Of the computers produced and prototyped by Commodore, most people are likely well-acquainted with the PET, VIC-20, C64 and C128, as well as the never released Commodore 65. Of these systems many examples and plentiful of documentation exist, but probably among the most rare is the Commodore 900, as recently covered by [Neil] over at RMC – The Cave on YouTube. The Commodore 900, conceived in 1983, was intended to become a microcomputer based on the 16-bit Zilog Z8001 CPU that targeted businesses as a UNIX workstation.
Only fifty prototypes were ever built of the C900 and no hardware was ever sold, even though the intended UNIX-based OS (MWC’s Coherent) had already been ported to the Z8000 and the rumor mill suggested a release in 1984. Although UNIX workstations were rather popular during the 1980s — with HP and Sun featuring prominently in this market segment — Commodore was more known for its home computers, which probably played a major role in C900 development being cancelled. At the time Commodore was also in the process of acquiring Amiga, with the C900 perhaps unsurprisingly featuring similar design language as the Amiga 2000.
Perhaps ironically, the Z8000 CPU that features in the C900 had a bit of a tragic history as well. Although featuring a range of interesting features, such as the ability to use its registers as 8-, 16-, 32- or even 64-bit registers by combining them as needed. Although this and the general performance of the Z8000 made it a solid CPU, it could not compete against the Motorola 68000 and Intel 8086/8088 CPUs when those appeared on the market.
In the video, [Neil] takes us through a detailed history of the C900, its feature list and the hardware inside the C900 prototype he got his hands on. It’s a fascinating glimpse at a part of Commodore history where this company almost went toe to toe with Sun, HP and other workstation giants.
Love that he starts out with the engineers names: Shiraz Shivji, Frank Hughes and Bob Russel.
It’s kind of funny how UNIX has been viewed as the savior over the years. Even Apple (A/UX) at one time had it.
“Even Apple at one time”
MacOS is Unix.
System 1-7 and Mac OS 8/9 weren’t, though.
The first ~15 years of the Macintosh were pretty much *nix free.
Then it got addicted. Poor thing. š
Jobs’ fault. He did a little too much BSD in college.
And why not? About 1981 there were articles about how Unix wss the operating system to seek. There was Microsoft Xenix, which was Unix. Mark Williams cloned it for Coherent. Various high end systems had Unix. About 1990 there was Minix, and the ever popular Xinix. And soon 386BSD and the ever popular Linux.
I used Microware OS-9 starting in 1984, which was “Unix-like”. Almost forty years ago. I couldn’t afford Unix or something like it until 2001.
You didn’t miss out, don’t worry. Unix was meant as joke by university students. Unfortunately, the others didn’t get that and took Unix serious. It’s like one if those hilarious scenes out of “The Hitchhiker’s Guide to the Galaxy”. Let’s take it with humor. It’s the only way to fight Unix.
Microware OS-9 was actually a real-time operating system, not just simply a multitasking operating system like Unix. :)
I know because I programmed a lot of CD-i titles in my second job, and the CD-i ran on OS-9. :)
It was kind of a way of legitimizing your computer. It might LOOK like it’s only good for secretaries or for playing games, but if we port Unix to it, it’ll be taken seriously as a Real Computer too!
I don’t think that ploy every really worked in practice.
Bil, did you have any interaction with that team while building the 128 at all?
Yes, I was interviewed by Russell and Hughs, and hired by Shivji. Our offices were within feet of each other, at first in the MOS building in Norristown/KOP and then in West Chester.
I’m really intrigued by the decision to mount the HDD controller right next to the HDD itself, and then cable the system bus up to it. Seems to presage the development of ATA/IDE, which wouldn’t come along until ’86 as I understand it.
It’s a bit like ESDI, too. ESDI moved a part of the controller intelligence to the HDD.
Btw, the old name of ATA/IDE was AT-Bus at the time (“AT-Bus HDD”).
The AT bus was what we now refer to as ISA (ISA bus).
The terms “fixed-disk” or “Winchester drive” was also used for HDDs.
They were called Winchester drives because the first one, at least the first popular one had 30 megabytes fixed and 30 removable.
30-30. Get it?
They were called Winchester because they were invented by IBM Hursley House in Winchester, UK
Beyond the circumstantial evidence that Hursley House is near (not in) Winchester, UK, there’s not much I can find to support this conclusion. The first hard disk drive (IBM 350 RAMAC) was made by IBM in California in 1956. The first storage product produced at Hursley seems like the IBM 5444 (code named Dolphin) in 1969, after IBM had already made many other HDD products. The IBM 3340 (code named Winchester) was made in 1973 in San Jose. It featured a new read-write head technology that was much lower cost and could remain parked on the platters even when spun down. This technology becoming the new standard was what spread the name “Winchester” to subsequent HDDs. The code name was said to come from the Winchester rifle due to the initial 30/30 storage configuration.
This was the standard I remember in the computers before the PC. The controller was the same size as the Hard drive (for 5 1/4″) and was usually mounted in the case. This also kept the cable short. I have a WD controller that supported up to 4 ST-504 (ST-506?) interface drives. I used a similar Xebec controller in systems like the Gimix Ghost running OS-9 Level II.
Loved my Timex Sinclair 1000, C64, and C128
Is it possible to dump the rom(s) and create an emulator?
http://www.zimmers.net/anonftp/pub/cbm/firmware/computers/c900/index.html
This machine is a primary example of Commodore’s failure. Trying to have a machine that suits everyone rather than staying focused on developing their own ecosystem. After the 64, they simply could not find a focus of their own. “let’s do another 8 bit with the 128…let’s do a 16 bit, not backwards compatible graphics machine.. Let’s do some pc clones…let’s do a Unix-like machine…”. The only thing most of these machines had in common was the company logo/name.
Too true, once Jack Tramiel left there wasn’t any real direction.
His management style might have hindered things.
https://en.wikipedia.org/wiki/Jack_Tramiel
That might be true enough but at least it was leadership. :)
As a German, I better make no comments about leadership here. š
Bought Jupiter Ace.
Reason: Using 8080 figForth at Sandia National Laboratories.
c/C++ defeated basic/forth software technologies in ~1991.
We are back?
https://www.prosefights2.org/irp2020/p022623/julka.mp3
I was in West Chester at the time this was being developed and spent some time with the team. I got to know Frank quite well. The whole thing was generally chaotic, there were several teams and we all seemed to be doing our own thing. Jack had just left and there was discussions going on about Amiga. Bil was working on the TED there were meetings about a MODEM but they were all independent and no thoughts about comparability. Having said that, I don’t think anyone thought much about that back then.