the future of GAMES

WHO WILL BE KING?

Time travel with us and find out what games in the 21st Century will be like. BY OWAIN BENNALLACK

"There's always been a question of what the future of computer games is," says Peter Molyneux, famously of Bullfrog, now co-founder of Lionhead. "That's one of the weird things about the industry. Whenever you start making a game, you know that by the time you finish it, the whole of the industry will have changed."

Indeed, it's difficult to think of any other medium which is as perpetually up in the air as computer games. Movie studios have to guess what kind of films will be popular in two years time, which they appear to do by phoning each other up and agreeing to make the same ones. Rock music depends on the whims of an ever-changing youth, but can be quickly knocked out to order. No-one ever made a computer game in a weekend of whisky, women and recording.

Perhaps only dance music is as dependent on the underlying technology as computer games, which could explain its marginal status and the way it dates in just a couple of summers. So, as we approach the end of the century, can we say anything about the games to come? Can we pick out the Chemical Brothers from the chemical toilets? The Kraftwerks from the cheesy house remixes? Can we say anything about the games of the future at all?

No, but we know a few people who can. In order to discover what the future holds for gamers, we went straight to the folks who will actually be making them. We interviewed over 20. Over the next few pages we'll tell you about the graphics to come and explain why they're not the whole story. We'll look at how 3D games are affecting game engines, the new hardware technologies that will go hand in hand with the new games and the changes to come in the games industry. And we'll put our necks on the line to make some predictions about future games. Who will be king?

Pretty as a picture

Let's begin by comparing this month's PC Review to an issue from two years ago. One thing jumps out at you. The graphics. Games are ever more beautiful, realistic and three-dimensional. The visual quantum leap enabled by 3D cards has taken hold of the imagination of every designer, artist and paying punter.

"I think more of a revolution is actually happening than people realise," says Peter Molyneux. "You question whether Red Alert was graphically stunning or even Doom - now it's inconceivable to release a game that isn't amazingly beautiful."

This rainbow revolution, with the 3Dfx chip as its mascot, has shaken-up the PC gamer's vocabulary. No-one needed to know about hi-colour, Open GL or tri-linear filtering when the real difference between playing Duke Nukem or Quake was your PC@s processor speed. Now, as well as taking the workload off the processor, 3D cards have stolen its limelight.

Intel's Wolfgang Petersen admits it's been the 3D card manufacturers' year. Witness, for example, the dismal interest in MMX. "In the beginning, everyone thought 3D was the big thing MMX would be used for," he says. "But what has happened is that 3D cards have provided a better solution, because they offload the CPU from primitive tasks like putting textures onto polygons. That's the reason why we'll always have graphics cards doing these dedicated things."

[Panel]

The Crystal Gaze: The Sensory Experience

	Movie quality graphics	VR in the home	Haptic interfaces	Speech in games
Glyn Williams	2030	Never	Never	1999
Bill Stealey	2020	Never	Never	2005
Andy Hollis	2000	2001	2005	2000
Laura Wheeler	1998	2002	2007	1999
PCR Predicts	2015	2020	2010	2000
	Glyn Williams thinks we'll probably never achieve true photo-realism, but that Toy Story style visuals could arrive within a decade. Our experts agree that restricting the player is the key to further advances. "The more you let him range around into different areas or altitudes, the harder it is to give the player a real experience," says Andy Hollis. In contrast, Laura Wheeler thinks we're almost there with Blade Runner.	"There will always be demonstration products for £10,000," says Bill Stealey. Otherwise, he and Williams are at odds with Hollis and Wheeler's estimations, arguing that VR screens just can't compete competitively with flat-screen technology. PC Review doesn't believe that helmets need to match flat screens for resolution. But the visual quality must not be so poor that the added benefits of immersion and control don't compensate for it. We think two decades should do it.	"It's one thing to get the visuals, it's quite another to get the input," says Hollis. He points out that designing a feedback joystick for a flight sim is far easier than making a decent Haptic interlace for Quake. Williams is even more sceptical. "Motion flight sim platforms are already available but no-one's going to buy one," he says. "We saw it on the Nintendo, it was crude and rude," Stealey reminds us. "Most games won't be able to handle it "	"The radio concept is coming and it's going to be a huge step forward for multiplayer gaming," says Hollis. "But voice recognition is a little bit further away in games." Incidentally, only Dave Perry rallied unprompted to speech recognition. "I already have a game design using speech," he says. "But we need a chip to tell the game what you said, without impacting the speed of the game. Then, by God, games will change."

When the first accelerators arrived people were most interested in the number of polygons they could shift. But a few games, like Eidos' Terracide and Psygnosis' F1 quickly revealed the power of the card's more specialised silicon. Special lighting effects and texture filtering have proved as important as faster frame rates. With nearly every game sporting the 3Dfx look for Christmas, it's worth remembering just how amazing these effects seemed at the start of 1997. Fortunately, we are going to be amazed again.

