This post just represents my thoughts about how people get into some hi-tec trap. They know how to create some eye candy effects, like fish eye or chromatic aberration with GPU shader with 5 lines of code or less - that's cool. But when it comes to a situation when people don't have those shaders for their support, they get confused and say "that's impossible, you can not create such effects without modern GPU" and that's bullshit. A good example is the platform, that I'm currently working on - Wii. I can not say anything more technical about this platform due to some legal rights, but I can say (and everybody know it) that it's rather a low-end hardware. Let's get back to those effects that I've mentioned. You can create some approximation (simple fake, but that's what graphics are all about, ain't it?) using simple linear texture filtering - just create a grid of quads with regular mapping and distort their positions in some manner. With fish eye - just create radial distortion with strength proportional to distance from center of the grid. It looks quite good and is extremely cheap both in performance and implementation.Chromatic aberration - draw red channel on such grid with horizontal distortion with some strength, then green one with 2/3 of this strength and blue one with 1/3, mix them using additive blending et voila!
As a conclusion - this can be really simple, just don't ever think that something is possible only on new GPUs or CPUs or whatever, stay open-minded :) I think that this is quite what low end hardware demoscene is all about.