiel, previously the head of Commodore and now owner of the reborn post-Great Videogame Crash Atari, knew very well the potential of the European market, and pushed aggressively to establish a presence there. In fact, the very first STs to go on sale did so not in the United States but rather West Germany. By the end of 1985 STs were readily available in Britain as well and, at least in contrast to the Macintosh and Amiga, quite inexpensive. A British software industry looking for a transformative machine to lift home computing in Britain out of its doldrums placed its first hopes — admittedly largely by default — in the Atari ST.

Still, it was far from clear just what sort of form the hoped-for new ST software market would take. The ST may have been a bargain in contrast to the Macintosh and Amiga, but it was still a fairly expensive proposition within a country just getting back on its economic feet again after what felt like decades of recessions, shortages, and labor unrest. A reasonably full-featured ST system could easily reach £1000, many times what one could expect to shell out for the likes of a cheap and cheerful Speccy. The ST would seemingly need to attract a different sort of buyer, with more money to spend and perhaps a few more years under his belt. This expectation was one of the calculations that led to Rainbird, one of the most significant British software houses of the latter 1980s.

Rainbird was born from Firebird, a slightly older label that has plenty of significance in its own right. In 1984 British Telecom, solely responsible at the time for the telecommunications grid of all of Britain, was privatized, becoming a huge for-profit corporation as part of Margaret Thatcher’s general rolling-back of the socialist wave that had followed World War II. Even before the first shares were sold to the public on November 20, 1984 — the largest single share issue in the history of the world at the time — the newly liberated management of British Telecom began casting about for new business opportunities. It didn’t take them long to notice the exploding market for home-computer software. They thus formed a new division of Telecomsoft, whose first imprint was to be called “Firefly Software.” That name was quickly changed to “Firebird” — it seems marketing manager James Leavey had just been listening to Stravinsky’s The Firebird — when they discovered a potential trademark conflict with another company. Firebird made its public bow in time for Christmas 1984 with a whole raft of mostly simple action games, selling for £2.50 (Firebird “Silver”) or £6 (Firebird “Gold”). Many turned into bestsellers.

Whether you considered British Telecom’s entrance into software a necessary result of a rapidly maturing industry or you were like Mel Croucher of Automata in considering them nothing more than “parasites” on software’s creative classes, it marked a watershed moment, a definitive farewell to the days of hobbyists meeting and selling to one another at “microfairs” and a hello to a hyper-competitive, corporatized industry destined someday to be worth many billions. If anyone was still in doubt, in December of 1984 another watershed arrived when newly minted software agent Jacqui Lyons presided over an unabashed bidding war for the right to publish Ian Bell and David Braben’s Elite on platforms other than the BBC Micro. Firebird, with the deepest pockets in the industry by far, won the prize.

Although published on the Firebird label, Elite would prove to be something of a model for the eventual Rainbird. Unlike Firebird’s previous releases, which had used the colorful but minimalist packaging typical of British games at the time, Elite‘s big, sturdy box contained not just the cassette or disk but also a thick manual, an equally thick novella to set the stage, a glossy quick-reference card, and a poster-sized ship-recognition chart (all licensed and reproduced from the Acornsoft original). All this naturally came at a price: £15 for the cassette version, fully £18 for the disk version. It marked a new way to sell games in Britain: as luxury products aimed at a classier, more sophisticated, perhaps slightly older consumer. In spite of the extra cost of all that packaging, the profit margins on its higher price points were to die for. If the Elite approach could be turned into a sustainable line rather than a one-off, British Telecom just might have something.

