By that time, OS/2 Warp 3 was plugging along nicely, gaining ground in large and stable industries like banking, insurance and telecommunications. It powered tens of thousands of ATMs across the world throughout the 1990s and well into the new millennium. It kept accounting and auditing companies running.
But somehow, it failed to create a buzz among consumer-level software developers. They were spitting out Windows programs, which OS/2 Warp ran like a pro, but many people failed to see the advantage of getting Warp when Windows was pre-installed on their PCs.
OS/2 soldiered valiantly on until IBM pulled the plug at the end of 2001 and withdrew support five years later. We may not see it at work when we pull cash out of the money machines anymore, but those of us who liked it still have the box on our shelves for old time's sake.
What NeXTStep?
By 1989, the brave new world of windows, icons and menus was getting a bit stale. Then Steve Jobs came along with the NeXT Computer, and we took a collective intake of breath so deep that our ears popped from the loss of air pressure.
NeXT hardware -- the original NeXT Computer, a.k.a. "the Cube," and its younger brother the NeXTstation -- was black, sleek and beautiful. The machines' gray-scale displays were so subtle and clear that we could get up close and stare at them without hurting our eyes.
And the operating system, called NeXTStep, was frankly exciting. Its graphical interface was built around Display PostScript, so it was sharp and scalable. Underneath, it was built on a solid structure of Unix, including a Mach kernel and BSD code. And for the developers, it had an object-oriented application layer and tool kit. This made it much easier to code for than other platforms.
NeXT hardware didn't take off as meteorically as Jobs had hoped, but it did find a place in higher education and academia. In fact, it was a favorite at a Swiss research facility called CERN, where an English researcher named Tim Berners-Lee used NeXT products to develop a little project of his called the World Wide Web. NeXTStep has earned its place in the stars on the strength of that alone.
NeXT's sluggish hardware sales meant that applications developed for this cool platform had fewer computers to run on. So the company focused its attention on developing a cross-platform operating system. This is how NeXTStep was reborn for the ages.
In collaboration with Sun, NeXT turned its NeXT-branded operating system into OpenStep, which could run on Sun Solaris systems and other hardware. OpenStep's spec was made public in 1994; this development became the linchpin of a 1996 deal that brought Steve Jobs back to Apple. OpenStep was the model for Apple's impressive new operating system, when the lurching old Mac OS classic gave way to Mac OS X.
And when folks at Stanford Linear Accelerator Center wanted to port their NeXT applications to another hardware platform, NeXTStep was re-reborn as GNUstep. Instead of rewriting the applications, they rewrote the NeXTStep object layer, which they laid on top of Unix code and glued together with X Window. Presto! A more open OpenStep than OpenStep.
Whatever will be will BeOS
In 1991, when Apple released its PowerPC reference platform so hardware vendors could make Mac clones, one company had another idea. Be Inc. decided to port its own operating system, BeOS, to the Mac platform.
Perhaps Be was anticipating that Apple would never quite deliver on the promise of Copland, its Holy Grail of a next-gen operating system, and that it might want to buy Be's off-the-shelf alternative instead. Perhaps Be was just trying to find a market for the operating system it had designed for a failing product line. Whatever the company's motives, BeOS became one of personal computing's favorite near-miss stories.
In 1990, an ex-Apple exec, Jean-Louis Gassée, founded Be Inc. to develop a new computing platform starring BeOS and a machine called the BeBox. But the AT&T Hobbit processors at the core of the BeBox were discontinued, so Be had to redevelop the platform to run on PowerPC processors.
When cash flow dictated that the company couldn't market its own hardware anymore, it retooled BeOS to run on other companies' PowerPC and Pentium platforms. The fact that this multithreaded, media-friendly OS could run multiple videos without a stutter or crash on clunky old Pentium IIs wowed many digital media developers and enthusiasts.
Sadly, Be didn't capture a lot of money. BeOS did attract interest from Apple in the mid '90s, but its price tag didn't. Be was firm on its asking price, Apple was firm on its offer, and the difference had a lot of zeros after it -- at which point Apple cozied up to NeXT and its OS instead.
