Hacker News Reader

Comment by bitwize | Hacker News Reader

Fun fact: The Intel 8008 was intended to be a single-chip CPU for the Datapoint 2200 terminal. It was delayed so ultimately wasn't used in the Datapoint 2200, but Intel had an agreement in place to sell the chip to other buyers.

So that means every x86-64 PC sold has an ISA with a compatibility lineage that goes straight back to Datapoint.

View on HN

More fun facts from "Programming the 80386", Crawford and Gelsinger (chief architect and designer of the 386 respectively), 1987, pp. 1-2

"The 4004 was quickly enhanced to the 8008. These devices, very trivial by today's standards, were novel but hardly taken seriously as computers of any worth...In 1978 the third generation was introduced--the 8086. This marked the beginning of microprocessors as 'real' computers...The 8088, the little brother of the 8086, was used in the IBM personal Computer...and this launched the personal computer revolution."

I wish someone else would have read that for me.

I think that's a pretty Intel centric view of things. Well before 1978 there were the 6502 (1975, MOS Technology) and the 6800 (1974, Motorola). Those were very much 'real' computers compared to the 8008.

Yes, I was surprised to see marketing, even criticism of their own previous products, in a programming guide. Once past the "history lesson" the book is actually very good--with concrete examples on building an operating system and a coherent explanation of IEEE 754 floating point format. Hardware floating point for x86 used to be offered as a separate coprocessor chip--the x87 series--otherwise it had to be emulated in software.

I would also argue that the 8008 was a real computer. The ENIAC was the first "real" computer (in the sense of general purpose, digital), with its ~.005 MIPS (integer add/sub) vs. the 8008 at ~0.05 MIPS, and was a bit larger and more difficult to program. I wrote programs on a 6502-based Apple years before the IBM PC appeared. I guess that was a phantom computer.

There's a book on Datapoint that I read, as research for my book. I read these things so you don't have to :)

I'm also in touch with Gordon Peterson, the chief architect of ARCNet, which was a pretty decent LAN, way ahead of everyone else.

Reply on news.ycombinator.com