"Everyone is saying they can do millions of polygons with 3Dfx, but very few people are looking at image realism. Specularity, how light bounces and reflects of things. These are all important." So says Glyn Williams of Particle Systems, who's upcoming game, I-War, is already pushing current technology to the limit. He points out that today's games are still a long way from the computer graphics of movies like Men in Black.

"There are a number of key technologies which you see in feature films which make the image look real," he explains. "Motion blur is one. If you're using film running at 25 frames per second, but it's motion blurred, you don't have a discernible frame rate. If you look at a monitor running at 30 frames per second, you still see discrete images."

The effect is caused by the finite length of time the shutter of a movie camera is opened. If an object moves whilst being filmed, it will appear as slightly blurred, as familiar from photos of cars streaking past. "If an image doesn't have that, it's very harsh on the eyes," he says.

Williams predicts we'll eventually see motion blur hardware supported by 3D cards. The most common blurring algorithm involves rendering 15 frames to make each final frame, by blending them all together. The process obviously takes either 15 times as long or 15 times as much processing power per frame. This, says Williams, together with the sort of polygonal processing power which will see real-time shadows creeping up the walls of some future Quake clone, is probably still about five years away.

Long before then, Intel's new Accelerated Graphics Port, found in the Pentium IIs, should inspire a new generation of 3D cards. The AGP, supported by the new 440LX chipset, gives graphics a direct path to the PC's RAM, instead of limiting them to the video memory. This will first make 3D cards cheaper, since low-end cards won't require expensive VRAM. But the ramifications could be much wider. Geometry processing will also be supported by the next generation of 3D cards. Current models help painting and filling polygons, but they still leave the CPU to work out where those polygons should go.

Polygons made to order

In the interim, software solutions might provide better graphics now. Shiny Entertainment is working on Messiah, the first game to feature its new, potentially groundbreaking 3D engine. Written by ex-Scavenger programmer Michael Person, the key to the engine is real-time tessellation. This technique enables the number of polygons in 3D objects to be increased or decreased as required In contrast, most 3D engines use three or four fixed polygonal models, which are swapped as an object changes size.

"Real-time tessellation is very important because of the ridiculous spread of processing power," says Shiny's leader Dave Perry. When faced with a less capable machine, the engine decreases the polygons to order."If you play Messiah on a P90 and then on a P200 you'll see it run at the same speed. And if you run it on a P300, you'll see polygons that we at Shiny have never seen before."

Another of the engine's advantages is its ability to meet changing game demands. "We can rob Peter to pay Paul," says Perry. "If we want a big explosion on screen, we can take polygons from other characters to create the explosion, so the game speed stays constant. Or if people walk past the camera, they don't turn into giant pixels, they stay detailed."

"The next thing we're doing is real-time tessellation of worlds," Perry continues. The game to use this technology will be called Sacrifice - Shiny's first multiplayer title. "In our office we can see for nine miles at 60 frames per second. John Carmack is quoted as saying that's the future of games - we have it in our office. If you look into the distance and you see mountain, you can walk for 30 minutes and you'll get to that mountain."

Perry is dismissive of other companies' work with voxel technology (such as found in Westwood's Blade Runner), arguing that polygon-munching 3D cards have decided the battle between voxels and polygons forever. He claims real-time tessellation is certain to become standard, and that Messiah puts Shiny a year ahead of the industry. In fact, Shiny is already moving into the next generation - real muscle and skin definition.

"Where a leg would normally be a lump of steel which bends at the knee, we have a bunch of muscles and skin which move," Perry explains. As well as improving realism, the advanced graphics format enables Shiny to plug 3D characters into any position and ensure they'll bend and stretch appropriately. As such, it might prove a viable alternative to motion-capture.

Despite Perry's plans for the future, few people who've seen Westwood's Blade Runner in operation would label it yesterday's graphics. In fact, it's probably the most attractive title we'll see this Christmas, and its millions of voxels make a compelling case for software rendering and voxels.

"Yes, with Blade Runner we did a complete left turn with the technology," admits Westwood's Laura Wheeler. "We didn't try to put band-aids over graphics issues with hardware acceleration. Intel's engineers told us we couldn't do what we wanted to do and we did."

What they did was to create a game with in-game graphics to rival its pre-rendered FMV cinematics. Millions of voxels crowd the screen while the camera tracks characters into previously fixed backdrops, dappled with a host of lighting effects. Newly patented 'true-colour emulation' enables Westwood to render colours in and out of the screen over time, so there appears to be millions of colours on-screen at once. All this on a P90, with a graphics engine that might actually slow down with 3D hardware.

Watch what space?

Wheeler's comments are slightly disingenuous, since when Westwood started work on its graphics software, 3D cards weren't really a force. Changing hardware is always a problem for developers and it might get even more complicated soon, as we start to see the first movements away from the age-old computer monitor.

