Inane and off-opengl-topic thread, maybe, but I just got an overwhelming feeling of how lucky I was to be able to program a computer. We (as programmers) have opportunities that mathematicians and engineers stretching back centuaries would have killed for. I’ve been playing with ODE (Open Dynamics Engine) recently, and I’m constantly in awe of the complex behaviour simple rules can produce…
I’ve thought this too…we are in a sort of golden age of computer graphics.
The real-time visuals we can get now are pretty damn good and the progress is still really fast, which makes it a very exciting time to be in this field.
I dont think we’ll even need graphics coders anymore in 100 years. Few more years down the line, and we’re gonna approach ray-traced quality one way or another. After that everything will just be done for you.
I think now, is the best time to be a graphics coder.
>>I dont think we’ll even need graphics coders anymore in 100 years<<
true, there is gonna be a limit reached ( i think about in 20 years time) where real time graphics are gonna be pretty close to ‘visualy perfect’ (then what are we gonna do? improve on perfection
‘visualy perfect’ defined as where improving the quality/performance doesnt have any human visual imporovements
eg take a 17inch monitor, have a 10000x10000 pixels on it does 1000000x1000000 look better to the viewer, it maybe measurably better but for joe blow both images look alike.
what are 3d programmeurs of the future gonna do then?
become artists
It’s certainly great having what we have at home now, at home.
At the same time, you have to think of the possibilities. How about some real 3D imagery. 3D holograms that you can perhaps touch, interact with. Androids that are very human like. And how about computers that don’t crash and that can fix their own problems or give us a clue
Watch Star Trek, you’ll see what I mean (Next Generation, or Voyager)
Also,
It’s pretty hard tracking bugs. Making mistakes is easy. I sometimes compile to find typing mistakes! VB is nice for that case.
i dont really care about what could be done etc, i just like what i have now. if we would not have pc’s i would like what we would have then etc…
and i dont think we get that far that we have androids and holograms, because our civilisation will die out before because of alergies… like the romans who all got white hairs and started to die out because of the tin/copper stuff they had in all the water in. look out and you see that today in the rich states there are more alergicans than normal peoples, and the rate people die because of it is growing. more and more people can’t move out into some grassfield with short trousers anymore without getting fully red else.
and ther are always our sweet terrorists and wars…
in the end i’m happy to be in a world where i can do what i want more or less, but human contacts are much more satisfieing than any virtual world. why i love pc’s is just because i love math…
Originally posted by Nutty: Few more years down the line, and we’re gonna approach ray-traced quality one way or another.
Ray tracing handles only one kind of light path: DS*E (D=diffuse surface, S=specular surf., E=eye). You have to use some kind of tricky integration method (montecarlo, etc.) to poorly simulate global illumination. Furthermore, raytracing is for rendering surface, not volume -in reallife there is not “matter surfaces” and “vacuum”, but a whole large matter density distribution along 3 axes. No rendering method can handle that nowaday
I think graphics rendering algorithms have not evolved so much in 30 years. We got raytracing, we got scanline, so what ? Is that all ? We’re still bound by computing power and data storage formats, IMHO.
Come on, what have lookups (a texture is nothing but a lookup) in common with reallife lighting ? I look forward for global procedural matter density algorithms, and who knows, maybe I’ll still alive when hardware can handle the thing
in reallife there is not “matter surfaces” and “vacuum”, but a whole large matter density distribution along 3 axes. No rendering method can handle that nowaday
That may well be true, but the better we get, the smaller the difference becomes of what we can render, and what real-life looks like. At some point you will get to a stage, that looks easily better than TV-reality. But to get as good as real-life, would require exponential amounts of computing power, for very subtle and almost negligable effects.
I do think once we reach simple global illumination systems, there will be a huge reduction in the amount of work a graphics programmer needs to do, in order to get a scene rendered. More time will be spent on authoring the data, with very little time setting up the hardware.
I think I am in the right time. I am not the best expert here, there are many far more talented in these forumns. My skills are all self taught, but find myself in an industry outside of games and film that I am fighting hard to push in the direction of realism. People here complain that things look “cartoonish” yet fight me for bringing techniques from the gaming and film industry to make things more “real”. I’m just a small fish, but the it is the struggle now that I enjoy. In a hundred years, I would be looking for another struggle, maybe AI, robotics or something similar because this fight will be done. But I am glad I am in this struggle, I enjoy it.
I’m not necessarily talking about graphics here… Just having this tool at your disposal. Have any of you done something like write a newtonian simulation of galaxies? Imagine what Newton would have done to be able to actually ‘play’ with such a simulation…I know he could play the simulations in his mind, but a computer would have given him more opportunities to experiment…
I’ve just realised I’m talking bollocks on an opengl forum - forgive me, I’ve been talking to a lot of non-programmers about programming, and it’s clarified how lucky I am to be able to program.
Dorbie, I randomly distribute the stars in the xz plane and then recycle those that are outside a certain radius.
I use Keplers law to calculate the stars velocity so that it stays in a stable orbit. There was a flaw in my code which is why some stars get flung out immediately (I have since fixed this).
The two parent stars have random mass the rest have zero mass.
The latest version (not yet released) loads a greyscale density map and the stars are distributed according to that. That way I can use real images of galaxies as templates.