Thanks, Michael

[From Bill Powers (930908.0700 MDT)]

Michael Fehling (930907) --

Michael, that was a review I would have killed for in 1974.
Thanks -- I hope chapters 16 and 17 don't cancel your enthusiasm.

You're quite right in saying that the higher levels need work. So
do the lower ones, for that matter. What we need are experimental
explorations, but it isn't easy to come up with experiments
demonstrating control of higher variables. If I could have
thought up 11 beautiful experiments, figuring on taking two years
per experiment on the average to write the proposals, get the
support, train the graduate students, get the data, and refine
the definitions until they were foolproof, that would have used
up all the time between 1973 and 1995. If 1000 people had been
working on this, it would have been done in a half a dozen years,
because someone would have come up with good ideas a lot sooner.
We PCT modelers have been stretched very thin and have worked
without support -- financial or moral. Up to a point this can be
an advantage, but as applications proliferate it soon becomes
crippling. Just look at the disciplines represented on CSGnet!
How could any one modeler, or any six modelers, develop and
explore the methods for using PCT for an entire scientific
community? We need a whole generation trained in the methods of
analog as well as digital modeling, in basic physics, in systems
analysis, AND in the lore of many disciplines in the life-
sciences, AND in PCT. Until Gary Cziko proposed and implemented
CSGnet, I had little hope of seeing all that come about. Now,
with people already on the net and with people like you starting
to take interest, the near future looks a lot brighter to me than
it did for the first 20 or 30 years. By God, it can be done. I
may yet live to see the founding of a real unified science of

A propos of that, and without pre-empting any formal
announcements, I can report that I have latched onto a young
genius of a biochemist interested in PCT. He has taken a feedback
model from Hayashi and Sakamoto (_Dynamic Analysis of Enzyme
Systems_), and working with me has done the first elementary
modeling demonstration showing that a system with an allosteric
enzyme as a comparator is a true negative feedback control system
complete with reference signal. Varying the concentration of the
reference-signal substance causes the concentration of an
unrelated substance to vary in exactly the same way with very
little lag and with high precision, despite all kinds of
disturbances including changes in the substrate and in drains on
the controlled concentration. The next step will be to
demonstrate an elementary two-level hierarchy of control. This
will open the door to a completely new understanding of the
organization of biochemical systems. I predict massive opposition
from conventional biochemists, but there are getting to be too
many PCTers to allow that opposition to stand for long. And my
young friend is very, very competent in his field; he can't be
ignored. At the very least, we are going to change the definition
of "controlled substance."



Bill P.