I could engage in some more speculative thinking.
I could be trying to design a hypothetical analog scheme for modulating color information, that belongs to the Y’UV system for representing colors. But I’d like my system to have as advantage over NTSC, that if the chroma sub-carrier gets phase-shifted, due to inaccuracies with the analog circuits, the result should be a shift in hue, which reverses itself from one TV scan-line to the next, as well as from one frame to the next. Just as viewers don’t normally see dot-crawl when they watch an NTSC-modulated signal on a monochrome receiver, they should also not be able to see the hue-shift, due to analog-circuit issues, with my hypothetical modulation scheme.
Consequently, the receivers for this type of signal should not have a Hue potentiometer.
But I discover a problem in my scheme. The U and V components are to be modulated onto a chroma sub-carrier, using quadrature-modulation, just like NTSC was. And yet, I’ll discover that I can only get the clockwise versus counter-clockwise reversal to take place, if I invert either the U or the V signal-component, but not if I invert both, nor if I just invert the sub-carrier, thereby inverting both U and V:
The problem follows, because every signal which gets modulated onto a sub-carrier, using quadrature-modulation, throws sidebands. Hence, if I was to place the sub-carrier frequency just-beyond the frequencies already being used to encode luminance, I would also need to invert both U and V, by default, to eliminate all the dot-crawl. What can I do?
Well, I can decide that my form of quadrature-modulation should also be a kind of vestigial sideband modulation. I could decide that the Human eye sees differences in blue-yellow content better than red-green differentiation, so that specifically my U signal-component is to have the greater frequency-response. This was quite a common assumption, back in the analog days. Then, the U component of the sub-carrier could extend further down in frequencies, into parts of the signal which once belonged only to the luminance signal. But I’d have to place my sub-carrier itself, high enough above the highest luminance frequency, so that both sidebands can be used to modulate V – in addition to U.
Together with that, I’d choose only to invert the U signal-component, from one scan-line to the next, and therefore, also from one frame to the next. My opportunity wouldn’t exist at the same time, to invert – or alternate – V .
But, I’d keep the idea, that the reference-signal that defines the phase-position of the chroma sub-carrier, should be in-phase with either blue or yellow, so that I’d also choose to plot U in the vertical direction on my phase-vector-diagrams, since the reference-signal is also a major point-of-reference.
P.S. It is more conventional today, to observe the regular sequence of U and V, and therefore, to plot those as one would plot X and Y. But, in the old days of analog TV, the symbols used, IIRC, were not U or V anyway…