Noise is everywhere. Last night was a cool night and the
windows were open. Unfortunately, conditions were just right
for transmitting the noise of traffic on freeways some two miles
from our house. I didn”t sleep well as a result. We”ve talked
about noise in earlier columns. That cosmic background
radiation from the Big Bang is responsible for some of the noise
on your TV picture. Lightning and a host of other sources can
cause static or other background noises when listening to a radio
program, particularly on one of the weaker stations. The
problem is to transmit and receive information, in the form of a
TV picture or Metropolitan Opera radio broadcast, without
distortion in the presence of all that noise.
I was recently in the main reception area at Lucent Technologies”
world headquarters, also home of Bell Labs, where I used to
hang my hat. I couldn”t help noticing a bust of Claude Shannon
displayed prominently near the reception desk. Shannon died
last February, having spent the last years of his life in a nursing
home with Alzheimer”s disease. A sad ending for one of the true
intellectual giants of the 20th century. I couldn”t help thinking
how ironic it is that, as Shannon was fading from the scene, so
too was Bell Labs being transformed from a national treasure
into a shell of its former self. I”ve seen a number of articles on
Shannon in the past few years, one of which I received from a
friend just a week or so ago. The article is by M. Mitchell
Waldrop in the MIT Technology Review and is titled “Reluctant
Father of the Digital Age. Claude Shannon”.
Picture the year 1948 at Bell Labs. Just a week before the year
began, John Bardeen and Walter Brattain had invented the “point
contact” transistor. William Shockley, incensed that he, the
leader of the group, was not in on the invention, was to put forth
intense effort on the theory and come up with the idea for the
much more useful “pn junction” transistor. Strangely, he had
already proposed the idea of what was later to become the “field
effect” transistor, which is today served up by the millions in
your computer. The ducks were lined up for Jack Kilby to
invent the integrated circuit some years later to begin the silicon
chip revolution.
In another corner of Bell Labs, Claude Shannon, at age 32, was
about to publish a theoretical paper that would also change the
world and complement perfectly the capabilities of the future
silicon chip. The paper was called “A Mathematical Theory of
Communication” and appeared in the October 1948 issue of the
Bell System Technical Journal. I also published my most widely
read paper in the same Bell System Technical Journal in 1960. I
wonder why it hasn”t received the same acclaim? Perhaps it”s
because my subject, the solubilities of impurities in germanium
and silicon, doesn”t quite stand up to a paper that revealed and
defined the very essence of information itself. Shannon showed
that any kind of information, ranging from music to text to
telephone or radio messages or TV pictures, could be encoded in
binary digits, what we now know as bits. Shannon had fathered
the field that became known as information theory. This paper
alone stands as one of the most influential papers of the 20th
century.
However, a decade earlier, Shannon was no slouch either. His
dissertation for his master”s degree in electrical engineering at
MIT was titled “A Symbolic Analysis of Relay and Switching
Circuits”. Not a title that would indicate that some would call it
the most important master”s thesis of the 20th century. In it,
Shannon said that if a statement is either true or false, you could
also characterize the statement by a 1 or a 0. Maybe not a
revolutionary idea, but he carried it a step further by noting that a
closed or open switch in a relay circuit could be designated as a 1
or a 0. Still not impressed? Shannon took this idea and showed
that a simple relay circuit could be used to carry out a logical
operation.
For example, one simple relay circuit is essentially just a couple
of switches lined up. Suppose you have two switches in the line
leading to your reading lamp. If a switch is closed it”s a 1, if
open it”s a 0. If switch # 1 is closed AND switch #2 is closed,
that”s a 11, current flows and the lamp lights. If either switch is
open, 10 or 01, or if both switches are open, 00, your lamp won”t
light. You”ve just performed a logical operation, the AND
operation with your simple circuit. If something is true (switch
closed) AND something else is true (other switch closed) THEN
something happens (current flows). You could also wire up your
two switches so that if something is true (switch is closed) OR
something else is true (other switch is closed) THEN something
happens (current flows). You”ve now performed the logic OR
operation with an electrical circuit.
