Trivia Tuesday: Xenix - Microsoft's forgotten OS

You would never guess it, but in the 1980s, Microsoft was king of the Unix world. That's right; while juggling MS-DOS, Windows, and a plethora of other products, Microsoft somehow found time to develop the most popular version of Unix.

Licensed from AT&T and based on Seventh Edition Unix, Microsoft started development of their very first operating system in 1979, just months after moving into new digs in Bellevue, Washington. Back then, AT&T wasn't licensing the Unix brand itself, only the source code, so Microsoft had to give their new OS its own branding. Xenix, as the project came to be known, was Microsoft's true operating system.

From 1979 to 1986, when they moved to Redmond, Washington, Microsoft was based out of Bellevue, Washington. Before that, they were based out of Albuquerque, New Mexico

In the days before MS-DOS (which was really just a last minute solution, but we’ll get to that in a minute), Xenix was Microsoft’s play for the OS market. Even though it came first, Xenix was probably a more powerful OS than MS-DOS, although both had their own advantages. Still, it was still a pretty amazing feat that Microsoft managed to get a real Unix experience running on a microcomputer, since it’d originally been designed for larger, multiuser systems. They even bolstered Unix 7’s feature set by including a few features from BSD.

In 1981 IBM was looking for an OS to power the Personal Computer. As it turned out, IBM had no desire to cozy up with the recently split AT&T, so Xenix was out the window. Microsoft hurriedly ran out and bought 86-DOS, the ‘Quick-and-Dirty Operating System,’ from Seattle Computer Products, and managed to convince IBM to use it in their new product. It was an off the cuff move that would have stunning repercussions for the industry.

IBM also considered powering the Personal Computer with an in-house operating system developed for their own IBM 801 chipset

The IBM PC could've been the ultimate Unix workstation for the consumer and enterprise market. With Microsoft's roadmap still pointing towards a Xenix-dominated future, the IBM PC could've been their Trojan horse, a well put together device that would set the stage for everything to come. Unfortunately for Unix fans, it turned out being the bullet that set Xenix on the path to a long, slow death.

Chances are you know the basics of what went down; the IBM PC launched a revolution, and ended up setting the standard for what was to come. Part of that standard was its MS-DOS operating system. This pretty much locked Xenix out of a huge chunk of the market. Still, Microsoft kept it alive well into the graphical interface era, releasing a port for the Apple Lisa, and launching Version 2.0 in 1985.

Check it out - Microsoft's using Apple in an ad

By ’87, the writing was on the wall, and Microsoft decided it was time to dump the aging OS. They entered into an agreement with SCO, a Xenix partner since the early days, and traded them the full rights to the OS in exchange for a 25% stake in SCO. They carried the Xenix torch into the 21st century in the form of SCO Unix, before filing for bankruptcy in 2007. UnXis, Inc. bought the rights to the assets and intellectual property in 2011, recreating SCO as TSG Group. Unfortunately, back in August they filed for bankruptcy again, stating that “There is no reasonable chance of rehabilitation,” so it looks like Xenix’s journey may finally be reaching its end, although we imagine very little of Xenix actually lives on in recent products.

Long after the sun had set on Microsoft's Xenix, the venerable OS was still the backbone of their internal operations well into 1990s.  As one former employee put it:

"I think the original DOS might have been developed on one of their old VAX mini's but by the time I got there everything including DOS 2.x, all their languages and applications, Mac Word and Mac Excel, Windows Excel and Windows Word were written in vi and compiled on those goddamn Xenix boxes, and all their documentation was written in vi and compiled in troff and nroff. I don't think [they] really moved to the PC platform for development until around the time Windows 3.1 came out."

If that says one thing, it's that Microsoft was pretty darn flexible in those days. Not only were they actively developing for three different platforms (Mac, MS-DOS, and Windows), but they were still hanging on to Xenix, and even using it as the primary driver of their operation! Makes you wonder if Bill Gates has a soft spot for Unix.

The ultimate minimalism

Even though it’s pretty obscure nowadays, don’t underestimate the influence of Xenix. It helped Microsoft cut their teeth on OS development, and somewhere deep inside Windows to this day, maybe there’s a little part of it still alive, as Bill Gates’ remarks at the 1996 Unix Expo hint:

"Well, Microsoft stepped back and looked at that situation and said that the best thing for us might be to start from scratch: build a new system, focus on having a lot of the great things about Unix, a lot of the great things about Windows, and also being a file-sharing server that would have the same kind of performance that, up until that point, had been unique to Novell's Netware. And through Windows NT, you can see it throughout the design. In a weak sense, it is a form of Unix."

