>
something that is pretty simple in realityIt’s not that simple: explaining it properly takes dozens if not hundreds of pages.
> RGB are the primaries in light.
This depends on what you mean by “primaries”. As far as I can tell, your definition is based on common color reproduction technologies, rather than the physiology of human visual perception. That’s fine, but recognize then that the “primaries” chosen for practical use are constrained by economic factors, &c. The three best lights for additive color reproduction are indeed R (an orangish red color), G (a yellowish green), and B (a blue-violet color): this is because those are the colors which maximize the differential responses of different cone cells, as can be seen in this diagram in Hunt’s book The Reproduction of Colour: http://i.imgur.com/ZOdZc.png Of course, such narrow-spectrum sources are not economically/technically feasible, and so instead a typical computer or television uses lights like these: http://i.imgur.com/JHeGa.png By contrast, a typical subtractive system uses primary dyes like these: http://i.imgur.com/qri1f.png (the colored lines on the charts are the reflectances at various concentrations)
You cannot reproduce any color the human visual system can perceive through just three primary lights: every display system has a “gamut”, and for example computer displays have great difficulty displaying saturated blue-green colors.
As Hunt summarizes,
“It will be realized that these three expedients cannot correct for the fundamental limitations of the process, which spring from the nature of the colour mechanism of the eye and the shape of the spectral absorption curves of the best available cyan, magenta, and yellow dyes. What is claimed for modern subtractive processes is that they produce pleasing colour pictures, and that the inevitable inaccuracies are balanced in such a way as to be least noticeable.”
> The "four color" wheel he lists at the end is not wrong, it's just silly: you could pick any points on the wheel and their opposites and have the same thing.
This is not true. Color opponency and the specialness of the so-called “unique hues” have had a great deal of scientific literature about them (physiology, psychophysics, linguistics, etc.), and no, you could not just pick any four arbitrary points. Of course, there’s some learned/cultural component to people’s color categorization too, and there are individual physiological differences, so there’s inter-observer disagreement on precisely what color is “unique red”, etc. But it is indeed true that any color can be described as some combination of red, yellow, blue, green, white, and black: this is Hering’s theory of color vision, the inspiration for the Swedish NCS system, based on decades of rigorous measurements in the 40s–70s.
The current scientific consensus is that color vision can be modeled in a simple way with 2 stages: (1) trichromacy of 3 cone responses, (2) higher-level opponent mechanism. There are many more complicated effects beyond that, and it’s essential to consider adaptation, but those 2 cover the basics.
* * *
All of that said, I’m not completely satisfied with the original essay either. It’s pretty fluffy and hand-wavey, and the jokey language gets in the way. Conflating long/medium/short cones with red/green/blue colors is dangerous because it hides what’s really going on. I wouldn’t, as the author of this article does, call the red–yellow–blue–green anchored hue circle “proper”; there are other equally valid organizations, such as the Munsell system’s, which aims for perceptually uniform hue spacing.
Still, on the whole it’s on the right track.
* * *
I’m sorry that the Wikipedia articles about these topics aren’t clearer and more comprehensive, or I’d point you there. As the article says, the best resource online is Bruce MacEvoy’s handprint.com, but several books have excellent explanations. If you’re interested and MacEvoy’s site doesn’t clear things up I can suggest where to look in the library.