I'm not sure I understand what you are saying and how it is a refutation. The line you pulled was in fact his argument for why memory addressing is an important qualifier for his definition and bit-length of registers is insufficient in his mind. The line is from where he argues that even though the 68k is considered a 16bit machine due to how the ALU works, it can address 24bit addresses, making it closer to a "real" computer by his definition. I'd reckon he'd say (as he did :) ) that the 6502 isn't a "real" computer because it can only address 64K of memory.

I think "modern" computing is a better term than "real" computing for what he means, but it's merely presumptuous of a definition, it's nowhere as extreme as calling a table a chair. I'd suggest not letting a definition poison your mind too much as the post is interesting if you can get past the tone.

> Mostly everyone else seems to define the bitness of CPUs by their capacity to add numbers in one go, not by the address space they can address.

Nah, only historically. Yes, "8-bit" refers to the ALU. Back in the day, 16-bit was similar. But even then it was muddy, because when talking about operating systems like Unix or NT the key question about bitness would be in the context of a 32-bit flat addressing model, not really the data width.

By the time the 64-bit era rolled around, the "64-bits" definitely referred to address space.. The original Pentium had a 64-bit data path and had instructions (MMX, eg PADDQ) that could operate on 64-bit numbers - no one would call it 64-bit. By the early 2000s with the big push to mainstreaming 64-bit, it was all breaking out of the 4GB address space limitation - not the width of data.

> define the bitness of CPUs by their capacity to add numbers in one go

This is nebulous and therefore troublesome to define. Are we talking about the ISA or the internal circuitry (ALU and/or data path)? Is the 68000 a 16-bit or 32-bit CPU?