The basic difference between component (aka "Colorstream (tm)" - Toshiba) and pure RGB is bandwidth. In psuedo techie, psuedo laymans terms here is the basic info:
RGB uses a high amount of bandwidth, the same amount of bandwidth to describe each of the 3 parts of the signal.
Component uses a smaller amount (IIRC ~1/3rd smaller) of bandwidth, it allocates the largest amount of bandwith to the Luminance (Y) (e.g. black and white 'brightness') signal and then half each to the difference between Luminance (Y) and Blue (B) and Luminance (L) and Red (R). Green can be automatically determined in a basic alegebra equation given those three signals.
The basic reason this works is that the eye/brain is most sensitive to brightness changes in Luminance, NOT color shifts in chrominance (color), and is significantly less sensitive to changes in blue and red light so less bandwidth is used to describe the changes.
Truely, component is simply a slight upgrade to S-Video which is Y/C (Luminance+Chrominance), but it gives a bigger bandwidth to the Y-B and Y-R than the simple C in a S-Video signal, thus giving somewhat better image quality on SOME SETS.
Therefore you can produce a virtually undectable "lower quality" image using color stream and 2/3rds bandwidth instead of RGB using full bandwidth. Yes, they still both require 3 cables, but when stored digitally or transferred in other ways, the bandwidth is less.
In fact, color differential video is how the video on DVDs are actually stored in bits.. They're not stored in RGB form like many people think.
Here's a good site giving pretty basic explanations of these technologies:
http://www.projectorcentral.com/component.htmSide note (relative to component video and another thread in this forum):
Component video is just one of the reasons that "hacking a TV set to display arcade video" is such a hard thing to do. Almost all of the sets made since the mid-80's have IC's on them, and the IC takes the tuner or VCR color composite and/or Svideo input and outputs component video for the rest of the circuit all the way up to the neckboard.
Therefore on many sets there are no "RGB" points to hack into directly on the circuit board, just Y/Cr/Cb which of course isn't what computers or arcade boards output. It's not until late in the neckboard just before the tube itself do the signals actually become pure RGB to the 3 guns in the tube, once they are of course amplified up to 25-30 volt signals to drive the guns--far from the .7 volts a computer outputs or 5 volts a video game board outputs.
This is why it's NOT a good idea to attempt to willy-nilly hack a consumer TV set's existing boards to accept RGB directly, besides the whole isolation (transformer) issues as well.