Tony Rainbird

Tony Rainbird

One person inside British Telecom who paid a lot of attention to Elite‘s launch and its subsequent success was Tony Rainbird, a former software entrepreneur in his own right who now worked for Firebird. He began agitating his superiors for a new software label, a sort of boutique prestige brand that would sell more sophisticated experiences at a correspondingly higher price point; it would be, if you can forgive an anachronistic metaphor, the Lexus to Firebird’s Toyota. His thinking was influenced by a number of factors in addition to Elite‘s success. He was very aware of the Atari ST that was then just arriving in Britain, aware that the people who bought that machine and in the fullness of time its inevitable eventual competitors would be willing and able to spend a bit more for software. And he was very aware of the American software market, which at that time was enjoying a golden age of gorgeous game packaging thanks to labels like Infocom, Origin, and Telarium. Games from those publishers and many others in the United States were marked by high concepts, high prices, and, yes, high margins to match. Elite, the first British game that could really compete in the United States on those terms, had been the first game that Firebird exported there; it became as huge a hit in the United States as it had in its native country. A new luxury imprint could continue to export games and other software that suited the higher expectations of Americans, whilst trading on the slight hint of the exotic provided by their British origins.

After getting permission to give the new line a go, with he himself at its head, Tony Rainbird decided that all the games should be published in distinctive boxes done in a deep royal blue, a color which to him exuded class. His first choice for a name was “Bluebird Software.” But, once again, a search turned up a conflict with another trademark, so he allowed himself to be persuaded to give the line his own name. Just as well; it fitted even better as a companion to the Firebird line.

Rainbird was launched quietly at the end of 1985 with two 8-bit creativity titles, The Art Studio and The Music System, that echoed more than faintly Electronic Arts’s Deluxe line of high-toned creative applications in the United States. But it was the following year that saw things get started in earnest, with two splashy game launches for the Atari ST. One of these, The Pawn, is an adventure game we’ve met before, along with its maker Magnetic Scrolls; we’ll continue their story in the next article I write. It’s the other, a space shoot-em-up called Starglider, that I want to spend just a bit of time with today. It’s not really a great game, but it is an interesting one to consider in its historical context, not least because of the colorful history of the person who wrote it, a young hacker with the perfect videogame-character name of Jez San (“Jez” is a nickname for “Jeremy”).

Jez San had already had a greater impact on British computing before his twentieth birthday than most programmers manage in a lifetime. It all began when his father, owner of a successful import/export firm, gave him an American TRS-80 computer in 1978, when he was not quite thirteen years old. He first won attention for himself by coming up with a hack to let one attach the joystick from an Atari VCS — another piece of foreign exotica that came to him courtesy of his father’s business — to the TRS-80 for playing games in lieu of the awkward keyboard controls that were the norm. His skills had progressed so far by 1982 that his father agreed to become partners with him in a little software-development company to be called Argonaut Software — think “J. San and the Argonauts” — run out of his bedroom. Whilst writing software for whomever would pay him, San was also soon terrorizing the network of British Telecom. He became one of his country’s most skilled phone phreakers, a talent he used to become a fixture on computer networks all over the world. It was in fact as a network hacker rather than a programmer or game developer that he first did something to make all of Britain sit up and take notice.

On October 2, 1983, San hacked the email account of one of the presenters of a live edition of the BBC program Making the Most of the Micro, an incident that has gone down in British computing lore as the “very first live hack on TV.” Millions of Britons watched as the presenter’s computer displayed a “Hacker’s Song” from San in place of the normal login message. Like much involving San, it was both less and more than it seemed. What with War Games a huge hit in the cinema, the BBC wanted something just like what San delivered for their live show where, as the host repeatedly stated, “anything could happen.” San’s alleged victims were more like co-conspirators: “They knew I was going to hack, they were quite hoping I would,” he admits. Why else would they announce the password to all and sundry inside the studio over a live microphone just minutes before the program began? After that, it just took a phone call from a few of San’s friends who were hanging about the studio. Further circumstantial evidence of the BBC’s complicity in the whole incident is provided by the host and presenter’s weird lack of affect when the “Hacker’s Song” appeared on the screen — almost as if they expected it, or something like it, to be there. As for the “Hacker’s Song” itself, it was lifted not from some shady underground but from the very overground pages of the American magazine Newsweek, yet another gift of San’s importer/exporter father.

