Although I'm not very familiar with this code, my understanding is that the HyperDrive intercepted the Mac boot sequence. When the 68000 started up, it was hardcoded to jump to a specific I/O address. By playing some games with address lines, I believe that the HyperDrive swapped the memory map around so that its own ROM was dispatched to instead of the Mac boot ROM. Once our software had control, it could install the hooks which allowed it to intercept disk accesses. I don't know any of the details of how our code interacted with the original boot ROM, but that was the basic idea -- install hooks so that disk accesses were routed through the software in the HyperDrive boot ROM. All of the code which accessed the MFM drive resided in the HyperDrive ROM, so no INIT resource was required. Some of our other software (e.g. HyperNet), did patch the system file by installing additional resources, but I don't believe that the basic HyperDrive made any modifications to the System file.
I think that Brad Parker developed the initial HyperDrive boot ROM, although I'm not sure. I came to GCC slightly later on, when we were developing the HyperDrive 2000, and I never had much of a reason to go digging through our boot ROMs.
