#seg:offset
I think 8086 chips were available first. The 8088 was supposed to appeal to
cost- sensitive designers. Using an 8088 saved a lot of complicated and
expensive byte swap hardware on memory boards. The 8088 and 8086 have the
same basic architecture, and use the same addressing scheme. Software can't
directly tell the difference between them.
The 8085 used a 64k linear address space. In our embedded systems, we
migrated from the 8085 to the 8086 to gain address space without the
complexity of bank switching. The joke was on us. In just a few years, we
ran out of address space again, switched to PCs, and started having to deal
with extended and expanded memory, which are essentially bank switching
schemes, only much less clean than old fashioned bank switching.
The S100 bus was _the_ bus for CP/M systems for a long time. It had a 24
bit address space, local power regulation, and used a 100 contact connector,
hence the name. Early boxes had front panels with lights and toggle
switches, which were great fun to play with. Something in the architecture
limited you to 2MHz if you wanted a front panel. I have a gutless S100 box,
but never got one running; I have a bit more experience, and expertise, with
Multibus.
The PC changed things because it didn't even have a bus, not in the sense
that S100 and Multibus are buses; the 'bus' we now call ISA is just the 8088
CPU signals, plus some bus-like management signals.
-Mike-