San was forced to cloak himself in anonymity for this great exploit, but he got the chance to advertise his skills to the world and earn himself some real money in the process soon thereafter, when he was hired by a dodgy little company called Unicom to help in the development of a new, ultra-cheap modem for the BBC Micro. He wrote the software to control the modem, much of which was supplied not on disk or tape but as a new ROM chip to be installed in the computer itself. The modem lacked approval from the British Approvals Board of Telecommunications, meaning that, in one of those circumlocutions only a hidebound bureaucracy could come up with, it was legal to buy and sell but not to actually use on the British telephone network; it was required to bear a bright red triangle on its face to indicate this. Undaunted, Unicom took the non-certification as a badge of street cred, painting little demons on either side of the BABT’s warning triangle that made it look like just part of the logo. The Unicom modem quickly became known as the “Demon Modem.” At a fraction of the price of its more legitimate competitors, it made outlaws of many thousands of Britons and earned San tens of thousands of pounds. Perhaps all those punters should have been more cautious about the people they did business with: in the book Grand Thieves and Tomb Raiders, San makes the eye-popping claim that he imbedded backdoors into the bundled software “to take control of a computer using his modem, to make it play sounds, or type words to the screen.” This sounds frankly dubious to me given everything I know about the technology involved, but I offer it nevertheless for your consideration. At any rate, San claims he mainly used his powers to do nothing more nefarious than cheat at MUD.

San first came face to face with the executives at British Telecom not, as you might expect, because he was hauled into court for his various illegal activities, but rather when he was hired by them to help David Braben and Ian Bell port Elite from the BBC Micro to the Commodore 64. Having accomplished that task in a bare couple of months, he parlayed the success into a contract for a 3D space-combat game of his own, to target the new generation of 68000-based home computers that were on the horizon. Eager to get started, and with the Atari ST and Amiga not yet released, he rented a Macintosh for a while to start developing 3D math routines for the 68000, then shifted development to the ST as soon as it arrived in British shops. When Rainbird came into being, this 16-bit prestige project was quickly moved from Firebird to the new imprint.


The finished Starglider that was released by Rainbird in October of 1986 was once again both less and more than it seemed. With Bell and Braben having already started squabbling and proving unable or unwilling to deliver a timely follow-up to Elite, Rainbird clearly wanted to position Starglider instead as that game’s logical successor. Just as Acornsoft had for Elite, Rainbird hired an outside author, James Follett, to write a novella setting the stage for the action. Its almost 70 pages tell the story of an alien invasion force that disguise themselves as Stargliders, a protected species of spacefaring birds, in order to penetrate the automated defenses of your planet of Novenia. You play Jaysan — didn’t I say he had the perfect name for a videogame character? — who with the assistance of his hot girlfriend Katra must save his world using the last manned fighter left in its arsenal. How’s that for a young nerd’s wish-fulfillment fantasy?

That said, connecting all of the texture provided by the novella to the actual game requires quite an effort of imagination. Starglider lacks the huge universe of Elite, and lacks with it Elite‘s strategic trading game and the slow-building sense of accomplishment that comes from improving your ship and your economic situation and climbing through the ranks. Most of all it lacks Elite‘s wondrous sense of limitless freedom. Rather than a grand space opera, Starglider is a frenetic shoot-em-up in which you down enemies for points — nothing more, nothing less. Get to and destroy a faux-Starglider, the “boss” of each level, and you advance to the next, where everything becomes a little bit harder. With no save facility beyond a high-scores table, it would fit perfectly into an arcade.


Which is not to say that Starglider wasn’t impressive in its day within its own more limited template. The game’s most innovative feature may just be its missile-eye view: when you fire a missile you can switch your view to a camera in its nose and guide it to its target yourself. There’s also a modicum of strategy required: you need to return to a depot periodically to repair your ship and restock your weapons, and you need to replenish your energy supplies by skimming over power lines located on the surface of Novenia (shades of Elite‘s fuel scoops). But mostly Starglider seems more concerned with showing off what its 3D engine can do than pushing boundaries of gameplay. Its wireframe 3D graphics aren’t exactly a revelation in comparison to Elite‘s, but there are far more enemies now with more complex shapes, which move more smoothly — the Stargliders themselves, enormous birds that smoothly flap their wings, are particularly well-done — and which are now in color.

