[From Bruce Abbott (980309.1245 EST)]

Bill Powers (980309.0929 MST) --

Tracy Harms (980309.06)

Bruce, you've brought up the example of a thermostat several times now.
You have explicitly called it an example of a control system, but that
is just not so. A thermostat is described much better by a
lineal-causal model than by a control system model. A thermostat
changes as a function of temperature and setting, period. There is no
preference to be found there. There is no control, because there is no
feedback. There is nothing in the output of a thermostat, alone, which
affects the input to the thermostat.

Tracy, instead of arguing with you I think I will just recommend that you
reexamine your concept of control, and of how a thermostat actually works.
It doesn't look to me as though you understand either one. You are going to
find that the entire massed might of CSGnet is arrayed against you here.

Not the _entire_ massed might; I think that Tracy is talking about the
thermostat itself. See Tracy's last sentence above. The thermostat is only
a component of the control system; it is not the entire system.