more thoughts on torch height

ok experts,
Why does the torch height control have to be separate from the
normal z axis? I understand that it must be controlled by the
changing distance from the work as opposed to being controlled by a
cad drawing input, but if an encoder can tell a program to correct
its course then tell a stepper or servo to respond, why couldnt a
short program be written based on variances in torch voltage( as in
the rutex system and others) to give step and direction signals. It
seems it would be only a simple if/then program ( ie:if voltage
increases then go clockwise, if voltage decreases, then go counter
clockwise). All the torch height controls I've seen out their
iclude a mechanism for moving the torch and are therefore quite
expensive. If someone were to come up with a system to utilize an
existing z axis for reliable torch height control, I would think
there would be a good market for it. It seems a natural for someone
like Marris or Dan Mauch. I can see where it might create an
interface problem with current moter driver software, but, if you put
a plug on your Gecko or other driver that could be disconnected from
the computer and plugged into a black box. Or a more sophisticated
programmer might write it into the torch control program that when
THC is active Z axis commands from existing moter driver software are
toggled off, all being run from the computer, rather than a black box.
So what about it? Does my lack of knowledge of the facts have me in
fantasy land? or is this doable?
Thanks Dale