The problem with a game that lives and dies on its technical innovations is that once those innovations are incorporated into and improved upon by other games it has very little to offer. Writing about Elite, I noted that Braben and Bell could easily have stopped after they had a workable 3D action game, the first of its kind on a PC, and been assured of having a sizable hit on their hands. What made Elite a game for the ages was their decision to keep going, to use that 3D engine as a mere building block for something grander. The lack of a similar grander vision is what makes Starglider, as reviewer Ashley Pomeroy put it, “a period piece.” Within a year or two other games would offer 3D engines that used solid polygons instead of wireframes — including, ironically, later versions of Elite itself. Many of Starglider‘s other aspects that were impressive back in the context of 1986 are most kindly described as quaint today, like the poorly digitized voice of Katra that occasionally screams out a monosyllabic exclamation. Most embarrassing of all is the brief digitized snippet of a studio-recorded theme song that plays as the game starts; it sounds like a particularly cheesy Saturday-morning toy advertisement.

The use of digitized sound from the real world is of course a signpost to the future of multimedia gaming, and represented a real coup in 1986, as San himself describes: “On the Atari ST Starglider was the first game to use sampled sound. I sat with my ST open, measuring voltages off the sound chip, and modulating the volume controls in real time on the three channels to find what voltages came out so that I could play samples.” Technically brilliant it may well be. Timeless, however, it’s not.

In its day, though, it was more than enough to make Starglider just the big hit needed to get the fledgling Rainbird imprint off the ground. Its sales soared well into the six figures once ported, with greater or lesser degrees of success, to quite a variety of 8-bit and 16-bit machines (a list that includes the Amiga and at long last the Macintosh, the platform where its development first began). The sheer number of ports illustrates what would soon become Rainbird’s standard business model: to release games first as prestige titles on the 16-bit machines, then port them down to the less capable but more numerous 8-bitters where the really big sales numbers could be found. Rainbird quickly learned that an Atari ST or Amiga game on the Commodore 64 still retained some of the cachet that clung to anything involving a 68000. (Cinemaware in the United States would quickly learn the same thing and engage in a similar triangulation.)

Jez San used the income Starglider generated to put his one-man-band days behind him, bringing in additional programmers to establish Argonaut as one of the mainstays of British game development for almost two decades to come. Argonaut became one of the leading lights of a certain school of game programming, centered in Europe, that would continue to program the new 16-bit machines largely as they had the older 8-bits: in raw assembler, banging right on the hardware and ignoring operating systems and all the rules of “proper” programming found in the manuals. The approach seemed to demand young minds. Indeed, it seemed to delight in chewing them up and spitting them out before their time. In 1987 a 21-year-old San was already starting to feel his powers fading in contrast to the young turks he was hiring to work for him; he declared he’d likely be “over the hill” in about two more years. He was therefore eager to complete the transition he’d already begun into a purely managerial role. Even professional sports didn’t worship youth like this brutal meritocracy.

San and his colleagues and the many other developers like them positively swaggered about their prowess at down-and-dirty to-the-metal assembly-language coding, treating those who chose to work differently with contempt. “I don’t believe you can write performance software in C,” said San bluntly in that same 1987 interview. What he apparently failed to understand or didn’t consider significant was that, in being forced to focus so much on the trees of registers, opcodes, and interrupts, he was forgoing a veritable forest of conceptual complexity and design innovation. Higher-level languages had, after all, been invented for a reason. It’s very difficult, even for an agile 20-year-old mind, to conceive really interesting systems and virtual worlds when one is also forced to manually keep track of the exact position of the electron gun painting the screen. Thus the games that Argonaut and houses like them produced were audiovisually spectacular in their day but can seem underwhelming in ours. The fundamental limitations of their designs are all too painfully apparent today, long after even the best of 1980s graphics and sound have lost their ability to awe. For that reason I don’t know that we’ll be hearing a lot from this school of game development in the years to come on this blog, but rest assured that they’ll be beavering away in the background, brilliant in their own ephemeral way.

