I'd like to spend at least 30 years on a killer game. By that time, everything can be raytraced and made from millions or even billions of polygons. At least for me, hacks like texture mapping or even bump mapping will be less of a priority for me. Basically, I'd like to know a basic universal 3D format that will still be around in 30 years time so I can begin designing the polygon structure of the game. Preferably, it will allow scope for materials too (but I want super low-level definitions, not just presets like 'glass' or 'wood' etc.) Finally, the ideal 3D format would be programmable through C. In other words I would want to program any point of any polygon to move in any way. I'm guessing something like OpenGL would support this kind of thing. What 3D format would be best? And what (preferably cheap) 3D software can get me started on this 30 year-long quest? Many thanks all for any answers to these questions!

Not a joke - I just have big ambitions :)

Quote:terms of design, 3d formats, platforms, genres, 3d engines or whatever, and they'd be lucky if they got it 5% right.

That's why I want to work on as low a level as possible. You surely can't go wrong with designing a super-detailed polygon mesh for a game, even if it's 30 years away.

To take 2D graphics as an example, one can (almost) be sure that pixels will still be around in 30 years time. Pictures will still probably be defined by a 2D array of pixels. Even if they aren't defined that way, the quality will still be exceptional given 30 years of work.

Let's say I do work on billions of polygons, and it turns out the future is a bunch of purely mathematical surfaces, one could still convert it I'm guessing.

Also bear in mind that there are graphics today which look amazing, but only run at 0.01 frames per second (i.e. unplayable). I hope to design to that level of detail, and when the future comes (say 30 years from now), it will run at a super smooth 60fps or more.

Quote:Besides, even if you ignore the fact that whatever you come up with won't work in 30 years, whatever you came up with would be crap in 30 years.

I'll be aiming more for highly detailed and tasty, but at least partially abstract graphics. I'm not going for super-uber realism, at least not at first. Even something like the Tron film had graphics which in many ways would look 'polished' in 30 years, even if they're not as sophisticated.

So again, which current low-level universal 3D format would most fit what I'm looking for?

[Edited by - TwinbeeUK on May 2, 2009 10:41:12 AM]

No, really. You can't say anything about it. For example look at the games 30 years ago: here. These are all 2D, with the pixels visable and some just lines. Nobody could imagine a game like Crysis (with milions of triangles) to emerge within a few years. Nobody could even imagine that we would have triangles in 3D space in our games in just 30 years.

The same is true for the future. We can't "see" what lies in the future of game development. Thirty years is just too much. The splines that sybixsus proposed could be just as wrong as another representation.

Edit: reply on your edit.
2D graphics were then defined as lines/shapes and for the exeptional games as pixel arrays. It's still the same but the procedural textures are gaining field.

Emiel1

My point is that even if that is true, one can convert it whatever format exists at that point. There are many awesome artistic/highly detailed 3D pictures around today, and the future will animate those so that they run in realtime. We can certainly expect something along those lines.

With enough polygons, and maybe some texture mapping, results will eventually reach a point where the eye can't tell any difference anyway.

Forget C, forget DirectX, forget OpenGL, forget raytracing, forget polygons. None of those are gonna matter (probably won't even be around). Here's my best uneducated guess about the lowest level file format for computer games in 30 years. If we don't just go down the road of cyberbrains, instead. Then I have no clue. Maybe neurons?

Quote:Original post by TwinbeeUK
With enough polygons, and maybe some texture mapping, results will eventually reach a point where the eye can't tell any difference anyway.

It will probably be like that but go into the direction of augmented reality, but that's not 30 years out, more like five years.

Quote:Original post by sybixsus
It used to be said that todays FMV would be realtime in five years, and it's probably less now.

I already can't tell the difference between some of the in-game cutscenes and FMVs in PS3 games like DMC4 and Uncharted. The only way I can tell in Uncharted is if I put on a custom costume and it disappears during the cutscene.

Good point. Though I thought even if polygons are 'empty inside' as the representation inside the computer, they are still *treated* as if they were solid when doing the final raytrace.

lightbringer, the qubit looks complicated, but hey maybe. I like the idea of using a bunch of voxels, or 'atoms' for a game. Make every part truly interactive.

Naturally, one can convert between polygons, atoms, voxels, qubits etc., given a high enough res to start with.