Back in the old days, printing was done by setting individual typeblocks in a frame, one block for each letter, number, punctuation mark and space. Even in the hands of an experienced typesetter, the process was relatively slow and prone to errors. The type was all backward in the frame, so spotting typos was tricky. One especially pernicious typo was the substitution of "p" for "q", since they are mirror images of each other in many fonts. Hence the origin of the phrase, "mind your p's and q's".
This system also meant that a printer had to have many, many thousands of individual pieces of type, enough to print an entire newspaper. After the day's printing, the frames were broken up and the letters redistributed, ready to be used again. Similarly, books were printed in jobs of some defined number of copies, then the type was reused elsewhere. Printers couldn't afford to leave type framed up for very long, since it meant having a big chunk of type inventory unavailable for other uses. As it was, the inventory of type was a significant part of printing.
Then, in a rush of technological innovation came LINOTYPE. It's pronounced LINE-O-TYPE, since that's what it did: make an entire line of type in a single, freshly cast piece of metal, called a slug. An operator typed in the text of the line, ready from a normal piece of copy, and the machine cast the line of type. The molten tin cooled while the operator was working on the next line. With the pull of a lever, the finished type slid down onto a frame, automatically in perfect position for loading into the printing press. The line could be in your choice of fonts or mixtures thereof, with specialties of bold, italics, etc. available with the push of a button.
You could set as much type as you wanted, and keeping a frame of set type in the warehouse didn't prevent you from casting more type for other print jobs. When the print run was finished, the type slugs were plopped back in the feed hopper where they were melted down and re-used. It wasn't exactly Print-On-Demand, but there are the same concepts at work.
It took a while for this technological innovation to evolve and mature, and even longer for the traditional publishing houses to adopt it. They'd invested millions in type inventories, old-style printing presses and employees who knew how to run them. Making the change took generations, and there were still publishers who insisted that the old ways were better. In this environment, newer and smaller houses, without the fixed costs of a legacy technology, were able to come quickly onto the scene and do smaller runs, faster publication rates. These would have been impractical for bigger houses with the older technologies, but with the new linotypes and automated presses, the game changed completely.
Before there were indie e.books, there was linotype.
===== Feel free to comment on this or any other post.