Microprose co-founder, Andy Hollis is now a Longbow 2 developer at EA. He believes an image is only as realistic as the VDU allows. "You can have a postcard of a perfect photograph held up in front of you, and you don't feel like you're there because it's a postcard," he points out.

As well as improving immersiveness, new display devices can improve our control over games, Hollis argues. Virtual Reality headsets, for instance, have probably not yet had their day. "If trackers ever get crisp and the visual quality gives you more than a big headache, they will really take off We need the movement of the head to be an input device, so you're controlling the game in an intuitive fashion." Others are more sceptical:"The technology won't ever be there," argues Interactive Magic's CEO Bill Stealey. "The resolution I can get on my flat screen will always be so much better than what I can get on two screens, in my helmet, at a reasonable price."

Stealey suggests that large screens, possibly curved, are more likely. Hollis agrees that they're a useful stop-gap, but believes multiple monitors provide a better solution. To be supported by Windows 98, such monitors provide a scaleable display solution. Three medium sized screens might make a nice display, with a fourth used for instrumentation and avionics. Such a setup would be very similar to those found in professional flight sims

Pretty vacant

But are relentlessly advancing graphics all we can look forward to? Few developers seem to think so. In fact, there's a general feeling that gameplay is due for a renaissance.

"Everybody's doing these fantastically real things," says Simon Byron of Bastion, a leading new media PR agency. "It will soon get as close to reality as you can get. There's only so much motion capture you can do, there's only so fast a frame rate you can get into a flight sim. When I first got into computers with the ZX81, a driving game consisted of an inverted A moving between some blocks."

For Byron, the graphical war has been largely won, meaning a decreasing payback for developers concentrating on graphics over gameplay. "Actua Soccer 2 is not quite like watching soccer on Sky but it's getting close," he says. "Once people hit that point they're going to have to innovate."

Ironically, stunning software like Shiny's real-time tessellation technology might also take the emphasis away from graphical innovation, "Basically, everyone wants more polygons," says Dave Perry. "Lara Croft would have about 700 polygons on her, which gives a basically real looking character. We use 180,000 polygons, compressed into our weird Shape format."

Of those 180,000 polygons, up to 8,000 survive to potentially feature in Messiah's on-screen characters. On today's average PC about 2,000 will. It's like scanning a picture at a certain resolution. Currently, the Shape format 'scans' 8,000 polygons from the 180,000. "Once everyone's got as many polygons as they want, the argument's over," says Perry. "We can put photographs onto models as well. Once you can do photo-realism, creativity comes back. Once you've done a photo-realistic soccer game, it's very tedious when ten more hit the market. At the moment the big race is to get there first."

John Hare created Sensible Soccer on the Amiga. It's a sorry sign that a decade on, his top-down sprites are still arguably the best computer footballers around. Now working on Sensible Soccer 2000, he says his aim is to blend the gameplay of Sensi with today's graphics."Sometimes understating the graphics is necessary to get the best out of gameplay," he says. "You can have both, but it takes a lot of design."

But there's a good reason why graphics have continually improved at the expense of gameplay and that's the continuing evolution of the PC. The old 8-bit and 16-bit machines were effectively set for life. The games of the future will be played on PCs that we can't yet imagine.

Still, as 3D cards and 3D games have come to prominence, so Microsoft and others have worked to provide programmers with APIs, making such 3D graphics coding easier. "The programming task is changing," agrees Andy Hollis. "Parts of the game are going to be dealt with at a higher level now. The DirectX stuff has helped in that way. We're going to be spending more time on the parts that matter - the gameplay and the AI."

Writing a tight software rasteriser just isn't important in the age of 3D cards. Just as happened with soundcards, graphics will be done almost wholly in hardware. Basically, if id Software's John Carmack isn't writing geometry processing routines for 3Dfx in two years time, he's not going to be involved in computer games any more.

OK Computer

Where then, as Hollis asks, are the areas for innovation? While we ponder that point a little tidying up might be in order.

Cavedog's Total Annihilation looks like breaking all previous standards in the real-time strategy genre, with a host of new features including the genre's first true 3D terrain.

Yet what really sets Total Annihilation apart is its humble interface. At every stage it makes life easier for the player. Compare this to Acclaim's Constructor, a splendid game, full of ideas, yet rendered almost unplayable by its appalling interface. Good interfacing doesn't just make the game more enjoyable, it enables you to achieve more. By allowing you to stack orders and automate repetitive tasks, Total Annihilation opens up your tactical options.

"Interface is absolutely one of the top three changes you're going to see in games," says Total Annihilation supremo, Chris Taylor. "In Total Annihilation, if you place the cursor over a unit, you see what its status is without clicking on it. You can wave your cursor over a group of units and in a matter of seconds you're educated. When weapons fire, you see shells on the radar map. Normally, you wouldn't care, but in our game you see dots moving down and you know they're fighters, then you see dots streaming up and you know your missile towers are working. You don't have to go there now."

