According to my ARRL Antenna Handbook, the Windom Antenna was described
by Loren Windom in QST in 1929.

The design is a horizontal half wave (ie single band) fed by a vertical
single wire feedline attached just off centre (~14%). Explanations go
that this approximately matches the feedline Zo (which is quite high)
with the horizontal wire. It is single wire (ie ultimately unbalanced)
feedline and therefore radiates. The Antenna is fed between the source
end of the feedline and ground, and the load impedance should be
somewhere in the many hundreds of ohms. The feedline carries an
appreciable net current.

More recently, the Off Centre Fed (OCF) Dipole design emerged,
principally as a multi-band antenna. The OCF Dipole is a horizontal wire
with a coaxial feed and coupling transformer (often called a balun)
attached offset from the centre of the dipole. The feedpoint excursions
at a half wave length and harmonic frequencies are much lower than centre
feeding, and may be operated as a multiband antenna with reasonable
efficiency, though it probably really needs an ATU at the tx end of the
coax. The OCF dipole feedline does have current flowing on the outer of
the outer conductor, at least as a result of the assymetric coupling to
the dipole legs, and to some extent because the ineffectiveness of
practical coupling transformers to isolate the feedline ends from the
differing voltages on each dipole leg. The feedline carries an
appreciable net current.

Then along came the Carolina Windom, which appears to be a OCF dipole
with a proprietary (ie secret, undescribed) coupling transformer, a
vertical coax section (feedline and radiator) and a proprietary (remember
the meaning) "isolator" located at a given distance along the coax to
prevent the current flowing on the outer of the outer of the coax from
flowing further towards the transmitter. The isolater would appear to be
a ferrite choke and it would introduce a series impedance (reactance and
resistance) to current on the coax, so influencing the establishment of
the standing wave pattern on the outer of the outer of the coax. You
might naively think that this isolator prevents current flowing into the
shack, but that is unlikely.

In all these cases, there is an expectation that the feedline carries a
net radiating current, and it seems to me, that if you don't want to
bring that into the shack, you need to design an appropriate solution.

In the case of the true Windom, it seems the easiest solution is to end
the single wire feedline outside the shack and place a matching unit
connecting to ground and the single wire feedline at that point, and
transforming the load to something suitable to coax or balanced feedline
to the shack.

In the case of the OCF Dipole and the Carolina Windom, shunting the
current on the outer of the outer to ground outside the shack is a
potential solution. Series chokes might help, but the magnitude of the
choking impedance is limited, and their effectiveness could be improved
greatly by a low impedance shunt to ground.

Comments?

Owen

PS: In todays paranoid world where rules in many jurisdictions restrict
the maximum permitted exposure to electromagnetic radiation, antennas
such as these with radiating elements that are close to areas accessible
by people are a safety challenge.