September 04, 2005
Gabe Newell Interview Transcript
Editor: Let me start with a really hard hitting question, it’s one I gotta ask.
[…interlude …]
…has anyone actually talked to you? Because I mean your stuff is very cinematic.
Gabe: Yeah, we’ve met with a bunch of people. I think… I sort of have two minds about it.
One is sort of as somebody who’s running... who’s supposed to be running a business -- which is that a lot of these film projects end up not being very additive. It’s hard to see how the Mortal Kombat franchise has really been moved forward by the Mortal Kombat movies.
And then as a gamer, I think most of these movies are terrible and weren’t worth making, and I would hate to see any of my favorite games sort of mangled like that. [Read More]
We’ve tried a couple of different times to see if some sort of movie project would be possible. So we’ve had people in Hollywood try to do scripts, and we’ve actually taken a stab at writing the script, and at the end of the day, none of the movies were worth the rights.
Maybe someday, but it’s not going to be made until -- I mean -- our goal would be that the movie is as good a movie as the games are games.
[… interlude: Half Life 2 xbox…]
The episodic content will be better than Half Life 2 was. They may be shorter but the content itself is going to continue to move forward and hopefully move forward faster than we’ve been able to do with these large scale releases.
People really enjoy the closely coupled interactions. A lot of people, their favorite moment of the game was playing ball with dog, and this idea that you have this other creature in the world who’s interacting with you in this fine grained way, where you’re doing something and they’re reacting to it right there.
That’s something we say we need to do a lot more of -- we need to make the player feel like they can rely on, or are dependent upon, other people in the game -- and not at some high level story telling way -- but in an actual “I need you to do something for me right now”.
So there’s a lot of stuff where you can’t do anything, you’re relying on Alyx to do it for you and what you can do is get her to do stuff for you, so you’re feeling like the two of you are working closely together.
Editor: That’s really interesting because, at least in terms of console games, that’s almost the opposite of the paradigm where you have a helpless, sort of female consort that you’re guiding through things. Resident Evil 4 was one of the more recent things to add that. Price of Persia did…
[an edit?]
Gabe: You don’t want to have the sense that there’s this box around the NPC and you see these boxes bumping into each other. You want a sense that they’re in the world interacting with things closely, like they can reach out to stuff, they can push things to the ground, they can kick things, stuff like that, and have it not be that sort of fakey box-box interaction.
[…interlude: Future Tech… Image Based Rendering…]
Gabe: There’s this technology that was really exciting that I’d like to see us get into production, which is a different approach to rendering complexity: Moving things into and out of an image domain and then seamlessly interpolating between those motions as you move around. So that everything close to you is physical and geometry, and everything really far away from you is an image, but you have no way of telling that if you do it properly and things can fly out and come back -- so as far as you’re concerned it all feels like the world to you, but as far as your rendering is concerned, you’re keeping your polygon budgets and your shader budgets and your fillrates under control.
We’re focusing on things that have more obvious gameplay significance like NPCs and vehicles and stuff like that earlier, because it’s always good to do gameplay stuff, but I would like us to come back to that -- I’d love to be able to put people down into cities where they can navigate the entire city and have them feel like “my god there’s a huge amount of stuff going on around me”. That anything I can see I can walk over and in the background that stuff is going from being essentially a skybox to a sort of a low res version of geometry to a high res version of geometry as I get closer to it.
[…interlude: Lost Coast…]
Gabe: What do you think of Lost Coast?
Editor: Looks great, I mean you can tell the difference in two seconds and actually I was just playing on DoD Source, between a non-HDR machine and an HDR machine and it looked really good, you can tell the difference in like a second.
Gabe: Good.
Editor: And the director’s mode… like documentary notes, it’s a great idea.
Gabe: Great, great.
Editor: Incorporate that into wargames.
Gabe: Hopefully that’ll be a standard feature of all of our stuff now. That’s the thing you know, we like to be able to get stuff out to people, like HDR. We can solve all the technical problems get it out and make sure it works with everybody’s display adapters and then DoD takes advantage of it right away.
So rather than waiting until Half-Life 3 ships for our multiplayer games to get a feature like HDR, we can do it in Lost Coast, and then DoD Source picks it up right away. So HDR now and maybe this image based rendering stuff later. It does make it possible for us to do that.
[…interlude: Changing the Game…Episodic Content via Steam…]
Gabe: The thing that we don’t understand -- and that’s going to be really interesting -- it’s like the entire industry has been making feature films, and the door is starting to open for TV shows.
TV shows are totally different than movies in terms of the people who are good at it, and how you structure them. In movies it’s the directors that are preeminent, in TV shows, it’s the head writer, the show runner who’s preeminent.
I think we’ll all be learning a lot. It’s like, "do we need to release 22 times a year?", you know, does it need to be a TV show. Is once a week the important frequency of releasing or is once a quarter ok?
[…interlude: Day of Defeat Source…]
Gabe: There were two sort of big painful periods for Steam. One was when it went from being optional to being the way that everybody needed to get the updates and that was pretty painful for people. And then the Half-Life 2 launch, where we got swamped with not having enough capacity. So those are both painful memories for us.
At least right now, we’re not setting anybody on fire. We seem to be getting updates out on a really regular basis and we seem to have that process really smooth right now.
