Looks like I didn't miss much by being out of town for a few days. The
same arguments and hand-waving are still raging. But I'll try to add one
constructive bit of information:

NEC calculates field interactions by means of very fundamental and very
well established electromagnetic principles. As a result, it does a very
good job of predicting currents in an inductor which is modeled as a
helix. There are some caveats as there always are with any modeling
process, particularly:

1. The spacing between the outsides of the wires in adjacent turns
should be at least one wire diameter, and ideally several. NEC does not
account for uneven distribution of current around a wire (proximity
effect). This mainly impacts effective resistance.

2. The whole model must do some radiating, although even very
inefficient radiating structures are analyzed with good accuracy. A
check of the average gain usually reveals if there's a problem with the
overall calculation. I've gotten quite good results with a model
consisting of a fairly small coil with a wire through the center
connecting the coil ends, and a source at the middle of the wire.

To the extent that the program is providing accurate results, it can be
used as a verification or refutation of measurements, and to test
whatever alternate theories one wants to propose. Cecil likes to split
currents into sets of traveling waves, which in itself is fine. However,
when all the waves are added together to produce the actual current, the
result should agree with measurement and with analysis by established
theory. This means that, again to the extent that the model calculations
are being done correctly, the solution using traveling waves should
agree with NEC modeling results.

One huge advantage of using NEC results is that they're not limited by
lumped constant, traveling wave, waveguide mode, or other approximations
which hold only over some range of conditions. Results should transition
from one to another smoothly since the same fundamental laws apply
regardless of the regime. (One additional caveat, though, when doing
analysis at extremely high frequencies: the wire circumference has to be
no more than a small fraction of a wavelength, because as mentioned
earlier the program does assume equal current distribution around the wire.)

Of course, comparison of rigorous numerical measurements or modeling
results with vague, hand-waving theories with no supporting equations or
other mathematical tools is a total waste of the reader's time, and
that's pretty much all that seems to be happening here. But if anyone is
actually seriously interested in investigating alternative theories or
analytical methods, NEC or derivative programs such as EZNEC can provide
powerful assistance in confirming or refuting them. Unless, of course,
the objective is an attempt to refute the validity of Maxwell's
equations, in which case disagreement with NEC should be expected.

Both NEC and EZNEC provide simple ways of generating a helical model.

Roy Lewallen, W7EL