With classic microprocessors, it was generally a simple logic problem figuring out how to attach components not intended specifically for the MPU in question. It wasn't rocket science. It's much trickier with microcontrollers, which, generally speaking, were not initially designed to have anything external used to expand them. It didn't take long for manufacturers to figure out that one MCU wouldn't solve all problems, hence, they developed ways, e.g. SPI, IIC, etc, for doing what previously was done with logic and external "LSI's". In each case, of course, there's a price to pay.

RE