Amount of radials for Verticals

It does have primarily vertical radiation, and it doesn't use ground radials.

I modeled it in EZNEC and compared it against a ground plane with the same
maximum height using 4 elevated radial wires about 4" above ground. The GP
had a bit more gain for identical ground conditions - in the range of 0.6dB.
I didn't try putting a set of radials under the C antenna, using more radials on
the ground plane, varying the height of either antenna, different ground
conditions, etc. But they give essentially equivalent results, and minor
changes to one or the other might shift the advantage.

The model still estimates about 5dB of ground losses for both antennas.

It can be used on multiple bands with a tuner, though the SWR will be high
(the model suggests >100 : 1 on all bands except 20m and 10m, and the
pattern isn't exactly optimum on 10m.) I'd consider it a single band antenna
unless you have the tuner right at the feedpoint. Common mode currents
are likely to be a problem: one choke balun might not be enough, especially
on other bands besides 20m.

It probably is most similar to a vertical dipole with the ends bent around, but
a ground plane with elevated radials is basically a dipole anyway.

The dimensions aren't nearly as critical as some of the online calculators might
make it appear: I've modeled versions with narrow or wide spacing with little
change in overall performance. Wide spacing requires less overall height, and
affects the relative lengths of the wires on the open side for a 50 ohm match,
but you can customize the dimensions to meet your specific needs.

If you were to straighten out the top wire then you could get it to work as
a vertical dipole with one end folded on 20m, and a J-pole on 10m. (I have
a design for 80m / 40m that I want to try some time using that approach.)
That would require more height, of course.