AC,
Actually the "classification" subject came up while having a few drinks with our local "boat electrician," as he calls himself (he's really a degreed Marine Electrical Engineer). He's a stickler for proper documentation and definitions before beginning any project - even something as straight forward as plates and windings.
Of course he might of just been messing with me when he asked what "class" of generator it was! But, all the same, each sub-class has a particular winding scheme and I do not know which one fits Holcolm's DC generator portion, if any even do.
Playing a lot with the concept simulation wise and a preliminary thought is to run the system up into the KHz or MHz region (the SiC Fets and micro have the bandwidth to do that); just charge a capacitor; and use an old scheme I developed for use with Ruslan's device - a 2/4 KW inverter similar to those found in modern petrol inverter-generators - but without the internal combustion engine (have a few inverter pull-outs left from the Ruslan project).
Along the lines of your 1) and 2) - plus, at higher frequencies, the system can be made a whole lot smaller. I also agree about building stuff to get the feel, but initially I prefer to use CAE simulation first (after the napkin-cad) since it saves a lot of time and money up front; also provides the documents certain "boat electricians" seem to require.
It's a nice concept since you have a solid 60/50 Hz 120/240VAC with very low THD, plus you can easily parallel them. Published the details in the old (banned) Ruslan thread here a while back. Also works for 400Hz systems. All in all, it puts a more versatile back-end on the system as well.
My conclusion regarding the "extra magnetic field and/or induced current;" - by mitigating the system counter-emf (found in converntional designs) you automatically gain an extra 85% to 94% by rotating only the magnetic field in a concise manner. Some will argue but not loosing that inertia (drag) every time you move past the peak flux apogee costs a great deal of energy - call it combined cog torque, magnetic drag, back-emf, air turbulance, bearing friction, whatever! The trick is simply moving the mythical "flux" with minimal interaction (no in-the-loop Farady/Lenz).
Anyway, far too premature to conclude anything...
SL
Let me try this (?): A pulse creates a magnetic spike (it's inductive so it takes a bit of time to peak, 5 tau), then the pulse goes away (it hides from the system), mean while the magnetic spike "fluxes" the secondary or stator pole which, in turn, creates current in the lapped stator winding (this winding is not in the loop so to speak - it goes way over to the other magnetic pole) and doesn't cause an interaction with the original inducing pole since it's spacially segregated. Any reaction is not "felt" by the original magnetic "flux" instigator. Even if the secondary magnetic spike did get back to the primary, the primary is still hiding and wouldn't even notice!
Like a snowball fight - you pelt the guy and immediately hide behind a tree - he immediately retaliates (180 deg out of phase) but you're already hiding, so big deal. He can't hurt you!

Oh, for crying in the sink - good night!