Unix on microcomputers
This page is not about Unix itself - plenty of good sites cover every aspect of Unix, Linux and all that is associated with it. Rather, I'd just like to show some of my Unix machines and put them into the perspective of the evolution of Unix through the 80s and 90s. For many years, Unix was the promise of tomorrow and, as it turned out, it stayed a promise longer than anyone expected. Unix only became mainstream with its adoption by Apple (OS X is, of course, Unix underneath) and the adoption of Linux underneath tons of non_PC computers in the 2000s.
Comemco System Three: Cromix
The System Three originally was a CP/M, or rather, CDOS Z80 machine (see the page Early Micros). However, the S-100 bus lent itself to a lot of expansion. Cromemco pushed S-100 to its limits, and beyond, like no other company. The first stage was Z80 Cromix: pretty much the only functioning Unix-like OS that ran successfully on the Z80. It required lots of RAM, which Cromemco provided with their 256KZ and 1024KZ boards. And it required a hard disk, although the determined could run it on the System Three's extremely fast 8" Persci disk drives. The Persci, however, was not only fast in transferring data. It was also fast, very fast, in running into alignment problems. Later on, System Three's were fitted with normal-but-slower 8" drives. And then, Cromix really did need a hard disk.
Z80 Cromix was an astonishing feat, well worth Googling for further information. But Cromemco quickly developed into the technology leader among S-100 vendors. The DPU card of 1983 contained a Z80 plus a full 68000, and 68000 Cromix was a giant leap forward. Later on, the XPU offered 68010s and system Three's became some of the most powerful microcomputer systems available.
My two 68K System Three's are typical for the period around 1985: a dozen+ boards, chock full of RAM, the astonishingly fast hard drive controller Cromemco had developed and some of the last Cromix versions before the company switched to 'normal Unix'. They must have cost an absolute fortune when new... They also have a nice background story to them: they played a core role at Philips R&D labs in developing the CD player technology.
My next contact with Unix-like systems was Minix. Developed in 1987 by Andy Tanenbaum, it ran on a whole range of hardware. I tasted it on the Atari ST (which is why shocked readers are confronted with a teeny little Atari ST picture on a Unix page) and it was a very educational experience, but not a logical choice for daily work. The very interesting Minix 3 system is still in development here. Shortly afterwards, I upgraded to a Mac II and I remember struggling with Mach 10, an awful Unix version on the 68k Mac. That put me off Unix until 1992, when I got Linux on my 386 clone. From there, my interest picked up again and I collected the machines shown below.
SUN SparcStation IPC & IPX
The Sun IPC (1990) is mostly known for its cuteness. A very small box with the great promise of a proper Sparc CPU. In truth, it has the horse power of a 486, and users would be well-advised not to upgrade to newer versions of Solaris or the IPC
would slow down to a crawl. The 1991 IPX had more power, I run it under Solaris 2.5. There are a lot
of these machines still around, probably because of their 'cute factor'. They are very reliable -
but you have to know how to fix the NVRAM once the built-in battery runs out.
Sun SparcStation 10
This was the machine I used at work for a while in the mid '90s. Although it came out in 1992, SparcStation 10s were around for many, many years as Unix workhorses. Compared to the IPC/IPX, this machine still gives you the feeling of driving something powerful. It may not look so cute, and take a lot of space in your computer collection room, but it's the nicest Unix machine of its Sun generation by far.
Although NeXTs enjoy an absolute cult status, I am in a minority of one: I actually dislike the things. Maybe that is because I own a stock, basic NeXTstation (1990). It is so overloaded with just running the NeXTSTEP operating system that it cannot even run a C compiler on top. Can you believe that - a stock NeXT didn't even come with C installed. And this was supposed to be the ultimate machine for scientists? To be fair, the system would feel very different when upgraded with more memory and storage.
NeXT is famous, but the company did not build that many machines: only some 50,000 in total apparently.
I am sure that the OS is structurally beautiful. Of course, Apple bought back NeXT for it as they could not get their own next-generation opearating system going. But the GUI looks slightly clumsy nowadays: too much conceptual thinking bogs down the visual appearance just as it does performance. I dislike the NeXT to the point, where in fact, I forgot its password... ouch.
Epilogue - OS X and Linux
Unix machines remained at the fringes of microcomputing all along the 80s and 90s, with always great expectations but - frankly - disappointing delivery. With the demise of other 'non-compatible' machines (see the 16-Bitters page) it's actually surprising that it took so long for Unix to come to the fore as the main threat to Windows, which in the end it did become through Linux and Apple's OS X.
Why did it take so long and why did the top Unix brands not play much of a role in it? In the late 90s at the latest, some of the Linux distributions on PC clones were actually more attractive than either Sun or NeXTSTEP. My trusty old 386DX/40 of 1992, beefed up with RAM, provided a much nicer playing ground at the time than the boxes from the Elite Providers. True, that's a hobbyist perspective. But clearly that perspective became valid for corporate users too, only a few years later. Linux now powers everything from netbooks to supercomputing facilities at research labs. In the mean time, Apple's OS X was another way in which Unix became mainstream. Whether Linux or OS X (based on Mach) though, the roots of Unix's final ascendance are not found in the Unix versions of the big corporates. Sun's Unix came to a dead end - the same is true for the SGI, HP, and other closed-source Unices. In that sense, it's the Unix ethos that has succeeded - but none of the commercial hopefuls of the 80s delivered on their promise of mainstream Unix machines. Open Source did that!