Still not impressed? Remember, this was back in 1937-1938
when calculators that could add, subtract, multiply and divide
were large instruments about the size of typewriters. These
calculators contained an assembly of wheels and gears and made
an impressive amount of noise doing simple calculations. They
were definitely not digital! I was still using them in graduate
school in 1950. Shannon, in his early 20s, had anticipated that
computers could be programmed through such AND and OR
logic operations to “decide” what to do next based on the input
given them. Today, your computer is chockfull of 1s and 0s in
the form of open and closed switches in circuits carrying out all
sorts of logic operations even as you look at this on your
monitor. That was one helluva master”s thesis!
What about noise? Back to 1948, Shannon”s digital encoding of
information was astounding. But what really caught the attention
of many was something equally breathtaking. He also showed
that once information was encoded digitally, it could be
transmitted without error, even in the presence of noise. If there”s
a lot of static or other noise, you have to encode the information
in more digits or bits than if there were no noise. Shannon also
showed that with large amounts of noise, you need error-
correcting codes built into the system.
Younger readers probably won”t know how gingerly we had to
handle and clean those fantastic long-playing vinyl records in
those ancient times when high fi was coming into being. Now,
you can have noise in the form of fingerprints and maybe even
minor scratches on your CDs and the music sounds perfect,
thanks to Shannon and the error-correcting codes he inspired.
These CDs are loaded with 1s and 0s in the form of tiny pits or
no pits that are massaged to give you the perfect sound. (Your
speaker may not be up to reproducing what is fed to it, however.)
The MIT article compared Shannon to Einstein in having the rare
ability to look at a complex problem in a completely different
way than us ordinary mortals. One of Einstein”s achievements
was to show that an object can”t go faster than the speed of light.
Shannon also came up with a speed limit in terms of how fast
you can send information error-free. If you try to send
information any faster, don”t expect perfection at the other end!
You might have thought that the immediate public appreciation
of the magnitude of Shannon”s achievement in his 1948 paper
would have resulted in his enjoyment of his newfound celebrity.
On the contrary, he hated it and kept out of the public eye as
much as possible. In fact, he was quite upset that his ideas were
being greatly oversold. At least that”s what he thought. The
world was awash in articles dealing with ”information theory”. In
a 1956 paper, Shannon wrote that information theory had
“perhaps ballooned to an importance beyond its actual
accomplishments”.
Perhaps Shannon”s shunning of the limelight explains something.
I decided to look up what was said about him in the hefty
volumes of “A History of Engineering and Science in the Bell
System”. I found only a few lines devoted to him in one volume
published in 1978. Shannon”s 1948 paper was only mentioned
in passing. I used to ride the bus to Bell Labs with Mort Fagen,
editor of that volume. Were he alive today, I would chide him
for giving such short shrift to one of Bell Labs” most illustrious
figures!
MIT hired Shannon away from Bell Labs in 1958. I had been at
Bell Labs for six years by then but never saw him riding his
unicycle down the halls late at night. I do recall later seeing his
famous “Ultimate Machine” on TV, possibly on the Johnny
Carson show. This machine is a box with a switch on the side.
When the switch is thrown the lid of the box opens and a hand
appears. The hand then reaches over the side of the box, turns
off the switch and retreats back into the box as the lid slowly
closes! Shannon loved to make silly things like that.
Incidentally, as I was in the midst of writing this column, Brian
Trumbore brought over an article from the New York Times
about the death of Sir Fred Hoyle this past week. Hoyle, an
outstanding astronomer, was the one who coined the term ”big
bang”, about which we”ve said so much in this column. Ironically,
he meant the term to be derisive and didn”t believe the Big Bang
theory. In fact, in 1948, Hoyle published his own theory that the
universe did not have a beginning but was in a steady state. Stars
and galaxies had beginnings, but not the universe. His view,
held to his death, is not shared by many in light of all the
evidence supporting the Big Bang theory.
On the other hand, those CDs of Beethoven”s Fifth or Madonna
demonstrate every time you play them that information can be
encoded in digits and transmitted faithfully. Claude Shannon
told us that in 1948. Today, 53 years later, digital radio and
digital TV are just emerging to embed his legacy even deeper
into modern life.
Allen F. Bortrum
