I think any discussion of colour without mentioning the sensitivity curves of Cone cells is pretty incomplete. We don't somehow have perfect knowledge of the frequency of individual photons and build up 'imaginary' colours from that, we just map activation levels from our 3 (usually) types of Cone cells to a continuous space.
Another way to look at it is that colour is effectively an infinite dimensional vector, but we can only sense a 3 dimensional function of that. (Fewer dimensions if you're colourblind, more if you're a bird.)
