The Ivory Tower

March 2004

Ebony and Ivory

Experience from joint projects between game developers and researchers

Working at a games studio active in the field of pervasive gaming and location-based gaming, I've gained some experience working in projects involving games studios and academic researchers. I believe such experience is still rare within the game developer community: game developers and researchers working together, side by side. With that experience, I'll try to shed some light on the pro's and con's and success factors of joint projects between game developers and researchers.

My company's first contact with game researchers came late. This was a few years after the company was founded and well after it had entered its current niche within games, when we met a research organization that was researching the pervasive gaming field. We were quite surprised that there were game researchers at all; let alone researchers in our field of gaming. A few months later we were already involved in our first joint project, and since then the company has been involved in half a dozen research projects with different research organizations. In this column, I will use an early project as my main example: the Vranderer project (a collaboration between It's Alive and SICS). This project was about developing a PC client in 3D for our pervasive game Supafly.

The Pros

At first, the benefits of working with researchers seemed limitless - like an outsourced R&D department working for you! And in an industry constantly under cost pressure, R&D time is indeed a luxury. Even if the reality is not quite that simple, having researchers involved in your game development can do some amazing things for your end product.

Having researchers involved in the game development process will bring a fresh perspective to your game design discussions. Instead of just throwing references to other computer games in the discussion, researchers will find comparisons to totally different projects. In the Vranderer project, our own experience from 3D games failed to help us in designing a 3D client for a "real-world" game, especially in the area of content creation. The researchers could draw on experience from other projects involving VR and geospatial data, and that way we found the solution.

In some cases, cooperating with researchers can give you access to external funding - either as government grants or as funding from other organizations endorsing collaboration between companies and academia. This may sound greedy, but it allows extra R&D work that else wise could not been motivated by commercial reasons alone.

The Cons

It didn't take long before our first joint research projects started to get out of hand. Deadlines were missed, the quality of the work was not as good as anticipated, and the collaboration generated friction between the game developers and the researchers.

The core of the problem was the existence of two different sets of goals. The game studio basically wanted something with a commercial potential. We were happy to have something that was "good enough" in terms of features, but had the right quality and was realistic compared to deadlines. The quality and deadlines were paramount.

The researchers were looking to have something that would make an impact on the research community, and had higher ambitions with the project's general scope. Wider scope meant less attention to quality assurance and deadlines. From the researchers' point of view, if the solution worked well enough to prove the overall research point, it was considered "done". In the Vranderer project, as well as in several other research projects, we have found it difficult to use our commercial game platform, which is important to us from a quality and integration standpoint. Most often, we have succumbed to using project-specific prototype software - a quicker route to the solution.

For us at the games studio, it proved difficult to plan resources too far ahead in relation to the research projects. This boiled down to a matter of prioritizing. When a well-paid development project landed on the table, it was tempting to free up development resources currently assigned to a joint research project. This would of course disturb the time plan of the research project, and upset the other project members.

Success Factors

Since our first attempts, the results of our joint research projects have improved. However, in order to make the projects work, you must align your goals within the group. Together you must thoroughly define purpose, planned results, and deliverables that are useful to all parties involved. In our experience, starting a project with a joint workshop where these goals are defined speeds up the rest of the project immensely.

For the game developer, it's important to early on define whether the project is considered to have a direct impact on product development or have a longer R&D perspective. If the outcome is a commercial product, make sure to plan developer resources to allow plenty of quality assurance before the deadline. When the purpose is R&D, testing may not be as important but you must have a plan on how to integrate the results with another product or process. Otherwise, the game developer may end up sitting with a research report or prototype, but not having the knowledge or resources to assimilate it.

--
Tom Söderlund
Tom Söderlund is a game producer and co-founder of It's Alive, a Stockholm-based games studio focusing on pervasive gaming and mobile massively multiplayer online games. It's Alive started in 2000 and made its debut with the mobile location-based action game BotFighters.