From BeOS to Haiku: The Once and Future OS Contender?
The computer industry also has a number of such obvious what if scenarios. What if Bill Gates and Paul Allen hadn't lucked into the IBM contract for DOS, which was the basis for Microsoft's eventual hegemony? What if the UNIX operating system hadn't split into several minor variants, but gone on to become what Linux later became, only a decade or two before?
What if BeOS had gone on to succeed on the desktop...?
These days there are probably very few, even among the computer savvy, who have heard of the BeOS operating system. But once, not long ago, it was actually deemed to be a possible contender for the desktop. Microsoft itself took it seriously enough that in their defence in the anti-trust trial, when forced to cough up examples of credible competitors in the PC market place, they mentioned not primarily the Apple Mac (which anyway at the time ran on non-Intel hardware), any of the UNIXes, which had long split into minor mutually hostile camps, or even LINUX, which at the time was thought to be far too primitive and clunky to make any sort of bid for the mainstream PC, but BeOS.
Having used it as a court room alibi for how Windows faced real competition on the PC, Microsoft proved that they themselves at least, if no one else, partly bought into BeOS as a possible threat, by using their very monopoly power to crush the upstart operating system.
A dual 133MHz PowerPC 603 BeBox, with the
"blinkenlights" LED CPU performance meters
(Click for larger image)
But while the clunky old station wagon might serve the everyday family well, there's something to be said for the speed, elegance, and vision of the Batmobile. Getting from point A to point B is all well and good. But to do it in style...
A lot of that style came from throwing out the standard text book on how to construct an operating system. A fact that's often obscured by the razzle dazzle and eye candy of successive new versions of contemporary operating systems, is how old the basic principles and designs they're built on actually are. Windows started out as simply a Graphical User Interface, or GUI, on top of old DOS, which in turn was built on the older CPM, which got a gene splice of UNIX when merged with the Windows NT code base. LINUX is based on the its UNIX ancestors, who harks back to the late '60-early '70s phone switches. Even the shiny, lickable modern Mac OS X rests atop this comparatively ancient heritage.
Starting from scratch, and not having to worry about backwards compatibility, the people who began the development of BeOS in 1991, led by ex-Apple exec Jean-Louis Gassée, decided that they might as well really start from scratch, and reconsider a number of established conventions on how an operating system should work. The result was designed from the ground up to be a desktop OS with strong multimedia capabilities.
Its claim to fame was its close to real-time response, even on low-powered hardware. One of the strong points of the operating system was the way it efficiently distributed tasks across a number of processors if more than one were present in the machine through SMP, short for Symmetric Multiprocessing. At a time when Windows was still stuck with the FAT32 file system, which only supported files smaller than 2GB, BeOS featured its own 64bit journaling file system, BFS, that supported file sizes north of 260GB. Drivers were not installed, as such, but dynamically loaded, if present on the system on boot up if any given piece of hardware was detected. And its Graphical User Interface (GUI) was developed on the principles of clarity and a clean, uncluttered design.
The early prototypes of the so called BeBox were based on a dual processor design, using the Hobbit processor from AT&T, along with three Digital Signal Processors (DSPs). But AT&T halted production of the Hobbit in '95, forcing Be Inc. to move to the PowerPC platform, as had Apple. The problem was that Apple, moving comparatively huge quantities of machines, would always have first dibs on the new components, and profit from their economy of scale.
So the decision was made to drop the hardware part of the business altogether, and market BeOS as an alternative operating system on Apple hardware to the then seriously long in the tooth Mac OS. Around this time Apple was also desperately seeking for a replacement for their own OS. The natural course of events seemed to dictate a buy out of some sort, and BeOS becoming the basis for the new Mac OS. The investors in the company, who had never seen anything other than red inc on the corporate balance sheet, were looking at a substantial pay day.
But Gassée's Gallic pride got the better of him. He believed he had Apple, which had once sent him packing, over a barrel. "I've got Apple by the balls, and I'm going to squeeze hard until it hurts," he was quoted as saying. Apple were willing to go to US$200 million for the little OS that could. But Gassée held out for US$400 million just for the rights to use BeOS.