(Sources: the film From Bedrooms to Billions; the book Grand Thieves and Tomb Raiders; Amazing Computing of November 1987; Retro Gamer 86 and 98; Amiga Computing of June 1988; Your Computer of January 1985, February 1986, June 1986, October 1987; Computer and Video Games of February 1985; Popular Computing Weekly of March 21 1985, November 7 1985, November 14 1985, and March 27 1986; Computer Gamer of August 1986; Home Computing Weekly of April 30 1985; Games Machine of October 1988. The web site The Bird Sanctuary is full of information on Firebird, Rainbird, and their games. If you’d like to experience Starglider for yourself, feel free to download a zip from here containing Atari ST and Amiga disk images along with all of the goodies that accompanied them.)


Tags: , , ,

28 Responses to Fire and Rain

  1. iPadCary

    May 15, 2015 at 12:21 pm

    I can’t even BEGIN to list the Golden Age classics that Rainbird’s responsible for. Helluva lable, them …. Great work as per usual, Jimmy!

  2. Felix

    May 15, 2015 at 3:51 pm

    That’s odd… I could have sworn Starglider was flat-shaded in its 16-bit incarnations. Guess I was thinking of the sequel. But then, what exactly did the game lose on 8-bit machines?

    And you make a good point about technological gimmicks. As I pointed out before, the technical wizardry in Lords of Midnight can be duplicated nowadays with 300 lines of fairly simple Javascript. It’s everything else that makes the classic game so many people still love.

    As for software performance snobbery, that’s always been a thing. At one point it was, “we’ll code in C with a little assembly in critical portions”. Then C itself took assembly’s place for those cases. Nowadays it’s Java with C++ as the low-level fallback. And at each step there are people priding themselves on their ability to code clever bit-fiddling routines, while failing to see big obvious optimizations.

    • Pedro Timóteo

      May 15, 2015 at 4:51 pm

      Yes, only the sequel had shaded graphics. As to what it lost in the 8-bit versions, that would be mostly speed, and (depending on the machine) some features. For instance, the Spectrum 48K version was simplified compared to the 128K one (no missile camera, no speech, no special missions).

  3. Felix

    May 15, 2015 at 4:26 pm

    Come to think of it, the first performance snob must have been Von Neumann himself, who supposedly criticized his colleagues for “wasting computer time” with an early assembler. And if I’m not mistaken, it was around the same time that people discovered just how hard it was to write a correct computer program…

    You see, as of late I’ve started thinking that computer programming is just about the most complicated task the human brain can handle — right at the upper limit of our ability. And yes, there’s only so much we can pay attention to. Studies have shown long ago that productivity and code correctness, as measured in lines per day and bugs per line, are constant regardless of programming language used — and therefore of abstraction level. You can work in a team; you can specialize. But sooner or later you’ll hit a hard limit to what you can do, and the limit gets lower as you grow older. As for teams, they can only grow so big before communication overhead suffocates them.

    Cue software companies still burning people at both ends, so that few people grow old writing software, and there’s never enough real expertise to go around. All while software projects continue to grow in size and ambition, even as the success rate dwindles.

    • Andrew Plotkin

      May 15, 2015 at 8:11 pm

      “You see, as of late I’ve started thinking that computer programming is just about the most complicated task the human brain can handle — right at the upper limit of our ability.”

      Not a coincidence!

      I don’t want to get into “lines per day” as a measure, but the complexity of a programming task can be as high or low as you want — depending on the task that you tackle. And in software, with no material limits, people tend to tackle the hardest program they’re capable of imagining. Midjudge the limit by a hair and you’re screwed.

      I like to say that the difference between a lousy software engineer and a great one is the difference between building the most complex program you can write, and the most complex program you can *maintain*.

    • Jimmy Maher

      May 16, 2015 at 9:18 am

      Obviously the benefit of higher-level languages, libraries, even operating systems is that they abstract away complexities that no human could possibly hold straight in her mind. I think that machines like the Amiga and Atari ST were right there at the bleeding edge of what a single person could manage, the last machines that one could know and control completely, at the raw hardware level. Doing so could be tremendously rewarding, but, as I noted, came with trade-offs.

