The Edge Archives: Gabe Newell on Valve in 2007
Every weekend, we're delving into the Edge archives, mining our rich 19-year history for features and reviews from years past. This week, it's an extensive interview with Gabe Newell from E172 in 2007, in which the Valve co-founder shared his vision for the future of his company, the Steam platform, Windows, and PC gaming in general.
Gabe Newell is the managing director of something we have no name for. The company he co-founded ten years ago, Valve, to make the game that would become Half-Life, is now part-developer, part-distributor, part-tools provider, part-publisher. Having embraced mod-making, eSports, online distribution and episodic content, Valve has consistently shown itself to have a progressive and inclusive vision of what PC gaming can be. So where does the man with no publisher to please, no retailer to serve and no licensor to obey think PC gaming is going?
Valve has become an increasingly hybrid company in the last ten years. Do you think that nowadays it’s a bit outmoded to still be thinking in terms of being solely a publisher, or a developer, or a distributor?
I just don’t think it’s helpful when you’re figuring out which decisions to make. I think you need to think: ‘How are we either helping our customers or how are we helping other developers?’ There’s been this notion, really pioneered by John Carmack, that your game is actually a starting point for other people to do stuff with, and that meant thinking about other things like graphics and interfaces, entities and threedimensional worlds, but nowadays it means thinking about things like payment systems and anti-piracy.
The Steam download service has evolved very rapidly over the last few years – has that been part of that same thinking?
I think with whatever we do there’s always what we’re going to do with it and what we think other people might want to do with it, whether it’s our development tools or our game code or our engine or friends or whatever. There’s what we can do with it, which is to create Half-Life, and then there’s what other people can do with it, which is to create Team Fortress and Counter-Strike. When we were thinking what we would like for Steam, we always assumed that in turn we’d have to make it available for other people, who would also figure out how to take advantage of it. Some people, like the guys who did Ragdoll Kung Fu, were unable to get a traditional retailing relationship, so it was super-useful to them. And other people say: ‘You know, this solves update problems for us because we don’t have an adequate mechanism for getting our patches out to all of our customers.’ So whatever we do we’re always going to look at it in terms of internal purposes as well as providing a building block for other people to build on.
Speaking of Ragdoll Kung Fu, what’s the selection process now for a game being distributed via Steam?
We don’t really think of ourselves as having that traditional publisher ‘here’s our certification’ process. We view ourselves more as creating tools for end-users and other software developers. So if software developers say: ‘Hey, that’s useful,’ then we’re perfectly happy to let them build on it. I think we tend to view ourselves as really in the tools business than in the gatekeeper…
So it’s predominantly a case of people coming to you rather than vice versa?
Well, right now I’m talking with pretty much everybody in the industry, so it’s not so much a case of people coming to us. We have pretty good relations with just about all of the developers and publishers that we’re aware of, and certainly if there’s somebody new coming along that we don’t have a relationship with, we’re always interested in talking to people. We’ve been pretty good at reaching out to new developers as they come along.
Do you think there’s a snobbishness in PC gaming that looks down on lower-spec games that aren’t cutting edge, but still aim to be a high-quality product?
I think that would be a huge mistake. I mean, look at Popcap’s ability to sell five million copies of Bejeweled, which, you know, by no stretch of the imagination is pushing the boundaries of DirectX 10. Certainly some of the most successful titles in PC history, such as Sid Meier’s Civilization line or the Sims line have not been driven out of their visual quality, so I think that it’s a big mistake to view graphics as being the end-all, be-all. Certainly, when I talk with other developers, they understand the
importance of the performance enhancements we’ve been seeing on the graphics side, but I think everybody understands the fundamental importance of good gameplay. I think the thing that’s interesting is that the same underlying hardware dynamics that have been allowing graphics to improve so rapidly, all the way back to the Voodoo FX 3D card, are now going to start being applied to the rest of the game engine as well.
The emergence of multi-core, which is essentially the CPU manufacturers adopting the same approach that the graphics card manufacturers have been using to increase performance, means that we’re now going to see that same huge increase in performance for everything other than graphics. And I think that we’ll see some more balance returning to where people can be excited about PC gaming. We’re going to see an order of magnitude, two orders of magnitude improvement in AI performance over the next couple of years, or physics, or animation, or whatever in the same way we’ve seen those performance improvements just in particle systems or polygon and shader rendering. We’ve got a little out of balance, and the fact that the CPU guys are going to go with multicore strategy means that we’re going to see all the other parts of the game catch up.
There’s often developer frustration about putting effort into the latter stages of a game that many players won’t end up seeing. Is there a similar frustration about designing highend visuals which most players can’t display?
We kinda like our customers and welet them tell us what they want. One of the reasons we’re experimenting with episodic is that it lets us track the hardware changes much more closely as they occur. For example, one of the things we’re seeing right now is that our customers are really starting to adopt multicore processors much more quickly, and that’s causing us to say we have to invest in that, and they’re holding on to their DirectX 7 graphics processors so we need to support that. I think everybody here is much more interested in making our customers happy than in pining for some idealised world of infinite graphics performance. That sort of permeates everybody here. If people want to run our games on a Nintendo DS, well that’s great, we’re going to run them on a Nintendo DS, because we like customers more than we like vanity graphics showcases.
It’s been a little over six months since the release of Half-Life 2: Episode One – how good has the uptake been among players?
So far it’s been great. We want to go through this three times, we want to make sure that people aren’t just excited by the novelty, but certainly, Episode One’s been really successful for us, and we’ll just sit down with our customers after we’ve done it three times. It’s like if you did anything once in this industry, you’d make really bad decisions both positively and negatively. We’ll certainly crank on that a few times and then sit back with the fans and talk about what the feedback is and go from there.
The episodic model means that you’re producing smaller chunks of narrative than before. Is this something that’s fun and works better now, or are the team feeling the constraints when they’re used to having the full length of a game in which to explore their ideas?
No, we’re having more fun, it’s definitely more enjoyable. At this point in Half-Life 2 we were still years away from even revealing it publicly, whereas with the episodic, we’ve already shipped Episode One. We’ve received about 7,000 emails from people that myself and people in the team all read, so it gives us a lot more confidence going into Episode Two: ‘These decisions worked, these decisions didn’t… we need to do less of this repetitive thing and more of this character development thing.’ It’s a lot more enjoyable to feel like you’re more engaged in that feedback process with the fans than still being years away from even announcing it.
So if the three episodes are successful, then this would be a model you’d like to switch to as your basic approach?
Yep. So far.