Hideo Kojima delivered his GDC keynote this afternoon, and we reported live from the event on this page.
It was the Metal Gear Solid creator's first time talking at the San Francisco event, but his nerves were absent and his message clear: "90 per cent of what is considered impossible is, in fact, possible."
Read on for the full transcript, the earliest post of which is presented first.
Our live coverage of this event has finished.
Good morning/afternoon from GDC! It's another perfect day here - San Francisco doesn't seem to do bad weather.
There wasn't anything like the huge queue that there was for yesterday's Nintendo keynote, but the huge Esplanade Room is filling up fast.
We've got front-row seats again, and there's a hum of anticipation in here, even though we've heard there'll be no new game announcements. It must be the Kojima Effect.
We can spot Suda 51 sat just over to our left - officially Japan's nicest videogame developer (and that takes some doing). Also the coolest. Only Suda could rock a black-and-white newsprint hoodie in his 40s and pull it off.
It's all Daft Punk and Gorillaz and sunglasses on heads in here. Not as many suits about as for Iwata yesterday.
The lights are coming down and GDC director Meggan is on stage introducing the keynote. Kojima is "renowned as one of the most influential videogame developers in the world".
Here he is! Dressed only as a rock star Japanese developer does - satin jacket and black-and-white spec frames.
He's talking in Japanese through an interpreter over the PA, and thanking everyone for the Lifetime Achievement award he picked up last night.
He jokes that he came because E3 has lost its punch over the years - and because they said they'd give him an award if he came.
It's his first time at GDC - he reckons "the atmosphere is really hot".
He's not going to talk about the technical side of game design - it's more about his philosophy of game creation. He promises entertainment - but no free games as presents. Boo!
He asked a couple of his friends what GDC was like. He was told that people leave boring sessions after 10 minutes, but there's nothing else going on right now, so we're stuck with him.
On with the presentation. He defines Revolutionary Creation as doing something nobody has done before - making the impossible possible (which is the title of his keynote).
It's a path - with a 2D black-and-white snake walking along it. Obviously. Sometimes you face little obstacles (a black box on the screen).
Snake vaults over it, and here comes a little 8-bit Mario, who jumps over with ease.
Now there's a huge wall. Snake can't make it. Mario comes along, turns into Super Mario and leaps over, much to Snake's dismay.
Snake thinks, he's not Mario and he can't jump over. Poor Snake.
Kojima thinks it's all about changing your preconceptions. Snake's stereotype is that he thinks he can't jump over - but we need to change visions.
Snake thinks of new ideas. Why not pole vault over the wall? Destroy it with a gun? Or fly over with a balloon? Or suddenly find a cartoon door in it? Or flood one side with water and float over?
But when you change the camera to a 3D view, you see that maybe you could walk around the wall, or dig underneath it. Making the impossible possible is changing your view, coming up with new ideas.
Kojima is clearly going for the world record in GDC keynote extended metaphors here. Iwata just doesn't have this kind of game.
The ground under the wall is hardware and technology. Software technology is a box that Snake can stand on to help him over. But it's still not enough - he needs a ladder to make the last few metres. That's game design.
Enough with the abstraction. We're now reminiscing about Kojima's first game - Metal Gear. The time is 1985 - when the MSX 2 was released, at the same time as the boom of the NES.
He joined the industry the following year, '86, and was given a mission - create a combat game for the MSX 2. The reason for it was the phenomenon of Rambo, and the coin-op games based on it.
Anything he's done before, he thinks is possible. Anything he's never done before, he thinks is impossible. It's only in his mind.
What was a combat game back then? 2D, with a human player character, and enemies - at least four. And bullets!
All this is rendered in animated sprites on the screen. Even in PowerPoint, Kojima's production values are through the roof.
To demonstrate the level of technology at the time, he's showing screens of an old side-scrolling shoot-em-up called Nemesis. He's explaining the concept of backgrounds and sprites for the young-uns attending.
You could only show eight sprites. He thinks we must be pretty glad we weren't born back then.
There's a little bit of footage of Nemesis - he's showing how some sprites blink in and out. "This is the actual product - it's not a bug." The ninth sprite on screen would always disappear.
With MSX 2, he explains, you could do four colours in one line if you layered the sprites. But now he's used up six sprites just to draw the player and two enemies. Any more than two bullets and they would start to disappear. So Kojima thought that creating a combat game for the MSX 2 was totally impossible.