      One problem that arises as the software stack on which we build our applications becomes ever higher is that we remain vulnerable to bugs at the lower layers. As I’m sure many other programmers can attest, this sort of problem is incredibly frustrating and seems if anything to be getting more rather than less common as time goes on and yet more layers accumulate below our high-level code.

      As a self-taught coder who mostly does less than mission-critical mobile apps these days — i.e., little more than a (hopefully) talented amateur in eyes of a real software engineer — I do kind of feel instinctively that much of the complexity of modern software development is unnecessary, that we’ve gotten off on the wrong track somewhere. But I also recognize that I am unschooled in even many fairly fundamental notions in computer science, and thus my feelings may very well be stemming from sheer ignorance and creeping grumpy-old-manism.

      • Felix

        May 16, 2015 at 1:43 pm

        I don’t think you need to be a software engineer or computer scientist to have an opinion about that. For what it’s worth, I’m self-taught too, and I worked on fairly serious business software. And after a decade and a half of butting my head against the Web’s infrastructure, I can safely say Rasmus Lerdorf didn’t know what he was doing when he created PHP any more than Linus Torvalds knew what he was doing when he created Linux. Both are hobbyist projects run out of control, and it shows.

        Look at it this way: the Minix 3 kernel has 6 thousands lines of code. The Linux kernel may well have 6 *million* by now. (Sans drivers!) Are you seriously telling me the latter is a thousand times smarter? More capable? What sort of advanced features can take *a thousand times more lines of code*? It’s still just an OS kernel…

        Nope, I don’t think you’re either ignorant or grumpy. I think you’re simply noticing that the emperor is naked. And you’re not the only one, not by far.

        • Carl

          May 19, 2015 at 11:13 pm

          To be fair, these things often follow the Pareto principle where 80% of the payoff comes from 20% of the work. As Minix (a teaching system) shows, you can get a long way with a minimalist design by ignoring edge cases. And that is where the complexity comes in, the edge cases. It’s like that in anything. Once the low hanging fruit is picked it takes bigger and bigger increases in complexity to get incremental improvements in performance. Then it is time to shift the paradigm.

          • Iffy Bonzoolie

            May 1, 2020 at 7:58 am

            In my modest experience, code complexity comes from a couple root sources:

            1. Resource Constraints. The classic “Iron Triangle” – Scope, Time, and People. Most of the software industry tries to squeeze down Time and People as much as possible, using a variety of terrible management ideas. So, coders move fast, take shortcuts, and leave behind debt. It takes more time to write less code. A lot of symptoms are rooted in this issue.

            2. Complex Requirements. Perhaps not surprisingly, complex specifications lead to complex code. When problems can’t be generalized, each bespoke requirement needs a bespoke solution. This actually grows polynomially. M components interacting with each other have M² connections.

            Sometimes, it turns out that a feature is actually not worth being built, because it messes the code up too much!

            OK, I know that sounds like a typical Prima Donna Engineer thing to say, but… keeping the code simple, general, correct, and performant also makes implementing new features easier and faster. I’ve worked in many code bases that had grown organically for years, and it can be a house of cards. It’s hard to build an extension on a house of cards.

            It’s a balancing act, and the person that understands the value of the requirement and the person that understands the cost are two different people who don’t really trust each other.

  4. Carl

    May 15, 2015 at 8:48 pm

    While it is true Acorn had difficulty coming up with a British 16-bit computer they were certainly trying, with custom silicon, no less!

    The outcome of this development hell was the ARM microprocessor architecture, currently the dominant embedded CPU architecture in the world, driving, for example, the iPhone and iPad and countless other devices.

    Interesting stuff!

    • David Boddie

      May 15, 2015 at 10:20 pm

      I think you could say that the development hell at Acorn was more to do with the operating system for the ARM machines than with the hardware. There’s one account of this by the lead developer of the system (Arthur) that was used, though sadly there’s no corresponding account for the competing system (ARX).

