>It goes without saying that complexity increase from physics to biology.
>However, the complexity of the subject studied ought not to be synonymous
>with taking a flight of fancy with mere speculations and disregard the basic
>sciences which underlie biology. I have never met an engineer who ever did
>that and did not expected his/her design to fall to the ground. The crux of
>biology and the questions raised there regarding the history of man will be
>answered eventually on how biochemistry connects the micro to the macro.
>
I'm not advocating "taking a flight of fancy with mere speculations and
disregard the basic sciences which underlie biology". My point was that as
systems become more complex, the interactions among subsystems become
important, and the methods used to analyze the components of the system
don't give all the needed information. George Murphy gave a good example
from physics. He pointed out that statistical mechanics does a very good
job of predicting the behavior of large ensembles of molecules. I believe
you can derive the macroscopic behavior of gases from statistical
mechanics, for example. And while PV = nRT is in the end the consequence
of the interactions of individual particles, you wouldn't consider the
motions of individual particles in the design of a pneumatic control
system, for example -- assuming you wanted to finish the design. Another
example that I experienced personally comes from a radar system I wored on
years ago. We needed a an interface to permit the computer to get radar
video, and we assigned a hardware engineer to the design job. He was a
good hardware designer, but he misread the driver manual and when we
started testing his interface it didn't work. He said, "Well, the only
thing this computer knows is ones and zeros," and started an excruciatingly
slow process of trying to figure out why the computer wouldn't talk to the
interface, confining his debugging efforts entirely to the hardware. I
didn't think that was going to work, so based on the specifications in the
driver manual I wrote some diagnostic software to tell me whether the
interface was responding correctly to each of the signals it was supposed
to recognize. Fairly quickly I found an incorrect response, and then he
and I went over the hardware design until we had identified what had to
change. He made the change and everything was fine. The point is that
systems of different levels of complexity usually require different methods
for modeling and understanding them. I suppose we could have started our
debugging process with the quantum mechanical band structure and doping of
the semiconductors in the IC's on the board too, but that wouldn't have
been appropriate.
Bill Hamilton
----------------------------------------------------------------------------
William E. Hamilton, Jr., Ph.D.
Staff Research Engineer
Chassis and Vehicle Systems MC 480-106-390
GM R&D Center
30500 Mound Road
Warren, MI
hamilton@predator.cs.gmr.com / whamilto@mich.com (home)