|
Photograph: Andrew Pothecary
|
One area that J@pan Inc hasn't touched on much so far is game consoles. But for a lot of people, that's the first topic that pops to mind when they think about Japan. We thought it was time to chat with someone in the know. Arka Roy, a Canadian born in India, has been based in Japan and working on games since 1991, when he joined developer 9003inc. There he worked on a team of four developers on AquaZone, an aquarium simulator (with tropical fish) that's still selling strongly nearly a decade later, including in konbini through DigiCube. Later he was the chief engineer for Pinna, a bird simulator with AI-based learning mixed in. After 9003 Roy became a VP at Vivarium, where he worked on Seamen, a bizarre but compelling game for the Sega Dreamcast where 5,000-year-old fish creatures have intelligent conversations with the player. After working in San Francisco for a few months on the US version of Seamen, Roy returned "home" to Tokyo and started helping a small game developer called inis -- "infinite noise of the inner soul" -- to create titles for the PS2. Editor in Chief Steve Mollman spoke with Roy at inis's small offices outside Tokyo.
inis is writing games for the PlayStation2. Assuming the games catch on, how will the revenue sharing work?
Well, let me start from the ABCs. There's first party, second party, and third party. First party would be if Sony is making the software. They're developing it, publishing it, and selling it. Second party is where there's an outside developer but Sony is still the publisher. Third party is when there's the publisher and the developer -- and they could be the same company or two different ones -- but neither is Sony.
So with our upcoming game, Guitaroo, it's a third-party situation. We're not the publisher. Koei is. Koei is a publisher that puts out a lot of historical games. They're a very big PC gamemaker, but they also make games for PlayStation. They haven't made much in terms of music games, but they were interested in doing one, so now there's a deal between inis and Koei.
The perspective to take is that the publisher is releasing it, and the question is how much do they pay the developer. The intellectual copyright is going to belong to inis, but the royalty structure is such that Koei gets the revenues and they pay inis a royalty. In this case, it's actually quite small. It depends on the number of players, though. If we were a really big, powerful developer with a huge track record, then we would have a lot more strength in negotiating with Koei and we could probably get a very high percentage. But this company doesn't have a track record of making games so far.
So inis has been in existence three years and has yet to develop its own product?
Right. Another thing that gives Koei the upper hand is that to develop a game, until you've developed it and released it, there's no revenue, yet the development takes massive amounts of resources. You need programmers, you need graphics people, you need musicians, you need to hire studios. So obviously we couldn't fund something like that on our own. Koei is giving us an advance up front, and that's usually how it works. This is a pretty typical situation for a small game developer.
Seamen [on which Roy worked before joining inis] was a second-party title in the sense that Vivarium was the developer and Sega itself was the publisher. Sega gave us the funding, we developed it, and for the first couple hundred thousand copies we didn't get any royalties, because the advance was applied against that. And then once it continued to sell beyond that, royalties started to come in.
How is this company funded? Is it all from the advance?
For Guitaroo, it's the advance from Koei. For the rest of the company costs, we've been doing a lot of outside jobs on an outsource basis. We did some work-for-hire for the Beatmania game, for example. We do a lot of stuff like that. So that's been bringing in some funds, but it's pretty much hand to mouth.
How often do VCs invest in game developers?
I haven't heard of too many situations like that. One interesting example -- and this was pretty long ago, which makes it even more interesting -- was 9003. We actually got some VC funding from CSK Venture Capital. But eventually the company wasn't doing well and the company that bought them out also bought out the equity from the venture capital. But CSK did well.
Right now you're developing a few games for the PlayStation2. How does that compare to other platforms?
PS2 doesn't come with much. Dreamcast and other platforms come with a lot of APIs [application programming interfaces] and give you a high-level mechanism for controlling the graphics and sounds. PS2 has a very low level layer, so you're programming to the metal in a way.
Why is it set up that way?
I think there are a couple of reasons. One is that it's new. In a lot of these cases with game console machines, APIs are created by third-party developers, so since the PS2 is still new there aren't tons of third parties that have made APIs yet. With the original PlayStation, anybody and everybody was making games for it. There are thousands of PlayStation games, which was one of the reasons they were so successful, but the flip-side of it was there were a lot of really bad games. So by making it a very powerful platform but not being all that supportive, maybe what they're doing is ensuring that only really good developers work on it. Which I don't really understand. I wonder if that's the way to go when you're trying to promote your machine.
Especially with Microsoft's XBox coming along. You know Microsoft is going to follow its usual strategy of having the most apps written for its platform.
Exactly. The games-to-consoles ratio for the PS2 is something like 1.6 to 1, whereas a healthy ratio is 4 to 1. So your average person who's bought a PS2 has bought one or two games, which is very bad.
I see a theme in your work: simulation. What's your fascination with it?
I like the idea of life. Computers can be pretty uninteresting at times, but life is never uninteresting. So if there's a way to imbue some of that interestingness of life into computers, that'd really go a long way.
Seamen uses speech recognition, which still sounds sci-fi-ish.
I think it's one of those things that's quietly seeping into our lives. For example, many DoCoMo cellphones have a speech recognition function: you have an address book and you say the person's name and it'll do an autosearch and dial the person's name.
The simulation and the pet raising are all well and good, but the speech aspect is enormous, actually. Seaman has sold about half a million copies. I think right now it's the biggest and most extensive use of speech recognition in any game. It's about 20 days worth of conversation with this creature. As you continue talking to him everyday you kind of figure out the mystery of who he is and where he's from. His first question to you is are you a man or a woman.
If you're a man, he says something like, "Oh, that's a bit worrisome. I hope you actually take care of me." [Laughs.] Meanwhile, this huge database profiling the user is being built up in the background. So as the conversations progress, he reaches into the database and uses that information.
And you went to San Francisco to localize this game for the US market.
We had to hire a team of programmers who weren't necessarily high end but could get the jokes. In Japan, gaijin programmers are usually really busy working at banks or whatever. They're usually pretty expensive.
Have you thought about outsourcing to India, where you were born?
It's more difficult with games. I guess if you're doing really straight IT and financial applications, and you've got very clear specs, then you can send the whole shebang overseas. But here the teams are very small; it's very tight. Because as you're doing it you kind of make judgments, and these are just day-to-day things you're just sort of embroiled in all the time. It's pretty hard to send out a portion of it.
|