Right now we’re sort of re-architecting Steam, so people are actually running two versions of Steam right now, they may not realize that but they’re running Steam 2 and Steam 3 along side of each other. The nice thing about having a system like this is there’s no reason not to have your old system and your new system co-exist while you’re migrating functionality from one to the other.
One of the problems that we had was where there were architectural problems that made Friends unreliable, and so the new version of Steam, its approach to connections and how it manages connections should make friends a lot more reliable.
So, just from a technical perspective it’s evolving from a “not annoying the hell out of our customer” perspective, I think we’re out of that hole. You know, it certainly turned out to be a great way to sell products, I mean we were worried that people wouldn’t want to purchase products that way, and now that’s not really a concern of ours.
Gabe: I think what we have to do is continue to look for ways to make Steam more valuable to people, to solve problems for people. There are huge numbers of support problems that we can just make go away by proactively solving them.
We still aren't getting display drivers out to people automatically, which makes me crazy. It's like one of the biggest problems our customers have is the fact that they don't have an automatic update facility for display drivers, and it's been obvious for a couple of years that needs to happen, and yet we still don't have that...
Editor: Isn’t Microsoft trying to address that with Longhorn?
Gabe: Great!
Editor: I mean, have them been talking to you about this? I figure they’d probably be talking to game developers….
[…interlude: Steam is the Future... the next-gen headache…]
I’d be more likely to be excited about NVIDIA or ATI saying they were going to have solutions than expecting that Longhorn is going to wave a fairy wand over anything. They keep cutting and cutting and cutting, and I think expectations for sales of Longhorn are getting less and less…
But that's an example of something that would be clearly be valuable to customers, would solve problems for our customers, and would globally reduce everybody's support burden and yet we haven't done it yet.
…Like I had a conversation with people at Microsoft recently, and I said "I cannot point to a single feature in Longhorn that I care about. There is nothing in Longhorn at all that solves any problems for us at all.”
You know, I had the same conversation with the xbox 360 guys. It’s like “The xbox 360 doesn’t make my life any better, and in fact it makes it a lot worse and you’re telling me I can’t rely on having a hard drive.”
When I look at what I need to compete with, the most interesting game property right now is World of Warcraft. Huge retail sales and huge recurring revenue. Not only that, but they have a great experience wrapped around it, whether it’s their forums, or community art, or whatever, they’re not only getting their customers to play the game, they’re getting their customers to make the experience more valuable for other people who play the game. So I’ll go up and download music and watch the movies that people have created, and see fan art and do all these other things.
So when I look at, sort of, what a platform needs to help a software developer do right now, it’s figure out how to beat World of Warcraft, and when I look at the strategies that are being put forward by Microsoft on the system side or the xbox side, or Sony, or Nintendo, they’re not making my life easier.
Like the Playstation 3 makes my life as a software developer much harder. All of a sudden I’m supposed to figure out how to have this asymmetric multithreaded game, right? And I’ve never written a single line of multithreaded code, ever, right? It’s not like I was lying around saying “I need to re-architect every line of code I’ve ever written in order to get it to work.”
So one of my junior programmers, who’s writing game code (rather than system code), could slow things down by, in a real world case, by a factor of 80, because they’re doing something out in the AI, or in the game DLL, which used to be totally safe, and now all of a sudden the whole system just slows down. And the one of the really experienced programmers have to go in and say, “Oh, you can’t tell but you’re doing this, you ran out of register space, and this other thing happened, and no there’s no debugger that shows this to you.”
Writing for SPEs and writing in a Playstation 3 environment, there are incredibly few programmers who can safely write code in that environment. You make tiny little changes to code running on one of the SPEs and the entire thing will grind to a halt. You have no visibility into why that’s happening -- it’s just sort of magically running really, really slow. It’s also incredibly hard to architect things at the beginning so that you can distribute all of your functionality on all of these different processing units.
This was not a problem that we were lying awake late at night saying “oh we would really like to take this on right now.” You know, we were worried about little things like billing, and forums, and wikis and things like that.
I totally see why Sony wants people to write code that runs on 7 SPEs and a central processing unit -- because that code is never going to run well anywhere else.
They’re saying “make your code not run on anything other than one of ours and we’re betting that we’ll have market share that’s so high that everybody will have to write code for our platform, and other people, you know, “we’ll just starve the air from other platforms by absorbing everyone’s R&D budget and making their code less portable.”
[...edit?...]
Steam was essentially “here is this set of tools that software developers need, focused on solving the problems we have with these next generation of games”. You know billing, updates, product support, connecting our customers together, and things like that.
I would think that for a lot of developers, things like Steam are going to be more interesting than and solve more problems to them than this next generation of hardware and operating systems platforms.
You know the Saturn came out, and that was intended to take this previous generation of games, and create this super complicated chunk of hardware that would help you make the ultimate sprite oriented game. And Sony came along and said “no, no, no it’s not a sprite problem any more, it’s a 3D graphics problem”…
I think a lot of developers are going to say “that’s not the problem, it’s another Sega Saturn". It’s "how do we connect to our customers?", "get data from our customers, get updates to our customers?", "have closer relationships?", "how do we compete with the customer experience that you get out of being a WOW customer?” more than “how do we blast another set of pixels at what is essentially a 640x480 screen?”. [Discussion] - Source: [1UP Videos]
