How come Linux replaced Unix? What happened to proprietary Unix?

[liam_on_linux]

The personal histories involved are highly relevant and they are one of the things that get forgotten in boring grey corporate histories.

Bill Gates didn't get lucky: he got a leg up from mum & dad, and was nasty and rapacious and fast, and clawed his way to industry dominance. On the way he climbed over Gary Kildall of Digital Research and largely obliterated DR.

 

Ray Noorda of Novell was the big boss of the flourishing Mormon software industry of Utah. (Another big Utah company was WordPerfect.)

Several of them were in the Canopy Group:

https://en.wikipedia.org/wiki/Canopy_Group

Ray Noorda owned the whole lot, via NFT Ventures Inc., which stood for "Noorda Family Trust".

https://en.wikipedia.org/wiki/Ray_Noorda

Caldera acquired the Unix business from SCO, as my current employers reported a quarter of a century ago:

 

https://www.theregister.com/2000/08/02/caldera_goes_unix_with_sco/

Noorda managed to surf Gates's and Microsoft's wave. Novell made servers, with their own proprietary OS, and workstations, with their own OS, and the network. As Microsoft s/w on IBM-compatible PCs became dominant, Novell strategically killed off first its workstations and pivoted to cards for PCs and clients for DOS. Then it ported its server OS to PC servers, and killed its server hardware. Then it was strong and secure and safe for a while, growing fat on the booming PC business.

But Noorda knew damned well that Gates resented anyone else making good money of DOS systems. In the late 1980s, when DR no longer mattered, MS screwed IBM because IBM fumbled OS/2. MS got lucky with Windows 3.

MS help screw DEC and headhunted DEC's head OS man Dave Cutler and his core team and gave him the leftovers of the IBM divorce: "Portable OS/2", the CPU-independent version. Cutler turned Portable OS/2 into what he had planned to turn DEC VMS into: a cross-platform Unix killer. It ended up being renamed "OS/2 NT" and then "Windows NT".

Noorda knew it was just a matter of time 'til MS had a Netware-killer. He was right. So, he figured 2 things would help Novell adapt: embrace the TCP/IP network standard, and Unix.

And Novell had cash.

So, Novell bought Unix and did a slightly Netwarified Unix: UnixWare.

He also spied that the free Unix clone Linux would be big and he spun off a side-business to make a Linux-based Windows killer, codenamed "Corsair" -- a fast-moving pirate ship.

Corsair became Caldera and Caldera OpenLinux. The early version was expensive and had a proprietary desktop, but it also had a licensed version of SUN WABI). Before WINE worked, Caldera OpenLinux could run Windows apps.

Caldera also bought the rump of DR so it also had a good solid DOS as well: DR-DOS.

Then Caldera were the first corporate Linux to adopt the new FOSS desktop, KDE. I got a copy of Caldera OpenLinux with KDE from them. Without a commercial desktop it was both cheaper and better than the earlier version. WABI couldn't run much but it could run the core apps of MS Office, which was what mattered.

So, low end workstation, Novell DOS; high end workstation, Caldera OpenLinux (able to connect to Novell servers, and run DOS and Windows apps); legacy servers, Netware; new open-standards app servers, UnixWare.

Every level of the MS stack, Novell had an alternative. Server, network protocol, network client/server, low end workstation, high end workstation.

Well, it didn't work out. Commercial Unix was dying; UnixWare flopped. Linux was killing it. So Caldera snapped up the dying PC Unix vendor, SCO, and renamed itself "SCO Group", and now that its corporate ally, the also-Noorda-owned-and-backed Novell owned the Unix source code, SCO Group tried to kill Linux by showing it was based on stolen Unix code, and later when that failed, that it contained stolen Unix code.

Caldera decided DOS wasn't worth having and open sourced it. (I have a physical copy from them.) Lots of people were interested. It realised DOS was still worth money, reverse course and made the next version non-FOSS again. It also offered me a job. I said no. I like drinking beer. Utah is dry.

The whole sorry saga of the SCO Group and the Unix lawsuits was because Ray Noorda wanted to outdo Bill Gates.

Sadly Noorda got Alzheimer's. The managers who took over tried to back away, but bits of Noorda's extended empire started attacking things which other bits had been trying to exploit. It also shows the danger and power of names.

Now the vague recollection in the industry seems to be "SCO was bad".

No: SCO were good guys and SCO Xenix was great. It wasn't even x86-only: an early version ran on the Apple Lisa, alongside 2 others.
 

The SCO Group went evil. SCO was fine. SCO != SCO Group.

 

Caldera was an attempt to bring Linux up to a level where it could compete with Windows, and it was a good product. It was the first desktop Linux I ran as my main desktop OS for a while. 

Only one company both owned and sold a UNIX™ and had invested heavily in Linux and had the money to fight the SCO Group: IBM.

IBM set its lawyers on the SCO Group lawsuit and it collapsed.

Xinuos salvaged the tiny residual revenues to be had from the SCO and Novell Unixware product lines.

Who owns the Unix source code? Microfocus, because it owns Novell.

Who sells actual Unix? Xinuos.

Who owns the trademark? The Open Group. "POSIX" (a name coined by Richard Stallman) became UNIX™.

Who owns Bell Labs? AT&T spin off Lucent, later bought by Alcatel, later bought by Nokia.

Was Linux stolen? No.

Does anyone care now? No.

Did anyone ever care? No, only Ray Noorda with a determined attempt to out-Microsoft Microsoft, which failed. 

Comments

[history_monk]

I saw the collapse of commercial Unix from the view of a software provider. When I started my current job in 1995, we provided software for OSF/64 on Alpha, HP-UX on PA-RISC, Irix on MIPS, Solaris on SPARC and AIX on POWER. Also Windows NT on x86-32 and Alpha, and VMS on VAX and Alpha. Everything was 32-bit except for OSF/64.

The only commercial Unix we added subsequently was HP-UX on Itanium. That was outlived, for us, by the PA-RISC version. We did 64-bit versions for both HP-UXes, Irix, Solaris and AIX.

The thing that killed the commercial RISC platforms was the lower cost and broad compatibility of PC-derived hardware, achieved through economy of scale. Once AMD introduced x86-64, the RISCs' days were numbered and without needing many fingers. The low price of Windows NT and Linux compared to the commercial Unixes meant that none of them were worth porting to x86-64.

Incidentally, SCO Xenix carried a Microsoft albatross. When MS sold Xenix to Real SCO, there was a clause that required it to remain capable of running applications for the then-current version. Forever. Since it was 286-based at the time this greatly handicapped its development. That's why people I know who had to port an X.400 package to SCO Xenix called it "ScumOS" well before the Caldera nonsense started.

We first did Linux in 1999 with Red Hat Linux 6 on x86-32, did 64-bit quite early, and added ARM64 more recently.

[liam_on_linux]

Thanks for this! I saw it at the time but forgot to comment.

Do you have any kind of a link for the SCO thing? This is news to me, and I deployed quite a few SCO boxes...

[history_monk]

I don't have a link, but I can ask the relevant people.