And so Be shot for the moon, missed and was eventually sold to Palm Inc., in 2001. Palm halted development on the platform, and it died.
However, Be enthusiasts kept the beast alive online at sites like BeBits.com. After Palm abandoned them, they began to improve their favorite OS in a series of reverse-engineered open-source projects unofficially called OpenBeOS. On Linux or BSD kernels, they built Be-compatible APIs and gave the results winking names like Blue-Eyed OS.
Palm wasn't keen on trademark infringement, so the Be fan community picked a name that wouldn't offend: Haiku. And it's in this guise that BeOS lives on. Without actually being Be, Haiku certainly seems to be Be -- and with the real thing on ice in Palm's vaults, that's as good as it's going to get.
The Spirit of '95
Oh, we know what you're thinking: The Windows 95 phenomenon was a lot of fuss to make over a steppingstone between 16- and 32-bit computing. The technical aspects of the thing were washed away in a marketing tsunami -- and on the subject of flooding, it cost more to develop than 1995's other bloated headline-grabber, the Kevin Costner film Waterworld.
But we appreciated Windows 95 back then, and we still think of it fondly. It enabled people on home PCs to name their files with something more flexible than an eight-character name and a three-character extension.
And it was the first time Microsoft had given consumers a graphical operating system with a decent foundation. Up till then, mainstream (that is, non-NT) Windows was just an operating environment -- an easy-to-navigate structure that was built on stilts over the wet-sand footing of DOS. The whole structure had a nasty habit of collapsing right before you'd clicked the Save button. Before 95, Windows really was the dog that ate everyone's homework.
Oh, we still had our gripes, and Windows 95 certainly didn't solve them all. We had been promised no more UAEs (unavoidable application errors), and we were annoyed to discover that something that looked and quacked like a UAE was still a regular guest -- but now it was called a GPF (general protection fault).
And it would be two more revs of Windows before we could make reliable use of those spiffy little USB slots that were beginning to appear.
But Windows 95 was a turning point in the world of Windows, and it brought us where we are today. Of course, there have been a few missteps along the way (Me? Vista?), but perhaps Windows 10 -- I mean Windows 7 -- will open new vistas for us.
Forgotten but not gone: X marks the spot
We know... the X Window System, or X Window for short -- or just plain X for shortest -- is not actually an operating system. But its creators started out with a manifesto, so for that reason alone, we can't ignore it.
While IBM and Microsoft and Apple were conducting parallel revolutions out in the marketplace in 1984, MIT boffins Bob Scheifler and Jim Gettys were crafting a work of philosophy: Cut away complexity. Don't get bogged down with every cool idea you can bolt onto your system. Leave the actual user interface to the user. Just make it work.
Consider these Ben Franklin-style nuggets of wisdom, laid out by Scheifler and Gettys:
"Do not add new functionality unless you know of some real application that will require it."
"If you can get 90 percent of the desired effect for 10 percent of the work, use the simpler solution."
"If a problem is not completely understood, it is probably best to provide no solution at all."
X ended up doing exactly what it set out to do: make a Unix operating system kernel and the user interface work together. So it's surprising we don't hear a whole lot about it anymore.
Or perhaps it isn't so surprising. The pace of development hasn't exactly whizzed by recently: In its first four years, X went through 11 iterations. In the following 21 years, it slid glacially up to release 11.7.4.
But don't think that X has actually gone: It's just lurking beneath the surface. And it lurks everywhere -- most famously beneath all of the free Unix and Linux releases, and the Panther, Tiger and Leopard branches of the Mac OS X family. And long may it continue!
Many thanks to the Personal Computer Museum, GUIdebook, Amiga Future and Wikimedia Commons for providing images for this story.
Next: Timeline: 40 years of OS milestones
Freelance writer Matt Lake's first operating system came on a 5¼-inch floppy disk. His most recent would only fit on a dual-layer DVD. His biggest regret is that he missed out on the cassette-tape computing era, because that level of propeller-headedness would really impress his colleagues. Computerworld is an InfoWorld affiliate.
This story, "10 operating systems the world left behind" was originally published by Computerworld.