In Red Alert and Interplay's MAX, it's possible to build exotic defensive structures. But nobody ever does because of the time and tedium involved. In Total Annihilation, one unit can do all the work for you. "In ten seconds you can lay down a path for a string of invulnerable dragon's teeth that go right across the map," says Taylor. "You can then go away and they'll be built over 20 minutes. Which means you'd actually bother to do it "

Automation and interfacing might not be as glamorous as graphics, but just remember how long Apple kept ahead of the PC by being easy to use. If we want to get more from our games in the future, one way is to get better control over them. As Taylor says,"There are huge lessons to be learned." Control is only half the story. Sending a unit back to the field hospital is of little use if he decides to get there through a minefield. Is there any hope for artificial intelligence?

The thinking man's thinking machine

Gremlin's Ade Carless thinks there must be. He's been working on Hardwar, a space trading game whose artificial life features have caused continual delays. But he's unrepentant. "AI is paramount", he says. "People are looking for more sophisticated things which will react to what they're doing. Netplay is big, and people want that kind of erratic experience reflected in a single-player game."

Jay Wilbur, now at Epic Megagames, was business chief at id Software when it released Quake. He says the AI in Unreal, Epic's new first-person shooter, will do for single-player gaming what Quake could only manage for Deathmatch.

"In Unreal, you'll be able to play against real intelligence," says Wilbur. "As opposed to in the past where you've only played against pure stupidity - when monsters come awake when they see you - or mock intelligence where they're running through points on the map."

[Panel]

The Crystal Gaze: Technical Talk

	3D cards become standard	The PC kills of consoles	The home PC converges with the TV.
Wolfgang Petersen	Now	Never	Now
Richard Darling	1998	Never	Never
PCR Predicts	1998	Never	2020
	"All the graphics cards you can buy now support 3D cards more or less well. 3D for me is a done deal," says Petersen . What's particularly interesting is that he envisages 3D to move out of the game ghetto to infuse everything else on the PC, from educational software to the Internet and Windows. "Interaction and navigation in the future will all be 3D. There will be no buttons and no menus soon," he predicts.	"I hope there will always be consoles!" says Petersen "The PC was always catching up, but now we exceed the quality of the consoles. But the consoles will strike back." Darling's support of consoles is less surprising than Intel's. "The consoles will always have the advantage of not carrying the extra burdens of the PC," he says. Imagine a new console packed with half a dozen 3D cards to understand why we'll forever be leapfrogging each other.	"It looks like it's via the PC that digital TV will come to the home," says Petersen. "Are the PC manufacturers going to dominate this market? No, the TV manufacturers will have to support digital TV." Darling is less sure. "Technically it's possible for all sorts of things to converge," he says "Televisions and video recorders are always right next to each other, but well over 90 per cent must be different units." We think they'll converge when all home appliances converge. DVD might bring that closer.

Epic puts such a premium on AI that it bought in Steve Poldge, who created the Reaperbots for Quake, to write Unreal's AI. The result is a new level of intelligence. "You can assign all sorts of different AI classes to the creatures. You can say 'Here, guard this cup.' And they won't mess with you until you mess with the cup," explains Wilbur. "There's a hierarchy within the creatures. You walk in on them and a boss shouts a command - you don't understand it because it's in alien but in time you will. That command was maybe 'flank him' and you'll see the group spread. These creatures maybe know a way to get around to the back of you. That's the kind of intelligence you'll see more and more in games."

We've played early versions of all the next-generation Quake clones. The one that's most impressed us is CUC's Half-life. At a recent demo, a PR representative was pointing out Half-life's graphical quality when suddenly a soldier jumped out and took a shot at him. He then disappeared behind a pillar, only to return later with several mates. In just a few seconds, our demands from first-person games changed utterly. And let's face it, running away when shot at isn't particularly complex. Yet it's taken so long to come.

Many people dispute what we actually mean when we say AI. Some developers, including Dave Perry, Jay Wilbur and Ritual's Harry Miller, speak of the improved intelligence of their protagonists. You might shoot one of Messiah's characters in the leg to get him screaming and lure his friends to a trap. But while very impressive, Peter Molyneux questions whether we should label such technology AI.

"That's not AI, that's nothing to do with AI," says Molyneux. "AI is a way of the machine looking at how the game's going and modifying the game accordingly, learning and adapting itself to a particular player. That's what's exciting. Character AI, like having some creature moving around a maze is incredibly complicated and incredibly boring."

Molyneux says that Dungeon Keeper's greatest achievement was the creatures' path-mapping. And it's true - we've seen most of Dungeon Keeper in some guise before, but the way the imps carefully prioritise building projects or the way the demon spawn scurry efficiently between the chickens and the training room smacks of smartness. Yet no-one really cared. "I said there's a revolutionary navigation routine in Dungeon Keeper," says Molyneux, "but it just doesn't sound sexy. And it's not AI."