  5. Bumvelcrow

    May 16, 2015 at 10:11 am

    I remember wandering into my local computer shop around the time Starglider was released. The Atari ST was in pride of place near the entrance and Starglider was running in attract mode with the digitised themesong/advert playing. Cheesy it may have been but everyone’s collective jaws were on the floor, mine included.
    BTW, I loved the novella but I’m sure Katra was the strong and determined leader rather than ‘hot girlfriend’ and Jaysan was a self-centred layabout. Their relationship through it and the sequel was grimly platonic.

    • Jimmy Maher

      May 16, 2015 at 10:36 am

      Milages may vary, but personally I’d say Katra was more the enthusiastic but subordinate hot nerd girl — a fairly established fantasy figure in this type of literature — rather than “a strong and determined leader.” She’s forever being described as “the girl” (Jaysan, presumably about the same age, is notably never described as “the boy”), “gasping” and “clinging to” Jaysan and trying to rally him to her cause, but she remains dependent on her roguishly charming man to ultimately save the day. This is an even older cliche; see movies like The African Queen and for that matter the Han Solo character in Star Wars.

      While their relationship is never explicitly defined (in either sense of the word), Jez San himself did consider her to be “his” girlfriend, as he states in his Retro Gamer interview and I believe other places.

      • Peter Piers

        January 18, 2016 at 6:00 pm

        I just watched the African Queen the other day, and am surprised at your description of the female lead! Hepburn’s character is certainly dependant on Bogart’s character’s know-how and strength, but she never gives up, she has a strong mind of her own. She’s almost a benign Lady Macbeth. In the end she also assists with the repairs as best she can.

  6. Cmdrfan

    May 17, 2015 at 6:19 pm

    Well, although I have to admit that your conclusion about the technical wizardry of these british programmers and how it quickly became obsolete, you should also shouldn’t be quick to dismiss their influence: as a perido Amiga owner and previous C64 owner I was very fond of the visuals and style of british software. You just couldn’t pile them with the other “arcade” (derogatory label) stuff coming out at the time, particularly on consoles. These guys managed to inject a certain artistic sense that’s quite hard to put into words and that made for a lasting effect on user’s memories.

    I sincerely hope you write something about this. Maybe a piece on Psygnosis ?

    Maybe it’s the same reason why we still like to look at Reanassaince-era portraits, even though their technical “wizardry” was made obsolete by photography long ago and in spite of the emphasis on “design” of modern art.

    • Jimmy Maher

      May 18, 2015 at 5:37 am

      I’m sure Psygnosis will be getting some attention eventually. Shadow of the Beast is practically the archetypical example of this hyper-technical school of game programming (dodgy gameplay included), and, as another commentator said a while back, rivals Defender of the Crown for *the* game people think of when they think Amiga game. And then there’s Lemmings, pioneer of a whole category of casual puzzler…

  7. The Larch

    March 4, 2017 at 2:30 am


    Firstly, thanks for the brilliant blog. Fascinating to read about the early home computers and adventure games in the US.

    I can understand why you’re cynical about the “backdoors” in the Demon COMMZROM, but I can confirm there was at least one. I hope he’ll forgive me for divulging some details.

    On one occasion I bumped into Jez the Wizard in MUD and he asked if I was using a Demon modem. When I answered yes, he replied with something like “music to my ears!” and my Beeb played a little fanfare.

    Knowing the addresses used for controlling the sound hardware, it wasn’t hard to locate the code responsible in the terminal emulation. It was triggered by a pretty unlikely sequence of characters (“\~^|” if I remember correctly), and interpreted the pairs of characters that followed as encoding pitch and duration to play as a series of tones. I guess Jez had a keyboard macro programmed with the appropriate sequence, just in case it came in handy.

    Of course, the next time I encountered Jez in MUD, (assuming I got everything right) his speaker played the death march!

  8. David C

    September 24, 2018 at 9:13 pm

    Had an ST and Starglider as a kid – the second computer in the family, the first being a Speccy 128 +2. As a kid would look at the credits on the opening screen and idly wish I could find out more on Argonaut Software, Jez San and Rainbird. I have indeed been waiting thirty years to find out more!
    Thank you so much!
    I have to mildly object to your assertion of the Starglider theme as being “cheesy”, for 1987 it’s actually pretty contemporary and perhaps most importantly, almost advertising level catchy. It sticks in your head, and brands the game quite memorably. My two cents.

