AFAIK, Dave Haynie wasn't directly involved in Hombre development. So even if there were conflicting statements about Hombre between Dr. Hepler and Haynie, I'd always believe the main developer over any bystander.
> that: "When I knew Hombre better, in 1993-1994, it was two chips, and > using the PCI bus and a PC-industry GPIO for all of the basic I/O stuff. "
There doesn't necessarily have to be real hardware for this statement to be true. You can very well implement chips, PCI and GPIO (and anything else, really) purely in schematics, software models and simulation, which is what Dr. Hepler managed to achieve according to his own documents he last updated two weeks before Commodore's bankruptcy.
