Hello, can you hear me?
What a tin can telephone can teach us about the digital revolution
Like many kids, I once constructed a rudimentary telephone by putting tin cans at both ends of a wire. This allowed us to speak into one of the tin cans (the sender) and listen with the other tin can (the receiver). It actually worked, over maybe 50 meters or so, if the wire was tight enough, because the sound waves caught by the tin would be transmitted to the wire, and the wave simply propagated along the wire, exciting the other tin can, where one could hear the speech quite distinctly again.
But what about longer distances? And would a third tin can pressed against the receiver allow for another 50 meters of transmission? As you may guess, probably not. The sound had become too weak, and the noise would thus start to make the conversation unintelligible. But here’s an alternate solution: A person could listen to the sound arriving at the first receiver, and if intelligible, repeat it into the second sender. If the speech is clear and understood at the first receiver, this will work! But now we would need to make pauses for the repeater to do his job. Another alternative would be to amplify the sound at the repeater level. But, the noise would then also amplified, of course.
The interesting thing is that the latter solution was the state of the art of telecommunications at the beginning of the 20th century. Cables for telephone communication were laid across the Atlantic, and repeaters were amplifying the conversation (and the noise!). This is what we call analog communication. However, the idea of a smart repeater as in the former solution was also pursued. But rather than having a human repeater (a sort of lonely job at the bottom of the ocean), the idea was to send discrete symbols, for instance a sequence of zeroes and ones of predefined lengths every second. Now, if the repeater could correctly decode that a zero or a one was sent, it would be able to exactly “repeat” that symbol and pass it on and, ultimately, the message of zeroes and ones would reach its destination irrespective of the number of repeaters.
The crux of the matter was to make sure the symbols would be decoded correctly despite the inescapable noise. Because the noise is random, even if the symbols are much larger than the average noise level, every now and then, in an unlucky situation, the noise would corrupt the symbol, leading to a decoding error. By mid 20th century, this was considered hopeless. Enter Claude Shannon, a polymath genius working at Bell Laboratories. In a masterful work published in 1948, he precisely derived when a message could be correctly decoded, despite the noise.
This was the true beginning of the digital revolution, because symbols (or bits as Shannon named them) could now be transmitted over arbitrary distances, and even be stored for later retrieval. As you know, this lead to all of the digital communications that we know today, be it wireless or over fiber optical cables. Tin can games can go a long way, I guess.