Revolution and Apocalypse

We computer technologists have made the term "revolution" a cliche over the last 15 years. We have attached scores of qualifiers to that word, yielding such phrases as "desktop publishing revolution", "personal computer revolution", home computer revolution" and so on. Yet I wish to drag that overworked word out one more time to describe a process that I believe will take place over the next ten years that I consider to be nothing less than well, revolutionary.

I believe that entertainment software will undergo a profound revolution over the next decade. When the year 2000 dawns, entertainment software will be a far cry from the games we now know. In this editorial, I propose to explain the causes and directions of this great revolution.

First Cause: Better Hardware
This is a no-brainer. Better hardware makes possible better software. The scale of hardware improvement is what is so impressive. The hardware of 1994 is two to three orders of magnitude more powerful than the equivalently priced hardware of ten years earlier, and another order of magnitude improvement is certain to come over the remainder of the 90s.

What is striking to me, as I consider this, is the magnitude of improvement in such a short span of time. Previous technological revolutions had profound effects with much smaller technological gains. For example, the Industrial Revolution was triggered by the development of the steam engine, yet the power made available by the steam engine was less than one order of magnitude greater than what was already available in the form of hydropower. Similarly, the internal combustion engine improved the power to weight ratio of the steam engine by only about half an order of magnitude, and yet that made possible automobiles and airplanes. Similar results arise from the development of the Bessemer process making available large quantities of structural steel, the application of mass-produced fertilizers in agriculture, plastics, and a host of other technological developments. The conclusion is that small improvements in performance characteristics can trigger gigantic changes in society. And here we have, with silicon technology, a performance improvement that has no parallel in human history. No other invention in all of human history has enjoyed such tremendous performance improvements in so short a time.

My astonishment leaps even higher when I consider the not just raw performance, but performance as a function of capital investment. We might compare the mightiest engines of today, the big power plants that generate a gigawatt of power, with the windmills of early Mesopotamia, noting that modern power plants generate about seven orders of magnitude more power than the windmills. Yet the modern power plants represent a much larger capital investment. If we compare performance of power plants per invested work-hour, the superiority of modern power plants falls dramatically. By contrast, computer hardware has made its gains at constant capital investment. The $2000 computer of 1994 is three orders of magnitude more powerful than the $2000 computer of 1980.

So here we have the mightiest technological leap in human history. If that doesn’t lay the foundations for some pretty big revolutions, I don’t know what will.

Second Cause: Mass Market
Consequent to the explosion in performance has been the explosion in sales of the machines. During the 1980s the primary market was in personal computers for business use, but during the 1990s the hottest growth is in the home computer market. There are already 25 million home computers installed in this country, and all projections show continued strong growth in the installed base.

The significance of this lies not in the raw numbers, but in the accompanying shift in demographics. During the 1980s, personal computers were purchased by early adopters (given the unfriendliness of the technology, "early adaptors" is a more apt term). These people shared a set of values that dictated the nature of entertainment software. Three traits in particular shaped first-generation entertainment software.

First came a need to justify the expense of the latest technological development. After laying down a significant chunk of their disposable income for the latest, hottest hardware, it was important to have software that utilized that expensive silicon to its fullest. This created a demand for entertainment products that pushed the system, primarily in the form of graphics, animation, and sound. Owners of VGA display boards were quick to complain about EGA graphics. Roland soundboard owners complained about SoundBlaster music, and SoundBlaster owners complained about AdLib music, and AdLib owners complained about internal speaker beeps. This pushed entertainment software to be strong on spectacle.

The second trait common to the 1980s buyer was a willingness to cope with complex rules. After all, in those days, just getting an MS-DOS machine up and running took a certain amount of expertise. Between cable nests and card slot assignments and CONFIG.SYS files and all that other arcana, it was quite a chore. The people who gravitated toward this were people who revel in it, and they brought their values over to their entertainment software. Some of the biggest hit games had rules manuals several hundred pages long!

The third common trait was a preference for puzzles over people. The 1980s buyer was a techie, an intellectual person for whom the computer was a refuge from social interaction. Entertainment software buyers have never been very sociable people, and their values are reflected in the software, which is not so much anti-social as simply asocial. There are no real characters in entertainment software, just a lot of fascinating intellectual problems.

