BeOS: Has-Been or Will-Be?

For an upstart operating system, BeOS hit a lot of marks: It outran Mac OS and Windows on their own hardware, supported modern operating system features before anyone else did, piqued Apple's interest as the replacement for Mac OS, ideologically succeeded the Amiga platform, started a rivalry with embedded operating system QNX, and did all of the above while booting in under thirty seconds.

Such innovation does not always equal success, however, and Be, Inc. was liquidated by late '02, its intellectual property being swallowed up by Palm (and now held by PalmSource, Inc.). The last release of the BeOS was in 2000, and in five years the computer industry has leapt ahead, while Be fans remain in limbo on an increasingly outdated system with no hope for reprisal.

Or is there?

Today there are three efforts to maintain and extend the BeOS experience on modern hardware, each with a different story but all holding the same basic goal: Keeping BeOS alive in a post-Be, Inc. world. And with three very different approaches, each of the BeOS projects is worth taking a look at, if not downloading and installing. Be fans, get ready to restart your engines.

Zeta

German yellowTab's Zeta is based directly on the BeOS 5 source code, as well as source to unreleased updates such as BONE (BeOS Network Environment) and BeOS 6, all inherited directly from Be, Inc. As such, it starts out of the gate with a working product and an easy ride to future updates and patches. Zeta is actually shipping at the moment despite still being in development. So what's to see?

Sadly, not much. Zeta is still doing work to bring BeOS 5 forward to compete with Windows and Mac OS X — bringing an operating system back to life after five years is a trying task. For instance, only in recent builds have SSE and HyperThreading been recognized, and both still remain un-optimized for in the operating system. Likewise there is a 2GB RAM limit and no FireWire support.

Zeta should hit a final version this Summer, and while BeOS fans would find this intriguing, there is one huge catch with Zeta: It comes only in German. Until the operating system returns to its roots in English, this son-of-a-BeOS will remain a German attraction only. Efforts to port the interface to English have begun, but aren't reported to be ready until sometime after Zeta 1.0's release.

Haiku

Haiku, once known as OpenBeOS, is an Open Source effort to first recreate BeOS 5 functionality (Haiku R1) and then to extend the experience through updates to the operating system's core functionality (R2 and beyond). Haiku has attracted many developers and fans, mostly desperate BeOS refugees, though numbers mean little — Haiku is far from being without problems.

Due to its Open Source development model, the same one the Linux kernel is based on, Haiku progress is slow and plagued with internal political struggles. The project hasn't fully reached beta yet, with most of its components still in alpha or development stages. To put that in perspective, it's taking five years to recreate a five-year old operating system. Haiku, upon completion, will be ten-year old technology.

The project claims that through various community channels many part of the system are already complete. Haiku uses NewOS, a kernel similar to BeOS's created by a former Be, Inc. engineer. There are also Open Source versions of the BeOS interface as well, the project site notes. Without the working middleware, however, the kernel and GUI are useless.

Another bottleneck in the Haiku roadmap is that without hardware partners, the project is relegated to old garage-sale systems. For instance, the main development system is a dual-Pentium Pro system over-clocked to 233 MHz with 512 MiB RAM. That's a ten-fold decrease in power and bandwidth from current Pentium 4 and Pentium M-based systems and another way Haiku will be ten years behind.

Overall, Haiku is an interesting hobbyist project with implications for clean-room implementation and reverse-engineering projects, but is of little use to anyone in the real world, even deprived BeOS zealots hungry for an answer to their operating system woes. It's worth keeping an eye on if you're part of the Be world but otherwise is best left to its own developers and sycophants.

BNX

Perhaps the most ambitious of the post-Be, Inc. BeOS projects, BNX was a project aimed at mating Open Source implementations of Be operating system services to a Linux kernel. Work began in late '02 when it was clear neither Be, Inc. nor Palm, Inc. would ever continue work on a proper desktop BeOS. In the last three years it has attracted a few dozen developers but not much else.

Progress on BNX has all but halted, perhaps due to the profound differences in Be and Linux kernel architectures. Changing Be's services to match a Linux kernel effectively took the wind out of the project, and Benix was left stalled. Today the web site hasn't been updated since December '04 and the CVS remains just as outdated. Reportedly, several BNX developers have joined the Haiku project.

End-All & Be-All

If you're in the BeOS world for the long haul, you have quite a trek ahead of you. None of the above three BeOS replacements are ready for daily work, and only Zeta is near being a replacement for five-year old technology—if you know German. BeOS fans might be better off sticking with BeOS 5 and grabbing the last few updates Be released. A sorry state, but probably the stablest bet.