A change in expression was necessary. The first idea he came up with was a combat game without fighting. Of course, that wouldn't sell.
So he thought about a combat game about escaping - just running around. "I thought that was totally uncool."
So how about hiding and sneaking? "Oh, this could work. This could be revolutionary, I thought."
But he needed to add another idea because it didn't seem heroic enough for the trend at the time. He needed to add another idea - infiltration.
He needed to add more tension for the players, so he added a story and a detailed world. The stealth game genre was born. All to save on sprites.
So the impossible mission (create a combat game) becomes a different mission (create a stealth game).
He's explaining the top-down design of Metal Gear - mazes, enemies with sight lines, screens that don't scroll. It was almost like a puzzle-action game.
That was Metal Gear. It was never released in the US. "You might know the NES version, but that was a crap game, because I didn't work on it."
Back to the wall metaphor. This one will run and run. The stealth game design was the ladder that got him over the wall, i.e. the mission; create a stealth game. "I think I completed the mission." Metal Gear became a hit, and they started on a sequel.
Their next mission was to create a stealth game on the next gaming platform that exceeded the first, and that was a deeper game. That was a mission they set themselves.
The wall rose, but the ground didn't - there was no new hardware. He had to create a deeper game without that advantage. The result was Metal Gear 2: Solid Snake.
To make the game deeper, he expanded the enemy's line of sight to a wider cone, and expanded the field of play to make it less like a puzzle game. You'd have to think about enemies outside of the screen.
Since there was no scrolling, he created the radar, so players could track enemies across the whole level. He thought that was quite a nice idea, if he says so himself.
Uh oh - a technical problem. Kojima's million-dollar PowerPoint is going backwards. "Please don't post this on YouTube, okay?"
We're back on track. He's saying that he also added an Evasion phase, and hearing to the enemies, bringing more rhythm to the hide and seek, and more tension for the player.
So that was Metal Gear 2, only released in Japan. There was another "crap" NES one that he didn't do.
So once again, the game design was the ladder that got him all the way over the "wall": create a deeper stealth game.
MISSION COMPLETE. Well done, Hideo.
Here's mission three: create a 3D stealth game for the MSX 2. Now that really is ambition.
He thought it would be more fun to change the camera to a first-person view. We're back to the 3D path. Snake's facing a huge wall with like two steps in it. We're losing him, frankly. This metaphor is like one of his cut-scenes.
There was "a big incident" though - in 1994, PlayStation was released.
The hardware elevated Snake! The ground went up. Ellie: "This is better than the DEATH SPIRAL!"
So Kojima changed his mission. Create a stealth game for the PlayStation. He went on to create Metal Gear Solid. There is some polite whooping.
He's explaining the game design of MGS. This PowerPoint has first-person and isometric views. It's got a bigger development budget than most iPhone games.
Kojima's talking about the long localisation process for Europe, and playing audio clips in six languages. David Hayter's unmistakable rasp is playing on the PA.
Japanese, German - "sounds like I want to eat some sausages!", Italian - "I feel like I want to eat some pasta", Spanish - "it's very passionate isn't it?"
Last but not least - the French version. "It feels romantic!" The crowd's lapping up this casual stereotyping.
Back to some game footage. Real-time cut-scenes and camera changes for the first time.
Snake has clambered over the wall, which is now a complex kind of 3D ziggurat of metaphorical challenges. We have no idea what's going on.
I added cut-scenes. "I guess these are not popular now, these famous cut-scenes."
Let's take a break. We're going to watch a Japanese TV advert for MGS. It's got some robotic Japanese chicks talking in unison. Weird.
A huge worldwide hit! Let's make a sequel - so we need a new mission.
A realistic-looking stealth game. Mission impossible, right? But PS2 came along, and the emotion engine, so everything's going to be OK.
Uh-oh - the ground under Snake didn't rise as much as we thought. Damn those Sony lies.
So we've changed the mission - create a more immersive stealth game. Voila: Metal Gear Solid 2: Sons of Liberty.
Kojima's explaining the game design again. The PowerPoint now has weather effects.
Game footage is being shown: 60 frames per second, motion capture, long cut-scenes. "I believe that some of you didn't like these." Silence. "You're supposed to laugh at this point." We do - we're polite, and he's nice.
