pouët.net

John Carmack doesn't understand nor make art. Thus he may be spanking YOU, but he's nowhere near us. He's just a toolmaker, the smart kind. When scene coders became greedy about efficiency and engineering (which are not are), they became toolmakers as well, and they were slowly made redundant as their tools replaced them.

We are not like that. Choose your option.

I guess I missed nothing while not reading much pouet.net threads. heh :)

do whatever you like/want.

ways of the past provide nowhere forward.

hehe, it's cute to see how 33's dream gets crushed :) 'What!? NO! DOPE WAS REAL! AND JIZZ HAS SOFTSYNTH!!!' hehe

fuck pc!! at this point in the scene, i think demoscene should start to code on horrible hardware. The typewriter demo from tapir was nice :) i'm waiting anxiously for a demo on eg. hometrainer, gprs system in car, or whatever obscure hardware there is :)

I'd like to see a pacemaker-demo.

how about making a prod for those new automatic houses? a smart-house-demo!

I actually enjoyed reading this thread. The discussion itself is pointless but the quality of the posts is unusually high for pouet.net (and I'm already used to ignoring anything saddam "lator" hussein says)

IMHO sw demos are being made in order to "make a 3d engine". The "effect" in these demos is the 3d engine (and maybe some others too in later demos). The hw demos from the other hand are being made in order to make "3d effects" (f.e. marching cubes). Of couse you can do them with sw engines too, but with hw acceleration you'll have more free cpu power, li-bili-trili-near filtering, hires, etc. You still have have hires and li-bili0-trili-near filtering and enough cpu power with sw for making cool effects, but then you'll need a 2GHz computer. And the demo will become "high end"...

33: what about that link you pasted? what should be untrue about it? Sure, it looks good, but variform / kewlers have similar effects in it.

BadSector: Demos are made to look cool, regardless of their code.

BadSector: Demos are made to look cool, regardless of their code.

hmm
demos for dildo!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
would rock!! and the syncs would also make vibrate it more!!!!!!!!!!!!!!!!!!!!!!!

a dildo doesn't vibrate dumbass!! a vibrator does! :)

Interesting to watch peple talking about demo-making history not mentioning C64... or Amiga.

I miss 2d

My personal opinion is demo groups should start pre-production work in readiness for first generation realtime raytracing gfx cards. It's time to look towards the inevitability of the tomorrow, not the reality of today.

I'd also like to have sex with that super hot chick I was talking to on Saturday, but I guess that's just the fantasy of yesterday.

Thom: Then you should penetrate her with your fallus.

í just care about other stuff too.

doom three - ~_~ - doom three

thom, I don't think there will ever be realtime raytracing hardware...
A Cray simply doesn't fit on an AGP board :)

Seriously though... Raytracing requires basically a complete general-purpose CPU, with complete addressing and all that.
It's much more complex than stream processing, which is basically what 3d cards do...
And 3d cards are getting so fast and advanced, that they are getting very close to raytracing results (you can actually get away with some recursive reflections using dynamic cubemaps... Rendering a scene 10 times works :).

I guess it will be quite some time until we have realtime raytracing hardware, if we ever get it at all. nVidia and ATi are clearly not headed in that direction yet.

Actually Scali, it is already possible to implement a simple raytracer using ps2.0 hardware.

I know that, sagacity.
But it's not a raytracer, it's more of a raycaster... And afaik there's no texturing/material/whatsoever possible.
The results are very slow and look extremely crap.
It's not realtime raytracing, it has very little to do with it, really. It has neither the quality nor the speed for that. It's just a kludge.

And in fact, it's even possible on R8500.

yes, *crappy* results are what you get on R8500-based hardware, which is why I said *ps2.0* hardware, which does allow for somewhat acceptable results. and there's already some flow-control (even more complicatedly so in ps3.0), for instance.

Still the ps2.0 stuff sucks afaik in quality, features and speed. No reason to even want to try it. And there's no reason to believe that vertex/pixelshaders will be more suited to raytracing anytime soon.
Proper raytracing just cannot be implemented on streaming DSP hardware.

We'll see realtime raytracing hardware within five years. Efforts are already underway. http://www.saarcor.de/