The big challenge in transmitting a television picture is that one wants it to begin with a fairly large picture, and end with a fairly large picture, but in-between have to squeeze through a very narrow bottleneck. It would be nice to do without the bottleneck, but doing without would require unreasonably large antennas or worse. Service to a 25-inch television would need a coaxial cable -- including the insulating cover -- over four feet in diameter. Larger screens would need even bigger cables; the new 108-inch television would require an input feed the diameter of a containerized cargo box. The solution is to focus on a tiny section of the original image, sweep over the whole image, and trust that the receiving television will figure out how it's supposed to fit together. That seemingly impossible task was made possible by a systematic approach: first transmit all the white spots of the picture, then the light grey spots, then the slightly different light greys, then the lightish-turning-to-middle greys, and by then you have to give up because there's another picture coming already and you have to get started on that. While this meant the darker images were never transmitted, the expected static in the signal allowed the viewer to be fooled into thinking he was seeing a picture, even when he had accidentally left the set off.
Early experiments tried just moving a very small camera really quickly across the original picture. This wore out the fingers of the camera operator very quickly, particularly when they got too close to the wall and their fingers got pinched. This inspired the invention of mechanical scan televisions, automating the sweep across a picture. This was pioneered by people like John Logie Baird in Great Britain, with a coal-fired steam-powered camera that could not respond quickly to changes in the original images, but had those fascinating little cherry bulbs on the thin rods that spun around and around and got faster and went out more as the pressure built up and slowed down as the pressure decreased, which you could watch all day before you knew it. Fascination with these little bits made it unnecessary to have good resolution in the received picture. Unfortunately the building pressure with the Abdication Crisis caused many of the experimental units to suffer boiler explosions, setting the invention of practical television back weeks.
Impractical television made fine progress with the inspiration that one didn't need to move the camera around; one just needed to have it see different small parts of the image. So one could simply take a picture of the image to be telecast, and chop it up into sufficiently tiny pieces without losing track of the order to fit them back together. This was too slow to be really satisfying, first because of the pause between taking the picture of the actors and stage and waiting for it to develop, and then the pause in waiting for the image to be chopped up. However, the high-speed and fine-grain chopping technology developed for this allowed manufacturers to make a fortune on paper shredders and in making confetti. Worse, the process of transmitting ripped into shreds the original picture, which is why the only record of early television are of programs never broadcast. All things considered they should probably have spent more effort researching practical television. However, the high-speed slicing technology is making a comeback with the new high-definition television images, which replace the old-style jig saws with lasers.
According to legend, television pioneer Philo T Farnsworth drew inspiration for electrical scanning by considering the movement of a plough across a farmer's field. This is slightly exaggerated in that actually, while watching the plough, he was attacked by geese and had to run into the storm cellar to escape. There he found bottles of illegibly labelled preserves, which didn't lead to anything particular. But inventing electronic scanning was much easier than dealing with the geese, and before long he stopped even missing the outdoors. Speculation is that toward the end of his life he was working on an experimental chat room.
Trivia: DuMont Laboratories bought a Passaic, New Jersey, manufacturing plant in 1935. It was formerly a pickle factory. Source: The Forgotten Network: DuMont and the Birth of American Television, David Weinstein.
Currently Reading: Inside Star Trek, Herbert F Solow, Robert H Justman.