I don't see how the wires are numbered can be important; it's how they
connect, isn't it?

You are probably right

My rational for always (almost always) using wire 1 to hold the source
is that them I can add or subtract wires without having to change the
source wire designation.

Good point.

What I have noticed is that similar structures (GP with depressed radials,
for example) produce erroneous TRP results.

Please explain "TRP".

"Total Radiated Power"

It will be interesting to try
such computations on variants of your sleeve antenna. My results did not
show significant current on the outer shield of the coax. This may be due
to my inability to implement the "Mininec" ground.

Run it without any ground. Run it without any sleeve. Just put a
trap (or a high value resistor) at one end of a center-fed halfwave
vertical to represent the sleeve (choke) and then add various lengths
of wire on the other side of the trap. With no ground, the current on
the added wire will peak at multiples of 1/4 wavelength. So much for
the trap "isolating" the rest of the antenna.

Ok.

I don't know what "S/M" ground is --

Oops, did I say S/M? What I meant was S/N (Sommerfeld/Norton).

-- but EZNEC supports perfect ground,
Sommerfeld-Norton and MiniNEC grounds. And I believe the object is to
*not* have to input files as NEC code.:-)

Ok, but I like to type it in cold, so see if I can get the cards in the
right order.

Frank