Also known as: Claude Shannon, Shannon
Claude Shannon (1916–2001) was an American mathematician and engineer who founded information theory, defining how much information a noisy channel can carry and setting the ultimate limit that every communication system, radio included, works against.1
Life and work
Claude Elwood Shannon was born in 1916 in Petoskey, Michigan, and grew up in nearby Gaylord, an inveterate tinkerer who built model planes and a telegraph line to a friend’s house. He took degrees in both electrical engineering and mathematics at the University of Michigan, then went to MIT, where his 1937 master’s thesis — often called the most important master’s thesis of the century — showed that the algebra of logic developed by George Boole could describe and simplify relay switching circuits. That single insight is the theoretical foundation of all digital hardware. After a doctorate applying mathematics to genetics, he joined Bell Labs in 1941, working during the war on fire-control systems and on cryptography, where he crossed paths with Alan Turing and proved the theoretical unbreakability of the one-time pad.
His playful, eclectic mind never stopped: he juggled while riding a unicycle down the halls of Bell Labs, built a mechanical mouse that could solve a maze, designed an early chess-playing analysis, and constructed a machine whose only function was to switch itself off. Behind the whimsy was the most consequential body of work in twentieth-century communications.
Contribution
In 1948 Shannon published A Mathematical Theory of Communication, and the field of information theory sprang into existence essentially complete.2 He proposed measuring information in bits and quantified the uncertainty of a source by its entropy. His central and most startling result — the noisy-channel coding theorem — established that every channel has a definite capacity, and that as long as one transmits below that capacity, error-correcting codes exist that make the probability of error as small as desired, no matter how noisy the channel. For the common case of a bandlimited channel with Gaussian noise, capacity is C = B · log₂(1 + SNR): raise the bandwidth or the signal-to-noise ratio and more error-free bits per second become possible. This drew a hard line — the Shannon capacity or “Shannon limit” — that no coding scheme can cross, and it turned reliable communication from an art into a target with a known optimum. Shannon also stated the sampling theorem in the general form now taught, building explicitly on the earlier signalling analysis of Harry Nyquist and Ralph Hartley.
Legacy
Shannon’s capacity theorem promised that near-optimal codes existed but did not say how to build them, and much of the next half-century of coding theory was a quest to reach the limit he had drawn. Forward error correction schemes — convolutional codes, then the turbo and LDPC codes that Robert Gallager helped originate — now approach the Shannon limit within a fraction of a decibel, which is why modern 5G and Wi-Fi extract so much throughput from so little signal. For a scanner like GopherTrunk the theory is the explanation of what it sees: whether a P25 or DMR frame can be recovered at a given SNR is, at bottom, a question about operating below the channel’s Shannon capacity, and the FEC in those protocols is engineering built to squeeze toward the boundary Shannon defined. Information theory now governs the design of every digital radio, storage medium, and data link on Earth.
Sources
-
Claude Shannon — Wikipedia, for biography and his founding of information theory. ↩
-
A Mathematical Theory of Communication — C. E. Shannon, Bell System Technical Journal (1948), the founding paper defining entropy and channel capacity. ↩