As an alternative to using a choke on the feedline, what sort of
results might one get with a standard J-pole by using a half-wave
coaxial balun and tapping up a bit further on the elements?

Still don't think that would clean up spillover currents...
If I were to use a J pole, and wanted decoupling, I would
add a 1/4 wave coax section below the feed, and then
have a set of radials at that point. They can be grounded
to the mast, or left free, as long as they are connected to
the shield. The 1/4 wave of coax is physical length, not
electrical. But I've found in testing, it's not all that critical.
Of course, he wouldn't be able to claim "no radials" at that
point, but it would be pretty well decoupled. This is the same
basic design cushcraft used with the ringo ranger. In that case,
the coax length was 50 inches long, to a set of 20 inch radials.
They used the longer length due to the dual 5/8 design.
They actually seem to claim that 50 inches of coax as a 3rd
radiating element, but I don't quite see it that way. If that were
truly the case, it would beat the isopole.
The comparison between the isopole and the ringo ranger 2
show how important decoupling is. They are both appx dual
5/8 designs. No real difference in element length.
The isopoles improved decoupling is what make it the winner
when you compare the two head to head.
A sleeve would also be easy to use with a J pole...
Actually, I sort of prefer the center fed 1/2 wave "sleeve"
dipole vertical, with a 2nd lower decoupling sleeve, over the
usual J pole design. MK