While he feels we should be clear about the differences between real AI and computer game AI, Molyneux doesn't believe the former is yet worth the cost. As evidence, he recalls an AI show in California. An eminent professor from Carnegie-Mellon University was showing off his latest project. Running on a supercomputer, it involved three creatures who walked about pulling faces and scaring each other away. Bullfrog showed him Theme Park, with its hordes of burger-chasing hedonists, and he nearly fainted. Of course, the Carnegie-Mellon project was true, scaleable AI the creatures were really learning. Interesting, but irrelevant right now to game designers.

"All the machine has to do, through tricks or real routines, is things that make you believe it's realistic," says Molyneux. "A creature doesn't have to think, it has to look like it does and thank God for that. People have been working on real AI for years and years, and nothing changes."

Bring in the rocket scientists

Instead of merely improved AI, Molyneux suggests that the real advances in the future will be the physics engines which power game worlds.

"As part of the gameplay, we're going to have to model a real world within the game, so that the computer knows what each thing is made of," he says. "Someone throws a hand-grenade at a car and the car explodes and a piece of car flies off and smashes a window and a shard of glass flies off and hits someone on the head and they fall over knocking a tree and a leaf falls down. That's the level of detail we need to go into."

[Panel]

The Crystal Gaze: The Future of the Industry

	First £10 million game	Games industry overtakes record industry	Most games sold over the Net	The industry consolidates like films and music
John Hare	2000	Never	2003	Now
Brian Walker	1998	1998	Never	1999
Ade Carless	1999	1999	Never	Never
Chris Taylor	Now	2007	1999	2001
Christiaan van der Kuyl	2002	2017	2000	Now
PCR Predicts	1999	2010	2010	1999
	"Right now we're seeing development being relatively austere and marketing going through the rough," says Walker. "Twinsen's Odyssey has a ten million dollar marketing campaign behind it." John Hare thinks that media convergence could drive up budgets. "While the industry continues along its myopic route, it can't justify it," he says. Chris Taylor says we're probably there now. "Final Fantasy VII is getting up to that. For the average game, four years."	Wildly different opinions suggest that while all developers would love to be rock stars, it will be some time before they get the limos and girls. The most commonly cited reason for pessimism is music's universal appeal. As more people get into gaming though, dominance seems inevitable. "Games cost a lot more, so the turnover is much bigger," says Ade Carless. But van der Kuyl sounds a note of caution. "The girls will never come," he says sadly.	"The problem is we're going to DVD," says Taylor "So by the time you can reasonably download 600Mbs of game, they might be 4Gb." Brian Walker agrees. "3D cards will soon quadruple the size of games," he warns. Meanwhile Ade Carless doubts there's even a demand. "There's something nice about browsing on the shelves," he says. We think software shops are doomed, but we'll need faster Net connections. Our estimate anticipates widespread satellite Internet links.	"People have got strong enough lineups to continue to exist. The gaffers of these companies are strongly independent," argues Ade Carless But with Virgin up for sale, everyone else sees the writing on the wall. Walker predicts the current rate of mortalities ad acquisitions will leave just a few major publishers standing. Van der Kuyl thinks we're there now. "There are five or six major publishers who own the sales and distribution channels and everyone else has to hover in the niches," he says.

"We've done 3D terrain now but we can't go into another dimension," agrees Chris Taylor. "We need to take the third-dimension to its fullest extent." Taylor says Total Annihilation shows the sorts of thing we can expect."Because we went to 3D we used real 3D physics. We have missiles that intercept missiles - it's very hard to do that in a fake world. When a shell hits a tank in the rear, it's going to do different damage to if it had hit the tank in the front. As a strategist, I now have to remember to hit the soft sides of units. That's the level of detail we're going to see."

Another company working on a realism, albeit of a different sort, is Interactive Magic. Its flight sims, while boasting perfect flight models, have often been criticised for looking old-fashioned. Yet Bill Stealey explains there's a good reason for that."Some of the things these guys are doing look beautiful," he says of his competition. "They look like paintings. Nothing I ever saw flying any aeroplane though, and I've got 5000 hours of flight time."

Instead of optimising its graphics engine for looks, Interactive Magic has concentrated on exploiting real-world data."With iF22 we used a new 3D system that the US government is now negotiating with us to use in a flight simulator for training US pilots," he says. "And this is not the fake 3D you see in other people's games. This is real Bosnian terrain with real elevation, real rivers and lakes."

Stealey says that there's no limit to the eventual accuracy we could see in flight sims, but that current technology is holding us back. "It takes us 900Mb to do Yugoslavia. And we're only using 100-metre data. The military has data accurate to one metre. That's ten to the ten as much data."