He might even have got it. But there was another ousted ex-Apple exec with an operating system for sale circling the tired company. It was none other than the founder of the company, Steve Jobs. Apple could pay off the arrogant Frenchman. Or it could buy Jobs' NeXTstep operating system, and get the Steve Jobs charisma and magic thrown into the bargain. They went with NeXT, which later morphed into today's Mac OS X, and Jobs, who went about clawing his way back into the top spot at Apple and the industry he enjoys today.
The clock was now ticking for Be Inc. Apple hardware, itself a limited market in the grand PC scheme of things, was now hostile territory. So Be Inc. up sticks again, and ported their OS to the dominant Intel based PC platform. In an effort to grab market share the R5 version of BeOS was even given away for free, in the form of a limited "Personal Edition" trial version, which could quite easily be installed as a fully fledged OS by users.
They also tried to get hardware makers to include the OS as an alternative to Windows. Hitachi was willing to pre-install BeOS on some of their computers as a dual boot system with Microsoft Windows. A stern talking to from Redmond, Washington put paid to that notion (an anti-trust suit was later launched because of this, amongst other shenanigans, which was quietly settled in 2003).
Even though BeOS had garnered a hard core following of enthusiasts, actual revenue was thin on the ground. Another focus shift, the last, was made. Be Inc. staked what was left of the company on that chimera of the late dot-com era, the Internet Appliance, dedicated small boxes that would connect to the internet and otherwise do bugger all. The flop in 2001 was the flip of Be Inc. into the mass grave of computing past. The remains were acquired by Palm Inc. for a song.
While Palm Inc. (later PalmSource and ACCESS) now owned the source and rights to BeOS, a German company called YellowTab, led by Bernd Korz, claimed that they had secured the rights to distribute and extend the BeOS R5.1 "Dano" version that Be Inc. had been working on at the time of their demise. They started selling the OS under the name Zeta.
But despite muted protestations to the contrary, suspicions were rife that the company did not have legal access to the source code the software was based on, and the ability and right to develop and distribute it. Statements by ACCESS made to Bits of News back in early April this year confirmed that Zeta was indeed infringing. By that time the project was all but dead anyway.
Another path for users who wanted to keep using their favourite OS on new hardware, was the community made "abandonware" distributions based on the free R5 Personal Edition, updated with new software and drivers. Foremost among them was, and is, BeOS Max.
But both of these solutions had a clear sell by date. There's just so long you can keep a closed source OS going through patches and gaffer tape. It was clear for many that if BeOS were to have a future, it must needs be through some form of of Open Source project. Some aimed to simply recreate some of the look and feel of the OS, and some APIs on top of another system, such as Linux. Most of these efforts have since run into the sand.
But others didn't want to settle, and set about recreating the entire operating system as a fully Open Source clone, first called OpenBeOS, but later renamed Haiku. Haiku's immediate goal is to recreate and extend upon Be Inc's latest released version of BeOS, version R5. This is at all possible due to the very modular nature of BeOS, where individual pieces can be redone, and slotted into a still working system. The plan is that, having recreated, and improved upon, the R5 version of BeOS, they'd then chart a future for the OS
Progress for the small team has been slow, with some observers having written the project off as quixotic. But Haiku is now in a usable state, though it is still very alpha and hardly stable enough for everyday use. That the small development team, some of whom will be getting together with enthusiasts at the FalterCon community gathering in Sunnyvale, California on Saturday, August 11, has made it thus far is an amazing feat of endurance, the only one that really counts in the end.
For sheer love of the work, and a vision of elegant computing, they have worked for years, largely unsung, with little more than a handful of contributions, and now the proceeds from the sales of Haiku merchandise, to keep the project going. But there can now be little doubt that they will succeed. BeOS, in the shape of Haiku, will have a future. What that future will look like, is anyone's guess.