  9. Leo Vellés

    October 15, 2019 at 4:21 pm

    Jez San had already had a greater impact on British computing before his twentieth birthday than most programmers manage in a lifetime. It all began when his father, owner of a successful import/export firm, gave have him an American TRS-80 computer in 1978.
    I think therr Is one verb here that should nota be: “gave have him an American…”

    • Jimmy Maher

      October 17, 2019 at 9:41 am


  10. Jonathan O

    January 6, 2021 at 9:37 am

    I know this is a minor part of the blog, but your comments on the state of Britain before the 1980s strike me as rather tendentious – it gives a picture of the country that doesn’t match with what I remember. Even the 70s, problematic as they could be, weren’t that bad most of the time, And I doubt if the post-war decades saw much real socialism, unless you regard anything designed to ameliorate poverty as socialism. (I’d recommend the book “Whatever Happened To The Tories” by Sir Ian Gilmour for a more balanced picture of the economic and political landscape in the post-war years.)

    • Jimmy Maher

      January 6, 2021 at 2:22 pm

      “That bad” is of course relative; certainly England was doing far better by any objective metric than the majority of the world. Among developed Western nations, though, I think there’s a pretty firm historical consensus that its economy didn’t perform terribly well in the 1970s and early 1980s, and the labor unrest of the time is quite well-documented. (See the 2014 film Pride for a delightful portrait of the London Gay Pride scene uniting with striking coal miners in 1984…)

      Of course, there’s always a regional component to these things; as usual, the north of England suffered more of everything than the south. The socioeconomic tenor of the times comes up again and again in interviews with game developers from Liverpool and Manchester; less so those from London. Much the same can be said about the respective music scenes…

  11. Mike Richards

    April 4, 2021 at 4:58 pm

    Another great page – thank-you for the blast of nostalgia.

    One small correction, the QL didn’t have a 68000 – it * should* have had one, but penny-pinching Clive Sinclair (who would have given Tramiel a run for his money when it came to corner-cutting), wouldn’t pay for one. So after looking at the Z8000 (the 16-bit successor to the Z80) and the Intel 8086, Sinclair went for the Motorola 68008 at 7.5MHz.

    The 68008 has an external 8-bit data bus with 20-bit addressing. The 68000, has a 16-bit data bus and 24-bit addressing. At the time of its development, the 68000 was about twice the price as the 68008. The strange bit is that Sinclair ended up paying more for 68008 chips than its rivals paid for 68000s even though Sinclair was ordering similar volumes.

    Though the processor was probably the least of the QL’s problems – the endlessly-delayed launch, wonky ROM dongle hanging out the back and those bloody Microdrives killed it. Had it shipped with a floppy it might have stood a chance, at least until the two big American 68000 machines rolled up.

    I remember the adverts and certainly in print the QL looked fantastic – especially when compared with the prices IBM and Apple were charging for their 16-bit machines. I was seriously impressed by the QL’s ‘hires’ graphics which looked so much better than we ‘d seen on 8-bit machines in the UK. And the QL looked sleek – Rick Dickinson’s design was seriously futuristic (well at least until the Spectrum +2 and +3 appeared). In the cheap plastic and that terrible keyboard – not so much. It didn’t seem long before it was being sold at £199, half the launch price. And then it was gone.

    • Jimmy Maher

      April 8, 2021 at 8:14 am


  12. Ben

    September 26, 2022 at 8:53 pm

    Jacquie Lyons -> Jacqui Lyons

    • Jimmy Maher

      September 27, 2022 at 12:10 pm


  13. Magnus

    July 14, 2023 at 8:45 pm

    Kind of mind blowing coincidence, that the guy was actually named Rainbird and worked at Firebird. Remember staring at those two logos on C64 game packages in awe as a kid, thinking they were so cool and wondering what they really meant.

    Thanks for a great blog!


Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.