Q: Do you recommend using the Boost-Buck Cascade for off-line rectification and regulation as is often done in industry? The Boost is used for PFC, an isolated Buck for load regulation.
A: No.

Q: Why not?
A: As mentioned earlier, the Buck is incapable of line rejection.

To see this, think of the unisolated Buck during D. The output inductor L2 and output capacitor C2 form a voltage divider. Any perturbation of Vg appears directly across the load in the ratio (1/sC2)/(sL2 + 1/sC2). Nothing but D=0 can prevent this. In other words, the disturbance has already propogated to the output--too late for feedback to do anything about it!

Q: What is the right thing to do?
A: Use the isolated Cuk Converter. It is easily isolated, providing multiple outputs of both polarities.

Note that since either P or Q is always on, there is no direct path to the output from Vg. This "complete conversion" equates to a D/D' DC gain, and gives the feedback time to correct for line variations, keeping them out of the load.

Q: The common industry practice is to Current-Mode Program the Boost Converter to achieve unity power factor. Without CMP, how can this be done?
A: Very easily. Remember that the current in the transistor of a Cuk Converter is just I/D' = Ig/D, which flows when Q is on during D. If the Source current is averaged over Ts = 1/fs, the result is the input current Ig, which is the desired current drawn from the line.

This signal may then be fed back, comparing it to the half-sinusoidal input voltage Vg.

Q: How is output regulation managed, then? In most systems used today, the Isolated Buck Converter runs on its own feedback loop to maintain proper output voltage.
A: A little thought will show you that the two loops cannot be entirely independent. Otherwise, the holdup cap, which resides between the two converters will sometimes sag too far, or else be overcharged and fail catastrophically.

In other words, unity power factor cannot be maintained under transient line or load conditions. Likewise, remembering the Buck's vhat/vghat Achilles Heel, perfect load regulation is impossible as well. In the end, a compromise must be struck between the two.

An isolated Cuk Converter would feed back both line curent and load voltage, regulating both in the best possible way.

Professor Erickson of UC Boulder has done some interesting work along these lines, investigating off-line rectifying regulators with Boost-like inputs. He points out that some capacitance is needed to make up the difference between the line voltage and the output voltage in such a circuit, which is clearly just the energy transfer caps in the case of the Cuk Converter.

But, in order to achieve unity power factor, there is also a fixed line current to be drawn, and an unpredictable load current. It is the corresponding difference in these currents that tends to pile up on the ETC, forcing a compromise in performance.

Q: Can't I just make the holdup cap bigger?
A: There would be no end to it. Most holdup caps are already big enough to constitute something of a safety hazard.

Q: So you're saying that I might as well "topologically reduce" my switch, since the duty cycles of the two converters aren't really independent anyway.
A: Exactly!