He's talking about all the new actions, the deeper use of the environment, location damage, first-person action view - these make the ladder that got him to the immersive stealth game he was after.
There's a businessman hiding in office toilets, cupboards and drawers. It's properly funny, actually.
New mission: complete a stealth game that surpasses the previous one on a new platform. But as with MSX 2, the new hardware doesn't arrive - the ground under poor old Snake doesn't move.
How can we advance what we've done using nothing but software and game design? By changing the environment.
There's a chart showing how the series was always set in closed, artificial environments (Shadow Moses, the tanker). The opportunity is to break out into open, natural environments.
It's not just the game design ladder this time. They remade their 3D engine to cope with the natural environment, bringing Snake the mighty box of software technology to stand on. It's brought along by Otacon, for some reason.
So we get Metal Gear Solid 3: Snake Eater. Kojima's explaining the camouflage and survival elements of the game design with his king of PowerPoints. It's got a nice wood effect now.
Tom says: "Otacon is the man!"
More self-effacing gags about cut-scenes, and we're into some MGS3 gameplay footage. CQC (close quarters combat) was added because it's not always easy to fire guns in the jungle.
The concept of the game design was "infiltration in a natural environment". The box and the ladder get us over. MISSION COMPLETE
TV commercial time again. The MGS2 businessman's in the jungle, fighting snakes off with an umbrella and doing his ironing. Why don't we get game ads like this?
"Although the series was a hit, I always said that I wanted to end Metal Gear's saga as a trilogy." But the world wanted a sequel. "So I came up with a plan - why don't I make the ultimate stealth game? Then I won't have to make Metal Gear Solid any more."
It's 2005 and Kojima hears a rumour of a monster gaming hardware that's so good you don't even need use a game design. A new mission: create the ultimate stealth game using the rumoured power of the ultimate gaming machine that doesn't exist yet so I don't have to make Metal Gear Solid ever again.
The machine, of course, was the PS3. The wall is impossibly high, but the ground soars up to meet it - and then drops halfway back down again. Snake clings to the ledge but falls back down.
New mission: use the ACTUAL power of the PS3 to create the ultimate stealth game. No, wait: a new infiltration experience.
The game design is to create an infiltration game in a warzone, where you can fight both sides or neither. Massive gun battles rage on the PowerPoint.
Presto, Metal Gear Solid 4: Guns of the Patriots, released last year. 8/10
"It's a monster machine, so the cut-scenes are monstrous as well, okay? You're supposed to laugh even more, come on, I'm talking about my cut-scenes!"
He's showing some footage of the first level, with Snake deciding to fight the militia as well, and them turning against him. The concept was infiltration into a "situation" - it's not about the place any more, it's about the situation.
MISSION COMPLETE. We're rewarded with a TV ad. A young man decides to play MGS4 rather than marry his girlfriend.
He's zoomed out to show the whole journey of Metal Gear up this path. Remember the first mission? If he'd given up, there wouldn't be any Metal Gear Solid - or Splinter Cell, for that matter.
It was because of design responses to hardware restrictions that Metal Gear came to be.
Hideo Kojima's philosophy is of designer-driven game design. Anyone who's over 40 years old is like this, probably.
Kojima's riding his metaphor all the way to the moon now. Ladders, boxes, isometric paths with multiple steps. Looks like it might be quite a good game, actually.
Especially Japanese game designers are like this, he thinks. It's the traditional way of making games. But now there's a different trend - more reliant on the software technology, as was the case with Metal Gear Solid 3.
This technology-based game design is used in American and European game development studios, he realised recently. It uses open-world game design where you can run anywhere without loading, ride any vehicle, destroy any building.
He wants to use both software and design to vault up a high, high wall, now. That's his vision for the next mission: the next MGS.
Developers who think Kojima is old-school with his game design ladder and want to challenge that with software technology are welcome to come and talk to Kojima Productions and do that. They're recruiting.
Summing-up time. By overcoming yesterday's impossibilities, they become today's possibilities. Nothing is impossible.
"90 per cent of what is considered impossible is, in fact, possible. The other 10 per cent will become possible with the passage of time and technology."
"Let's join together and make the impossible possible. I want to make great games with everyone."
And amazing PowerPoints. Thanks, Snake. Thanks, Mario.
It's all over. Warm applause from the GDC audience, thumping Britney on the PA.
Thanks for staying with us and stay tuned for more from GDC.