But in the last few years, the primary buyers of home computer have been Late Adopters, who are very different people. Late Adopters don’t feel any need to show off their hardware or push it to its limits. They don’t even know what the limits are. So the Late Adopters will not share the obsession that Early Adopters have for state-of-the-art graphics, sound, and animation. They will certainly enjoy well-executed cosmetic factors, but they will not judge an entertainment software product primarily on that basis.

Even more pronounced will be their aversion to complexity. Let’s face it: entertainment that requires huge time investment is an aberration from the norm. As more normal people enter the market, that aberration will be catered to less and less.

The most important change arising from the demographic shift will be an increasing emphasis on people rather than things. The general public does not devote much of its entertainment time to sterile intellectual puzzles. The overwhelming majority of entertainment, be it movies, books, television, or participatory, is essentially social in style. It focuses on people: their problems, vicissitudes, conflicts, and growth.

Third Cause: A New Class of Developers
During the 1980s, the creation of entertainment software required great technical expertise. This restricted the medium to technologists, who impressed their values onto their designs. There were games that sported macros, games with their own miniature programming languages, and a variety of other techie delights. Entertainment software was created by techies for techies.

But the dramatic improvement in hardware has made it possible for less technically adept people to create worthwhile products. In the early 80s we all programmed in assembly language; in the late 80s we could move up to C; nowadays it is possible to create commercial products with multimedia scripting systems that take care of most of the technical detail.

This has opened the floodgates to a swarm of people whose interests are more artistic than technical. They are creating a whole new style of entertainment software.

Fourth Cause: Stagnation
Entertainment software has grown stagnant over the last five years. There have been no major new ideas in the industry and few ideas that I would call even moderately innovative. Entertainment software design has been entirely a matter of refining old ideas. All the major classes of entertainment software that existed in 1993 (flight simulators, sports games, graphic adventures, role-playing games, strategy wargames, running-jumping-climbing games, "shooters", puzzle games, etc) also existed in 1983. The only changes that we have seen in these ten years have been embellishments. The graphics, animations and sound are better. The games have more internal detail, larger worlds, more complexity. But the basic designs have not changed.

This is truly striking when one considers the gigantic changes in the capabilities of the hardware. In the last ten years we have seen as much change in our supporting hardware as the field of mechanics saw in its supporting technologies during the entire nineteenth century, yet we have taken must less advantage of it than our more imaginative forbears.

I liken the situation to a fault zone before an earthquake. The pressures for change have all built up to titanic levels, but the entertainment software industry is locked in its current state by a variety of inertial factors. What this implies to me is that, when the fault does slip, the earthquake will be of magnitude 8.0 on the Richter scale.

The Shape of the Revolution
So what will this revolution produce? What will be the character of entertainment software ten years hence?

We need merely look at the shift in demographics to predict the shift in entertainment software. The audience will move away from the predilections of the Early Adopters and towards its own interests. Specifically, entertainment software will shift away from its emphasis on coffee-table cosmetics, and it will shift away from massive imposed complexity. It will shift toward more social themes. There will be characters in the software, characters that are not just marketing gimmicks but who actually act like people.

Technological Challenges
Creating entertainment software that focusses on people rather than things will not come easily. It’s quite simple to simulate the flight of a bullet or combat between an orc and a dragon. Presenting compelling interactive characters is a much taller order. Three primary technological challenges must be overcome before we can build entertainment software that appeals to the general consumer.

The first of these I call "artificial personality". This is a technology analogous to artificial intelligence, but its emphasis is quite different. Where AI has a mathematical feel to it, a tendency to tackle rigorously definable problems, I expect artificial personality to be more artistic in style. Unlike AI, there is no expectation that artificial personality will be right or wrong, merely more or less true to character. Human personality is widely recognized to be immensely complex; any system that models it will be an inspired approximation, not a strict simulation.

The second enabling technology is the modeling of the human face. Facial expressions are one of the most important media of human emotional communication. Their importance has been overlooked by the community at large. Yet facial expression is heavily used in other entertainment media. For example, consider Star Wars, widely acknowledged to be one of big hit movies of all time. Consider one of the most action-packed scenes in the movie, the final dogfight sequence culminating in the destruction of the Death Star. Nothing wimpy about that sequence it was pure white-knuckles action through and through right? A careful analysis of that sequence shows that there is almost as much screen time devoted to facial shots as there is to action shots. Thus, even the most action-packed sequences in one of the most action-packed films devotes nearly 50% of its time to facial displays. Someday software developers will learn that message.