Images via Makeuseof, Wikipedia, and OS/2 Museum

Previous Trivia Tuesday articles
Got ideas for future Trivia Tuesday articles? Let us know in the comments section below!

Report a problem with article
Previous Story

Nokia sending Lumia 920 to some consumers in limited edition packaging

Next Story

Characters from The Hobbit set to star in Windows Phone 8 ads

13 Comments

For a second there I almost forgot about xenix, I wonder what the PC world would be like today if Microsoft didn't split into DOS.

Pretty sure I remember reading that either MS or Bill Gates invest a sizeable amount of money in to Linux OSes('s?)(s?)
I can't remember which one(s) though

Nice article. I still have a copy of this here somewhere; I remember playing around with it on an old IBM PC/XT many years ago but I never had any software to run on it. I wonder if the disks are still readable.

Makes you wonder if Bill Gates has a soft spot for Unix.

I have no recollection of where I heard this, nor do I have any source, but I swear I once recall someone quoting Gates as suggesting that DOS was a "lovely collection of compromises".

I was always amazed at how powerful Xenix was - it was the first real operating system I worked with. Prior to that I had only seen the BBC Micro and a couple of different Commodore computers at school and home.
This was in the mid 90s so DOS and Windows were around (Win 95 was just around the corner).
Although it didn't have a GUI it always felt like I was using a proper computer.
Gone but not forgotten...

Good day. And here's a little contribution from someone who was actually there at the time.

Something to ponder: Xenix was the first ever 16-bit operating system for microprocessors. Up until then there was CP/M and Apple's funky DOS and little else. While the world was marvelling at the 8-bit 8080s, Z80s and 6502s, Microsoft ported the venerable Unix to 68000, Z8000 and a couple of others that have over time slipped through the grey matter - but all 16-bit chips.

Next thing you know, IBM shows up at the door wanting an OS. So why didn't they use Xenix? Why did Microsoft have to go out and buy QDOS and make it MS-DOS? Why not just use Xenix? Great OS. Well ahead of Berkeley by a couple of years. And it worked.

Actually there's a very simple answer. Intel couldn't make enough 16-bit chips to meet the expected demand. In fact, if you recall, the first IBM PC used the 8088 processor - a 16 bit chip on an 8-bit bus. Intel couldn't produce enough 8086 (16 bit on a 16 bit bus) chips to come anywhere near the expected demand for the IBM PC in late 1981. It wasn't until the next year, next generation PC that IBM moved to a 16-bit processor and bus.

If Intel had been able to meet the demand for 16-bit chips in 1981, Xenix might likely have been the first PC operating system. If that had happened we would today have no need for Linux ever to be developed and the software world would be a totally different place.

Yet another example of how Intel's incompetence changed the world. Likewise today, Microsoft cannot release the Surface Pro until after Xmas because of Intel's inability to produce enough chips to meet expected demand.

Some things never change.

Major Plonquer said,
Good day. And here's a little contribution from someone who was actually there at the time.

Something to ponder: Xenix was the first ever 16-bit operating system for microprocessors. Up until then there was CP/M and Apple's funky DOS and little else. While the world was marvelling at the 8-bit 8080s, Z80s and 6502s, Microsoft ported the venerable Unix to 68000, Z8000 and a couple of others that have over time slipped through the grey matter - but all 16-bit chips.

Next thing you know, IBM shows up at the door wanting an OS. So why didn't they use Xenix? Why did Microsoft have to go out and buy QDOS and make it MS-DOS? Why not just use Xenix? Great OS. Well ahead of Berkeley by a couple of years. And it worked.

Actually there's a very simple answer. Intel couldn't make enough 16-bit chips to meet the expected demand. In fact, if you recall, the first IBM PC used the 8088 processor - a 16 bit chip on an 8-bit bus. Intel couldn't produce enough 8086 (16 bit on a 16 bit bus) chips to come anywhere near the expected demand for the IBM PC in late 1981. It wasn't until the next year, next generation PC that IBM moved to a 16-bit processor and bus.

If Intel had been able to meet the demand for 16-bit chips in 1981, Xenix might likely have been the first PC operating system. If that had happened we would today have no need for Linux ever to be developed and the software world would be a totally different place.

Yet another example of how Intel's incompetence changed the world. Likewise today, Microsoft cannot release the Surface Pro until after Xmas because of Intel's inability to produce enough chips to meet expected demand.

Some things never change.

Awesome stuff

Major Plonquer said,
Good day. And here's a little contribution from someone who was actually there at the time.

