TGS 2003: Developer's workshop update

TOKYO--So you have three game developers: a Japanese, an American, and a Scot.

No, not the start to an inside joke. This was the lineup for Friday night's (in Japan) IGDA-sponsored Tokyo Game Show developer's workshop. In attendance were Yu Suzuki of Sega, David Weinstein of Red Storm Entertainment, and Graeme Devine of Ensemble Studios/Microsoft (Devine is the Scot).

If there's royalty in game development, this is it. Suzuki didn't work on every Sega hit of the last 20 years, but he came pretty close, with a resume that includes Hang On, Out Run, Afterburner, Space Harrier, Shenmue 1 and 2 and the Virtua Fighter series. Devine is an IGDA Chairperson Emeritus and worked on Doom 3 at Id software before moving to Ensemble, while Weinstein has been involved in Red Storm's hit tactical shooters based on Tom Clancy's work, a select list that includes Rainbow Six, Rogue Spear, and Ghost Recon.

Devine kicked off the session with a presentation on the evolution of game development, concluding with a proposal for a new approach to development.

He wrote his first game in 1978; the development team was "one teenager in his bedroom," and the publisher was an adult that did business from his living room. Since those days, two important trends have converged: more powerful software has resulted in more sophisticated games and much larger resource requirements (time, budget, and staff) for development. At the same time, beginning about five years ago, PCs started to catch up to the workstations that developers used, which once enjoyed a 10x speed advantage. Now developers work on the same high-end PCs that average consumers can buy and as a result, they have much less ability to predict what new functions might be possible in the next generation of consumer hardware.

The trend towards longer project lifecycles and larger teams and budgets have made project management the most important development discipline in development. Devine blames burnout due to poor project management for the fact that the majority of workers in the game industry leave the industry after completing just one project, and he points out that even when a development house with poor project management succeeds in shipping a game, there's a high chance that it will fold before it can line up its next project. This failure to retain talent is a severe handicap to the long-term growth of the industry and must be addressed.

Devine's proposed solution is revolutionary: he believes that when hiring producers, rather than promoting from within, game studios should seek out people with "business degrees" and management experience. Despite Devine's assertion that "We suck at managing ourselves," this recommendation flies in the face of current industry practice, and it's hard to imagine that this message will find a receptive audience.

Devine's next proposal, though equally unorthodox, has immediate impact on the bottom line, so could be a more likely candidate for adoption. He believes game design teams should consist of a small core of perhaps four people, with all other positions filled by contractors.

This would eliminate the need to pay idle workers, which currently arises when, for example, texture artists have to wait for level designers to finish before they can begin. Moreover, having a small core team makes it easier to decide a general direction, which each of the team members can communicate to the contractors in their area of expertise. Devine sees a parallel to the current Hollywood system, in which studios hire set designers and other production personnel on an as-needed basis, and he predicts that this model or something similar will be in place by the end of the decade.

David Weinstein's very technical presentation entitled "Case Studies in Multiplayer Games," challenged not only his interpreter's skills, but also the native English speakers in the audience. At one point, he showed a code sample and said nonchalantly "oh, you'll notice there's a bug in that code. I just noticed it when I was preparing this presentation," and then moved on, leaving some members of the audience scratching their heads. The meat of his presentation was recommendations on software architecture for multiplayer games, accompanied by the caveat that "even a solid architecture cannot salvage a mediocre game."

Weinstein also noted some important differences between console and PC development. Because of the ease of releasing patches, the availability of online user communities to provide support, and the higher level of technical ability among PC users, he has found that it's usually better to release PC versions of games before console versions. He points out that "on the first day of release, users will put in more hours of play than occurred during the whole development cycle," sometimes uncovering bugs or play issues that slipped through development. It's easy to release a patch for the PC, and then fix the console version pre-release. Patching a console game, on the other hand, is a much more difficult task--and may be impossible.

After Devine and Weinstein's respective presentations, Suzuki joined them for a round table discussion. They began with some comments on the difference between Japanese and U.S. games; Weinstein felt that while U.S. games focus on realism above all else, Japanese games have more art direction, and make an effort to attain a certain "look."

Devine felt that Japanese games took more risks, a viewpoint with which Suzuki agreed, saying "we used to make games without even thinking about the market. We just tried to make games that we would enjoy." Note his use of the past tense--it seems that the shaky economy has limited Japanese developer's freedom to some extent.

When asked about a recent wave of high-quality games from U.S. and European studios, Devine and Weinstein agreed that rather than a Western game development Renaissance, this is simply indicative of the fact that until recently, most Western games were PC games and weren't ported to consoles. Now Western developers are porting their games, or doing parallel development, so Japan is seeing more Western games than before.

Suzuki commented, "when I started in the game industry, arcade machines were the fastest, then consoles, and PCs were the worst. Now I think that may be reversed." He singled out PCs high resolution and rapid evolution as particular strengths of this platform, and pointed out that even more than the keyboard, the mouse is a superior input device to anything available for a console.

Weinstein summed up his view on the PC/console debate as follows: PCs have better performance, input devices, and positioning--typically on a desk where it's easy to use the keyboard. He sees standardization as a key strength for consoles, "and of course, some games like sports sims and fighting games are just better on consoles."

The conversation moved on to game development procedures, and Weinstein put it in a nutshell by saying "Designers have one of the hardest jobs. Artists work with Maya and Photoshop, and programmers use compilers, but the only tools a designer has to work with are artists and engineers."

He feels that the key to game development is communicating a clear vision to production staff so that they can work independently without referring every decision to the designer, and Suzuki agreed, saying that developing a visual concept is the key to his game design process: even if he wanted to, it's not possible for him to make all the decisions.

The session concluded with an interesting discussion on what the panelists looked for when making hiring decisions. Interestingly, their responses were almost diametrically opposed: while Suzuki said that their first criteria was growth potential and life experience, Devine said that a degree or other formal training in programming or design was an important part of hiring decisions for him. This dichotomy may help explain some of the differences the panelists discussed in game design as well--Suzuki's criteria do seem biased towards free spirits, a key part of the genre-busting games for which he is known.