The third enabling technology is a language of emotional expression that allows the player of the game to interact with the other characters in an emotionally meaningful manner. The ideal would be voice recognition combined with effective language comprehension, but this is far beyond the reach of AI today. For the foreseeable future, we will need an alternative. This alternative, whatever it may be, must permit expressions more interesting than "pick up the jewel", "open the door", or "fight the dragon".

All three of these technologies must be built before we can move into the larger world of character-based interactive entertainment. Happily, I can reassure you that such technologies can in fact be built, for I have done so.

But before the revolution comes, there may be some bloodshed along the way. I was struck the other day by the eerie similarities between the overall context of videogames today and the context just twelve years ago. Consider the individual terms of the comparison:

Gold Rush: Back in 1981/82, lots of companies were formed to get in on the videogames action. There was a broad recognition that this was a hot industry, and lots of money poured into the field. There were plenty of well-capitalized startups, large established firms creating videogames divisions, and other attempts to cash in on the grand new opportunity. Sounds like 1993/94, doesn’t it?

Schlock Wave: The Gold Rush mentality brought the crassest of opportunists out of the woodwork, and we saw a tidal wave of crappy product pour over the shelves. Yes, there were still plenty of good products, but there were so many bad products that the good ones were lost in the noise. By "bad products", I mean poorly executed products, products with poor design or bad taste. Does anybody remember Custer’s Revenge or Beat Em & Eat Em ? Much the same thing appears to be happening in the videogames field. Survivors of this January’s CES report a decrease in the quality standards for videogames.

Backlash: By 1983, we were starting to see a publicly expressed backlash against videogames. The newspapers carried regular articles about the bad effects of the games on children. I remember Atari growing quite defensive during this time. Ten years later, the issue is back with a vengeance. I’ve seen dozens of television, newspaper, and magazine stories about "the problem of videogames." We see Congress debating the matter of violence in videogames, and the industry is establishing a rating system for games.

It could never happen here: This is another thing that is particularly frightening to me. When I mention the eerie similarities with the first videogame boom, many people laughingly dismiss my observations. "Doom and gloom, Chris! We’re much too big an industry to just dry up and blow away! Don’t worry!" What’s really spooky about this is that these are the same reactions that people had in 1983. Some of the phrases people use to dismiss my Cassandra-feelings are the exact same phrases they used eleven years ago. I half expect to see Rod Serling standing in the corner of the room, bemusedly observing my consternation.

Put these together and it’s enough to give me the heebie-jeebies. Right now, things feel just like 1982/83 to me. That doesn’t mean that we will see a crash but I can no longer rule out the possibility. The last time, I mouthed vague warnings that perhaps our growth might not continue so strongly. I never anticipated the scale of the catastrophe that descended upon us. This time, I still can’t believe it will come, but I won’t be surprised if it does.

Implications for the Revolution
How would a videogames crash affect the multimedia revolution? The short-term effects would be seriously negative, but the long-term effects would be positive. First, we must recognize that a serious reversal in the videogames business would have negative repercussions throughout the rest of the industry. A great many companies in this industry straddle multiple platforms, and videogames often provide their largest source of income. Take away those videogame profits and many of the multimedia projects underway will be cancelled. Second, a collapse in videogames work would cost us the economies of scale that we have developed in software and development technologies. Third, and most important, it would cause a collapse of confidence in interactive entertainment in general, which in turn would mean a reduction in the amount of capital available to us.

The same thing happened with computer games in 1984. The primary collapse came in videogames, but they dragged down computer games with them. There was no fundamental reason why computer games should have been affected; people continued to buy personal computers. But the taint of failure surrounding videogames touched computer games and poisoned their marketplace. I think that we could see something similar happen here if videogames collapsed.

Nonetheless, I would view such a development as a temporary setback, not a long-term defeat. Indeed, a collapse like this would clear out the deadwood in our industry, challenge antiquated assumptions, and encourage desperately-needed fresh thinking. In the long run, I think the future of interactive entertainment is a bright one.