Stefan Wolfe wrote:
My reading would be, computer "CW" (with sound card output electrically
coupled to mike input and does not operate true A1A), is still legal in the
CW sub-bands provided the data coupling from computer to radio is electrical
and not acoustic.

Doesn't matter.

This would imply, of course, that computer programs for CW are not
acceptable (even if legal) across the entire CW allocation unless the output
actually "keys" the CW carrier. They are NOT equivalent to CW.

Yes, they are. Or rather, they can be.

If you have an ideal SSB transmitter, and you feed an ideal sine-wave
audio tone into it, you get a pure carrier output. Key the ideal
sine-wave audio tone, and you have a keyed carrier.

Now of course if the SSB transmitter or the sine-wave isn't ideal, you
wind up with unwanted outputs, such as the suppressed carrier or the
unwanted sideband. How much suppression is needed is another matter,
but I suspect that with modern methods the unwanted products could be
kept low enough not to make any difference.

The big question is whether the signals (keyed carrier vs. keyed audio
tone) look different on a spectrum analyzer. If they don't, why should
FCC care?

73 es KC de Jim, N2EY