Subject: 4004 microcomputer history--intel
Community Memory: Discussion List on the History of Cyberspace
The history of the first (Intel 4004) microcomputer is published in several
IEEE journals, but briefly capsuled here, as I recall it.
1. Intel was founded as a memory chip company in 1968 by Noyce/Moore.
2. Hoff was hired from Stanford as Intel manager of Applications Rsrch.
3. Early memory chip acceptance was poor and Intel needed some
revenue and decided to take custom chip work to augment revenue.
4. Noyce had 'friend' in Tokyo who need custom chips for calculator.
5. Intel took on contract to build calculator chip set for Busicom,
Hoff was assigned to manage the project.
6. Busicom designers (Shima, Tanaka) defined 8 custom chips for desk
top calculator (printer, display,keyboard, arithmetic, control, ROM,etc).
9. Intel had 2 chip designers and insufficient design bandwidth to do
8 custom chips; these 'random logic' chips are 'difficult' to design.
10. Sept 69 Mazor joins Hoff (from Fairchild where I was a computer
designer, and just finished decimal floating point arithmetic unit).
11.Shima was my officemate and I worked with him on calculator logic design;
I had just done similar decimal arithmetic unit at Fairchild.
12.Hoff proposes a computer CPU on a chip with other memory chips holding
'code'. Since microprogramming was popular, we considered it a micro-
programmed calculator, and called it a micro-processor.
13.I worked on 4004 architecture with Hoff and helped to write test code
to implement arithemtic and control programs to implement desk top
calculator project. Busicom objects at first since we cannnot prove
we can do a calculator, but by 12/69, Busicom abandons their custom
chips and we begin the detailed design of the first microcomputer.
14.Faggin joins Intel from Fairchild (2/70); has responsibility for
chip design project ---now behind schedule. He works over-
time and designs all 4 chips: (ROM, RAM, CPU and Shift register).
Mazor helps with logic checking (simulation); Shima helps in chip layout
checking. Design to production takes about 18 months.
15. Chip set is produced, code written, and ROM chip's built. Busicom
introduces several calculators using the 'MCS-4' chipset.
16. Hoff/Faggin/Mazor encourage Intel to refund $60k development fee to
Busicom and Intel to sell MCS-4 chip set as standard product.
17. November 1971 Intel runs ad for the MCS-4 in Electronic News with:
"Announcing a new era ..." a prophecy. Chip set is introduced as
an Alternative to designing with random logic chips...using programming.
18. Computer on a chip is not correct...actually CPU on a chip, and this
is not patented by Intel as it is an 'obvious' extension. Intel does
patent specific features of MCS-4. Idea of 4-bit computer seems silly
to real computer users; conventional computers were 16 minicomputers;
MCS-4 seems like a toy to them; a joy to those not computer literate.
19. Mazor proposes 8-bit CPU to Victor Poor of Datapoint, with Hoff and
Mazor writing the proposal and specfication (see separate history).
20. Please see articles by Hoff, Mazor, and Faggin for more.
Posted by David S. Bennahum (firstname.lastname@example.org)
Moderator: Community Memory
http://memex.org/community-memory.html A CPSR Project -- http://www.cpsr.org -- email@example.com
Materials may be reposted in their *entirety* for non-commercial use.