The general problem with modelling the real world is its complexity. When 3D graphics first began to oust cell animation from computer games, scores of previously competent 16-bit programmers were forced to seek new careers. They simply didn't have the mathematical know-how to work in 3D. And 3D polygon manipulation is a reasonably simple task. Working out how a table should appear on screen is a trivial compared to modelling how the people sitting around it are thinking, or what will happen to the furniture if a fight breaks out.

"I think there's going to be some very unpleasant changes," says Molyneux. "We've been talking about some amazing things, and that's going to mean we'll need some amazing people coming into the industry. We need to attract people with a real knowledge of physics and dynamics and mechanics."

"Every time you do something like that it changes the industry. It makes games one hell of a lot more expensive. The more expensive they get the more there is at stake and the less likely you are to get away with interesting concepts. It's a bit like Hollywood - when they do a $50 million film, they've got to make something they know will sell everywhere."

So, ironically, just as programmers are going to be free to spend more time on gameplay, we're hit on two fronts. Games will need to slowly come to model the real-world, and games will need to appeal to the widest possible audience to justify their cost. The result could well be more restricted games (modelling just a bit of the real world, like a race-track or a sports field, is far easier than the whole globe) which are designed to please the lowest common denominator of punters FIFA 2000 anyone?

Big games or bust

According to Eidos' Brian Walker, we are already starting to see the effects of this trend. "Two years ago a $1 million game was a rarity, now it's the low end. The age of the $5 million game as a standard is about five years away. Players simply want so much that you need 10 or 15 people on any development team. Simcity 3000 has 15 programmers alone."

The computer games industry is perhaps uniquely unfathomable. The time it takes to make a game means the shop shelves always lag behind people's perceptions by at least 18 months. Consider all the Doom clones that arrived well after Quake. Or this year's flood of mediocre titles, the result of projects first kicked off when the arrival of CD-ROM seemed to herald an endless summer for PC games. Today there are simply more games then games players, or indeed spaces on the shelf. Walker estimates that the ratio of games to retail slots is about three to one. If a game doesn't succeed in a week, it's chucked into the bargain bin.

"Sequels are the next step. Anything that's not a sequel is going to have a hard time," says Walker. "When you see the bottom line of a lot of the major publishers, who are losing money regularly, they may not be making bad games. They've lost the ability to get their product on the shelf, or they've lost their ability to hang on to their best technical people. Once that edge has gone, it's very hard to get it back."

Sensible Software's John Hare sheds few tears over the casualties in games publishing. "The marketing first approach works really well for short-term sales. Everyone's got what they wanted, short-term sales, on the back of what were long-term sales, on the basis of games which people spent years developing and people played for years."

"We can't support any more B titles, or even A titles, the industry is definitely going to triple-A titles," agrees Westwood's Laura Wheeler. "Less titles, quality titles."

Coming from Westwood, perhaps the most consistently profitable development studio in PC games, Wheeler's sentiment is perhaps unsurprising. It's similar to Nintendo's mantra, which has seen the N64 grab a significant proportion of UK gamer's money, despite the appearance of only six games so far.

Another thing which could shake up games publishing, and thus games in general, is the Internet. After all, why should developers let publishers ship their games all over the world or fight for shelf space when punters can download games directly from their Web site? Well, where are developers going to get the money from? Publishers, of course. Those publishers are likely to demand significant control for their millions, bringing us back to square one. The Internet could change game pricing in other ways too. "I'd dare to say that in the future almost all games will have online aspects to them," says Peter Molyneux. "There's one slight snag though, and that is that no-one's worked out how to make any money out of them."

In the States right now you can play Red Alert multiplayer free via Westwood Chat, but you're limited to two players. Alternatively you can play via TEN which will charge you an hourly subscription, or Mplayer, which lets you play free and charges advertisers instead. Which, if any, of these methods will prevail over the long term is anyone s guess. Yet the opportunity is there to revolutionise games. Over 18,000 people play Red Alert every day via Westwood chat. Not a bad incentive for cracking the puzzle.

Keeping a cool head

One company with a lot of new ideas is vis Interactive. The brand new Scottish start-up is working on a two eagerly-awaited games, Earthworm Jim 3D and a new and wonderfully inventive title, HEDZ.

"HEDZ, is the first of the collect-'em-up genre!" says Vis' hyperactive CEO, Christiaan van der Kuyl."The aim is to collect 225 different heads which you can wear. Each head gives you two different properties, and you can then go and fight against other heads."

What makes HEDZ yet more interesting is its online capabilities, and corresponding charging method. "Everyone saw online as the way forward, but then people just put normal games online," says van der Kuyl. "HEDZ is designed as something from the ground up that can go online. We've got the HEDZ stock market, where you can bid against other people for heads, you can buy random packs of heads and you can put your doublers up for sale."

