This is the classic vintage computing pissing contest...it's been argued by geniuses and cretins and it's hubris to think you have the One True Answer. The Z80

This is the classic vintage computing pissing contest...it's been argued by geniuses and cretins and it's hubris to think you have the One True Answer. The Z80 is a 4-bit processor (it's ALU is double pumped and 4-bits wide), right? So is the MasPar super (not 32-bits). Obviously, the 8088, mc68008 and ns32008 are 8-bits (size of data bus). And the i386SX is 16-bit. Oh, wait...the 8088 and 68008 are 20-bit processors (address bus width). Unless it's the PLCC version of the 68008, which is a 22-bit processors. IBM 360/20? 16-bit ALU, so there you go. Connection Machines? 1-bit (they're bit-serial) just like the Bendix a couple of generations earlier. While the 68000's registers are 32-bit, the data bus is 16 bit, the A1000, A2000 and A500 that defined the range had 16-bit fetching chipsets, they literally had 24-bit address buses. None of this says "32-bit". It can't be overlooked. Oh, please...you're the one overlooking things and marketing isn't a very good source for architectural definitions. It makes a shitload more sense to say "it's a 32-bit processor because the registers are 32bit" that it does to say "well...the 68000 and 68010 are 24-bit, the 68008 is 20 bits unless its a 68008FN where it's 22-bits, and the 68012 is 30-bits, and the everything else is 32-bits, even tho the fundamental unit of computation since the original spin is 32-bits". There isn't one canonical answer.