4-30-04Back in like '99 working on Genesis, I used to talk to Jason Wood and John Pollard, and we'd talk about increasing visual realism, and the things we really wanted were - floating point buffers, radiosity, fresnel & better surface lighting, and ambient occlusions. Now all this stuff is coming to be! It's sort of a let down in a way, back then there were very few clever people in game graphics who could actually think of things like that and implement them - now you have big teams of demo guys at the IHV's and researchers in academia, all paying attention to games now, and writing tons of papers and demos, and tons of game devs are implementing all kinds of crazy things, you can't really sit in a closet and come up with clever things and blow people away any more. Everybody basically knows what can be done and what would look good, and now it's all just a matter of actually doing it. Dumb issues like the art pipeline and getting the designers to make a game that takes good advantage of the tech are what dominate the results now.
Making technology-driven games is fun, and it's really not bad for gameplay. Basically game designers have very vague ideas anyway, so if you give them some heavy tech direction and constraints, it really doesn't make the game any worse, as long as they can still experiment and tweak and have good-workflow and fast turnaround, etc. Letting the programmers choose the primary structure of the game is clearly the way to go. Then the designers and artists fill that structure with fun and decorations. Not only does that make it easier to write the code, it makes the design & art much better, because you get a system which is doing something very well, doing what it was made to do, rather than just a random mix of ideas built on a system that was made to do something else.