Thus HEDZ is a cross between marbles and football sticker albums. By turning its game's stars into collectible items, Vis hopes to replace boring charging models with something new, lucrative and appealing to gamers. "You get this brilliant gamers thing, where you think 'I'm the best HEDZ player in the world, I'll never have to pay, I'll always win other people's HEDZ,"' says van der Kuyl. "In actual fact 90 per cent of the population will always be buying more heads."

[Panel]

The Crystal Gaze: Guessing on Gameplay

	Single Player Gaming Fades	AI takes up most processor time	Point and Click Adventures Die	The public creates its own games
Harry Miller	Never	Never	Never	Now
Charles Cecil	Never	Never	Never	Never
Dave Perry	Never	1998	Now	Never
Peter Molyneux	2002	2000	Never	Never
PCR Predicts	Never	2000	2000	Never
	"That's complete rubbish," says Charles Cecil. "When I come home from work and put the kids to bed, I don't want to play against other people. And I want to be able to save, and start again where I left off." Astonishingly, it's Peter Molyneux who's most cynical. "There's only one single-player card game, which is solitaire," he says. "But no-one would think of playing solitaire when people are around. Computer games will become like that."	Our experts worry graphics will always suck up power. And Charles Cecil isn't even sure more AI cycles are necessary. "Within a fighting game, you can give it the intelligence to fight really well, then a little AI to learn after that," he says. But Peter Molyneux believes the shift's inevitable, particularly once worlds start getting more complex. "We're going to be approaching 40-50 per cent within the next three years," he predicts.	"Real-time adventures I'm all for," explains Dave Perry. "But the days of pre-rendered Mysts are gone." Miller is less sure, saying that while first-person games will soon match the graphics and provide a more immersive experience, the genre is still unique. Molyneux believes adventures have a future, but they must change. "In fact, they've got to have a renaissance," he says. Charles Cecil agrees. "I'm nailing my colours to the point and click adventure's mast," he says. "They're going to evolve".	If we combine off-the-shelf engines with Quake-style customisation, do we get a recipe for gamemaking at home? "Soon people will be able to create their own areas and experiences" says Harry Miller. "When a million people are working at this, things happen really fast." Charles Cecil is more sceptical. "You don't have the teams and resources, and you can only do it on yesterday's technology. Look at Jon Romero's Daikatana. The guy's an industry guru, but he's using yesterday's technology and he's being slaughtered".

Van der Kuyl thinks the online gaming market will suddenly explode like the Web did three years ago. "The smart people will start building for it now." Nearly everyone agrees that the potential for multiplayer gaming, whether bought off the shelf or over the Net, is boundless. Games already in development will soon bring persistent megaplayer gaming to our PCs, before some people have experienced multiplayer gaming. And it's not just about having more opponents. It's also about creating new worlds to fight in.

"In Unreal, you can go not only level to level on the server but also from level to level across servers," says Jay Wilbur. "Unreal also comes with the ability to make your own levels very easily. So now you can create your own castle, and post it to the Web. You'll have links coming in from somebody else's levels and now you've got real estate on the Net. You can have miles of servers connected together. You could be walking for days through totally new stuff"

Eventually, this should move beyond merely blasting through various levels. Communities might eventually be formed, although for Unreal at least, that's likely to be a dangerous community of trigger-happy maniacs. Wilbur: "You'll be able to populate your own levels and tag me as a friend and him as an enemy. I have the keys, I can come in and we can hang. But when he comes in your minions will go and kick his ass."

Chris Taylor looks forward to a Total Annihilation sequel featuring 25 people on each side. He envisages huge worlds, where individual players assume responsibility for various areas of the game, such as resource management or weapon construction. But he's sceptical of totally persistent online games, where your world continues even when you're offline and an invasion might see you being emailed or even paged. "When you build a game that handcuffs you to the PC, it changes from being pure entertainment and into a job," says Taylor. "And who needs more responsibility?"

A word to the wise

In fact, some people have been making persistent online worlds on the Net for years now. Dr Richard Bartle, the inventor of MUD, watches the development of new online technologies with amusement. All his games are text only, but he thinks that far from being past it, text belongs to the future of games.

"We started off with people playing simple text games," he says. "They soon want more of an experience so we start putting graphics into games. Very impressive but they're not believable. So we go into more believable graphics - not photo-realistic, but that will come. Then people will be getting the little 3D goggle things. Then we'll get the full body-suit, then we'll get control of peoples senses but still people won't believe everything in the game because they'll still be interpreting things. What we really need to do is plug straight into their brains and imaginations. A technology that will bypass their senses so they have to see things as they really are. The kind of power we'll get from that will be remarkable. The technology was invented about 5000 years ago by the Assyrians and they're called words. They can speak directly to the brain. And that's why text has a future."