Something to ponder: Xenix was the first ever 16-bit operating system for microprocessors. Up until then there was CP/M and Apple's funky DOS and little else. While the world was marvelling at the 8-bit 8080s, Z80s and 6502s, Microsoft ported the venerable Unix to 68000, Z8000 and a couple of others that have over time slipped through the grey matter - but all 16-bit chips.

Next thing you know, IBM shows up at the door wanting an OS. So why didn't they use Xenix? Why did Microsoft have to go out and buy QDOS and make it MS-DOS? Why not just use Xenix? Great OS. Well ahead of Berkeley by a couple of years. And it worked.

Actually there's a very simple answer. Intel couldn't make enough 16-bit chips to meet the expected demand. In fact, if you recall, the first IBM PC used the 8088 processor - a 16 bit chip on an 8-bit bus. Intel couldn't produce enough 8086 (16 bit on a 16 bit bus) chips to come anywhere near the expected demand for the IBM PC in late 1981. It wasn't until the next year, next generation PC that IBM moved to a 16-bit processor and bus.

If Intel had been able to meet the demand for 16-bit chips in 1981, Xenix might likely have been the first PC operating system. If that had happened we would today have no need for Linux ever to be developed and the software world would be a totally different place.

Yet another example of how Intel's incompetence changed the world. Likewise today, Microsoft cannot release the Surface Pro until after Xmas because of Intel's inability to produce enough chips to meet expected demand.

Some things never change.

Except...

Xenix ran on 8088 and the reason DOS won was performance.

The Unix overhead of Xenix made it sluggish on 8088 and 8086 in comparison to DOS and for general consumption even in the 80286 era it was still slower than DOS and OS/2.

From someone else that was 'there'...

PS...

If that had happened we would today have no need for Linux ever to be developed and the software world would be a totally different place

Technically there is no reason we have Linux or that any mainstream OS is using a *nix model. It is in fact rather insane that a model with massive limitations is still even in use.

In the late 80s there was a lot of work in new kernel and OS model theories, and from all this brainstorming, the only truly new OS that emerged and not caught in the Unix minimalism was Windows NT. (Object based OS model, new kernel created from theory.)

It was during this time that everyone was caught up in OS performance and universal compatibility and minimalism was the foundation of virtually every OS project, which took everyone back to the simplistic and generic Unix concepts.

This locked in mindset is why everyone is still putting duck tape on Linux and XNU/OSX, instead of actually working from a richer OS model that doesn't need all the work around and 'new level' of complexity to exist in today's far more complex hardware and software demanding world.

Edited by thenetavenger, Nov 20 2012, 6:31pm :

thenetavenger said,

Except...

Xenix ran on 8088 and the reason DOS won was performance.

The Unix overhead of Xenix made it sluggish on 8088 and 8086 in comparison to DOS and for general consumption even in the 80286 era it was still slower than DOS and OS/2.

Technically there is no reason we have Linux or that any mainstream OS is using a *nix model. It is in fact rather insane that a model with massive limitations is still even in use.

Performance not withstanding (I don't know enough about that to compare them - it takes actual usage to really compare that), I think Unix was still a far more powerful OS than MS-DOS. I'm not saying that to bash DOS, I do have some experience and fond feelings for it, but it was definitely lacking in features.

I'm curious, though, what sort of 'limitations' Unix-like OSes have/had? I deal with NT and Unix based devices pretty regularly, and I haven't really hit any brick walls that Unix-likes can't handle. A lot of high performance supercomputers seem to favor *ix systems, too.

Makes you wonder if Bill Gates has a soft spot for Unix.

Let me answer that for you... Nope.

When developing OS/2 with IBM, Microsoft had the option of using Xenix and evolving it into a more robust graphical platform, and they kept hitting limitations in initial development, that they did not want OS/2 to have. Which is a rather sad threshold, as the original OS/2 model was rather simplistic.

Going forward...

There is a reason the *nix model was rejected in the Windows NT project, and the majority of their focus and goals was to avoid the pitfalls of the *nix model and create an OS without the limitation of *nix. Microsoft truly gave Cutler and the NT Team the option of starting with any foundation or model they wanted, and the one thing they avoided was *nix limitations.

This is why having XNU/OSX and Linux be the primary competitors to Windows NT 20 years later is beyond insane, as the *nix OS model wasn't seen as crude and limited to Microsoft in 1990/1991, and it still is in comparison to the NT OS model and architecture (which is an Object Based model that is far more portable and extensible than virtually every *nix by leaps and bounds.)


Seriously, if you are going to write about Microsoft's history, you should at least go find a 1st Edition of "Inside NT", you can find it on Amazon used. Especially if you are going to speculate on something you can find fairly well documented in books and magazines.

Commenting is disabled on this article.