He shows off some pretty impressive shit. Growing from a microbe to a sentient spacefaring race. All the graphics generated by code instead of teams of artists. The fully dynamic animations and tiny file sizes for content sharing are what really get me exited. It's kind of amusing that procedural content is being touted as the next big thing when 10 years ago, David Braben was able to fit a complete full-scale galaxy on a single floppy for Frontier: Elite 2 and Frontier: First Encounters using similar methods.

http://www.pqhp.com/cmp/gdctv/
You can use this login instead of registering(thanks bugmenot!)
p_samty@yahoo.com
yguest

They've been talking about Spore for a while now. I'll admit that it looks quite impressive - but also quite niche.

Your more standard games may use procedural for things like detail texturemaps, fire effects, etc, but artists are going nowhere. If you get fancy you could dynamically generate 3D maps and buildings, but you'd still need to model the basic building-blocks.

To be honest, I don't think Wright started with any real vision for a game, he just started throwing technology at it to see what popped out the other end.

I think it has too much of a lego/Sims thing going for it to be all that niche unless the SF-ish theme scares people away.

This game does a appear to have building blocks like limbs, torsos, and weapons. Modeling the building blocks of a creature or environment's geometry and textures still requires effort, but a lot less than hand carving and painting each feature.

FPSes expecially could benefit from procedurally generated, dynamic maps. The only recent one with random maps I can think of is Soldier of Fortune.

>Wright started with any real vision for a game

He says as much in the video, I think.

> FPSes expecially could benefit from procedurally generated, dynamic maps.

I'm not so certain about that. Maps are an important, if oft-overlooked, part of an FPS. An otherwise great FPS with bad maps is a bad FPS, and here lies a problem: procedurally-generated random maps have no guarantees of quality or consitency. I'd hate trying to think up a fitness function that could determine the gaming quality of a map. Some 2D and 2.5D game maps have used it, but true 3D?

I suppose it might work if they used a pRNG while developing the game, looking at the maps generated by many thousands of seeds, then settling on the seeds with maps that are actually decent.

I'm on dialup, so no video for me. :/

>procedurally-generated random maps have no guarantees of quality or consitency.

There will inevitably be some unbalanced maps in any randomly generated set, but I think a mitigating factor would be that neither side would have prior knowledge of their advantages or disadvantages. It would be up to them to adapt to the strengths and weaknesses of their positions on the fly. I like the idea of random maps because in many of the multiplayer games I've played, people tend to stick to the fastest, simplest maps. I like to call it the de_dust problem, after the horde of 24/7 de_dust servers in Counterstrike. This can also be seen in games like Team Fortress or Tribes with their maps that have bases facing each other across a short gap.

>I'm on dialup, so no video for me. :/

I feel for you. :( It really is quite impressive, especially how he can change a creature's limbs and it instantly adapts its mode of locomotion, still within the creature editor!

Multiplayer is the biggest game designer cop-out ever. "Hey, we have this cool graphics engine, but we haven't the first clue of how to turn it into a game!" "Eh, just stick some guns in there and call it mutiplayer!"