Interview: Emergent's John Austin

Atomic: Going back to the server technology for Gamebyro you mentioned before - can you go into more detail?

JA: For all of our products, one of the keys is flexibility and modularity. We're doing the same thing on the server engine side. So our server can be used for massively multiplayer online games as well as advanced 'casual' games - 6-player and 8-player. We think that's a really important trend in the market - it's not just going to be about MMOs. We have a team that's working on that, and it'll be delivered around the first [quarter] of next year. The product that we just announced at GDC and will be shipping is our metrics product.

Atomic: So the 'metrics' product is separate from Gamebyro?

JA: Yes. The idea is there's a set of probes - one line of code - that you basically can drop into your game that are general purpose and allow you to collect any sort of data you want. Then we have a web-based dashboard that allows you to display and analyse that data. These probes can measure anything from things as low-level as frames per second to higher-level things like player behaviour.

Atomic: Was NDL originally started to create a game engine, or as a game developer?

JA: NDL was actually started in 1983 by a guy named Turner Whitted. Turner is the guy that invented ray tracing and was into computer graphics rendering. So the first products the company built were not for games but for high quality rendering. What was interesting was that it was a middleware business - we didn't make an end user product back then either. For example, AutoCAD from Autodesk. Up until three or four years ago if you used AutoCAD to design a building or landscape and then went down into a menu and wanted to click and look at the rendered picture, that was NDL software doing that rendering.

So we were in the middleware business before anyone knew what middleware was. From 1983 until about 1996, the company just worked on rendering products. Then around that time, Turner started to see chips coming out from 3dfx and NVIDIA and Turner recognised the next interesting computer graphics stuff was going to be in games and the 3D chips coming out of PCs was going to generate a whole new market. That's when the company started working on a graphics engine for PC games. That became the NetImmerse product in 1997 and move to Playstation 2 and Xbox. We then enhanced it and changed the name to Gamebryo and added tools. Now we're going from what was a graphics engine to something that's much more of a complete game engine.

Atomic: So you're designing systems like physics as well?

JA: What we're doing is either designing it ourselves or licensing technology from another company. In Gamebryo, we do have collision detection and particle systems and there's a little bit of physics in there. In fact, with the current release of Gamebryo, there is complete integration with AGEIA. If a customer wants to do full-blown AGEIA physics there's a completely integrated layer.

Our pitch to game developers is that you should license an engine rather than build it yourself. We feel that same way [about middleware]. If someone is out there with physics that's really, really great we shouldn't build it ourselves, we should partner with the company that builds it. There's also company called Allegorithmic [that specialises in procedural texturing] and what they can do is generate very beautiful textures with essentially no memory footprint - and that's really important for things like Xbox 360 Live Arcade where there's a size limit. If we partner with them and make it so Gamebryo can work with that, that's really great.

Atomic: Do you see PhysX and the physics processor as something that's going to take off?

JA: I think what will really drive it is if there's a really compelling game. [The customer doesn't] care if it's generated by a physics chip, software algorithms or magic pixy dust. What was it that drove Playstation 2 sales? Grand Theft Auto. When the console manufacturers come out with their next-generation consoles, it drives their sales if there's a really compelling game that people want. They buy the game - they don't buy the console. I think it's the same way with this. What will drive how fast these chips get adopted is whether there are great games that take advantage of that technology.