Of course computers boot from HDD, that's where the operating system is stored!
Yes, that's what we commonly say, but that doesn't make it true (in detail). Just like the HDR-FOX, a computer can't actually read anything from a HDD until it runs some code (ie instructions) to tell it how, and those instructions have to be available in the computer's memory from the off. So, straight after a reset (or power-up), a computer is reliant on the code in non-volatile memory to boot, and only after that does it look for more instructions from the HDD (or not, in the case of the HDR-FOX, because the HDD only stores data as standard).
Think yourselves lucky. When I started playing with computers in the 1970s, there was no non-volatile memory so the first thing the operator (ie me) had to do was load instructions into memory manually, using a bank of switches. You could load the whole program that way, very laboriously, but it was error prone and slow, not to mention tedious! In fact, the code you would load manually would be just enough to make the computer fetch data from a paper tape reader, and then the rest of the code would read in from pre-prepared punched tape. The paper tape might then contain instructions to read from the disk pack, from where the full operating system (what there was of it) finally got loaded.
Bear in mind that a top-of-the-range 14-inch four-platter "Winchester" disk pack held 35MB, and if you needed more storage than that you had to physically swap disk packs!
