I ran TSR-80 systems and I never had a S-100, I was always fascinated how it could have multiple motherboards on it with different CPUs. I think S-100 was more similar to SCSI or Ethernet than to a PCI bus.
The Jonathan bus would probably be a lot more robust, however, as power delivery was an issue on loaded Apple II's (with too much current flowing through too few VCC and GND pins).
there was a single address space. so every board had a set of dip switches to give it its address. which was always a big source of pain
I can't imagine apple shipping a computer product like that, so maybe you get the high bits based on what slot you're in and a config eprom to do discovery?
I suspect if Apple had done this they would have had someway for the cards to auto-negotiate.
A few years after the Jonathan prototype, I remember fighting with the DIP switches on PC ISA cards and setting the the correct IRQs, and then seeing a friend drop in a NUBUS card into their Macintosh II and the hardware was magically configured. I wonder if NUBUS could have multiple masters?