I'm confused—the author keeps referring to rotations. The very first sentence is "To represent 3D rotations graphics programmers use Quaternions", even. But I don't think you don't need quaternions to represent 3D rotations. You need quaternions to represent translations, because translations aren't linear functions in 3D. (To be a linear function, v * 0 = 0, but that isn't true for translations). My understanding was that the extra dimension was to be able to represent translations. Am I mistaken?