It's an extreme view, but people have been playing Bartle's games for nearly 20 years now, more then any other game can boast. A version of MUD recently put onto Wireplay has proved very popular. So, is the thrust of the rest of the industry all wrong? "All the things we're seeing are fine for an industry that's going forward, but if you follow the logical extent of it, we'll just end up in the past," says Bartle. "They don't give the experience that they could give if people looked at what was behind what was being presented instead of just looking at what's just being presented. What it looks like is what makes you buy it, it's not what makes you play it."

Other people questioning today's interactive technologies are Cyan's Robin Miller and Richard van der Wende. They've just finished Riven, which brings some of the most impressive pre-rendered visuals we've ever seen to the PC. But they're not exactly enamoured with technology.

"The future of games is just like the future of any other artistic medium," says Miller. "It's not always dependent on technology. The mistake we often make is thinking that because a game is on a computer, it's the technology that's important. It's still about concept. That s what affects people.

Miller says that while the interactivity of computers remains uniquely inspiring, he's come to doubt the power of the medium to engage people's emotions."I don't think Riven is as provocative as a movie, and I'm not sure the medium even has the potential at this time. I think that story works because of people and this medium cannot express people in detail. You can only get a gross understatement. Story is about people, and they aren't here yet."

Miller and van der Wende question whether non-linearity, which seemed so exciting when creating Myst, can actually deliver a compelling story when the main character is effectively out of control all the time.

"As we made Riven we became quite intimate with the limitations of the medium," says van der Wende. "I think that Robin and I both have a desire to affect people on a more profound level. I'm not sure that the technology can achieve that."

Bastion's Simon Byron agrees that emotion somehow has to be brought into games in the future, "How can we get people more emotionally involved in a game?" he asks, "To me the perfect game would be something you could only play once, just like life. With stuff like Quake, there are no real pressures to get things right first."

Byron believes that the first a-life programs might show the way forward. "Take something like Dogz," he says. "Because you grew this little thing, you'd really care for it and you wouldn't want anyone else messing with it. We need more emotional attachment."

Story might be a harder nut to crack. Indeed Dave Perry wonders if story has any place in computer games at all."I think that interactive storytelling is a farce," he says. "I sit in a room and I listen to 100 people and they all agree, interactive storytelling is wonderful and games suck without the story. It's just not true. I want to be the story. I don't want to only have an ability in a certain place in case I break the story. I don't want to not be able to open one door until I open another one - I just want to kick down the door and work it out later."

Charles Cecil, Revolution's adventure guru who's just finished Broken Sword 2, stands at the other end of the spectrum, but he's sympathetic to some of the frustrations. "The problem is people want to feel like they have great freedom, but they also want to be told a good story, and the two are diametrically opposed. What we do at the moment in adventures, because it's the best we can think of, is we give pseudo-freedom. That's because before you move a plot on, you have to know what the player knows - and you have to force them to know what you need them to know."

Despite the difficulties, Cecil believes he's working in the most exciting genre of computer gaming. "The story started with cavemen, and it was linear. Then the Greek plays came, then the 16th and 17th Century dramas then books and then films but they've always been linear. Now we've been given an opportunity that no-one's ever had before. I don't know what the answer is, but I think that's where the growth is going to be."

The game which might just show the way is Blade Runner. "I've played the game five times in the last two days and there's been different dialogue between the characters every time so far," says Westwood's Laura Wheeler. "There's not a limited branching, it's endless."

Even more than the graphics, it's this story engine that makes Blade Runner such an important game. All its characters have their own AI and agendas. Some are replicants, some not, yet they can all swap information about events that have happened in the game. In short, they walk and talk, and word soon gets around. Information is effectively treated like magical items in MUDs, or variables in a programming language. Anyone could know anything, but then again they might not. Knowledge will affect their actions, which in turn will change the situation once more. The geography of the game is restricted - you won't suddenly find yourself catching the plane to London - but the benefit is true freedom.

"It's a very real-time, organic game. You can save the game at any time, but you can play five minutes and then go back to the save game and play another game and have a completely different result," says Wheeler. She even compares it to Command & Conquer, and predicts that Blade Runner will do for adventures what C&C did for wargames.

The kingmakers

So, will the games of the future all be adventures? We're not convinced, although we suspect that perhaps the most creative will be. Indeed, we're only a little more enlightened then when we began our investigation.

Our soothsayers have cleared a few doubts but raised so many more. Are we really heading for a gameplay renaissance? Does real-time tessellation herald a new era in graphics engines? Will artificial intelligence ever switch on?

All we know is that our gaming wallets are ever more valued by the industry's dwindling yet bloating publishers. As budgets increase, games must reach a wider audience of ever older and more discerning players.

The increase in online gaming will see more people playing games before paying for them, which means that games must hit the spot straight away. And the industry has discovered it can only survive through long-term successes.

In short, the industry has grown up. With publishers finally listening to our needs, we'll risk one firm and final statement. In the future, it's the gamer who'll be king.