American companies and institutions tend to create projects based on either immediately practical applications or open-ended research. In Japan I encountered something else: comparatively long term application-oriented research projects.
Consider this project at the National Institute of Information and Communications Technology (oddly acronymed NICT). They want to hook up the real world and the cyber world by projecting live video of street activity into Croquet, adding computer-generated characters, and displaying these on large interactive touch-screen monitors in different street locations and on PDAs. Ostensibly, people walk up to a screen and ask for directions or other information. They are answered by interactive cartoon guides who might preview the route for the visitors in a cyberspace mixed with reality, and then point the visitors in the right direction at real street level. Maybe the guides jump into the visitor’s PDA if they have one. Then the guides meet the visitors at the big screen at the next waypoint.
This is not necessarily an important problem to solve, and there’s no well-understood big payoff. No VC would fund this. However, it is a specific situation that grounds the research. Open-ended design problems get resolved more easily when there’s a specific child-understandable task involved, even if it’s a silly one. I think the task is chosen for having interesting problems that we don’t yet know how to solve. The ideas is that it will produce fruitful areas of research.
I think this idea of five year research applications is a good thing to include in the mix between laser-focused problem-oriented practical applications and completely unfocused pure-research.