[JulianMorrison]

IIRC, Be was not technically superior. It was a one-user no-security pre-internet OS which had been highly polished in for one task: concurrent multimedia. It would have require a lot of work to bring up to scratch as a graphical design platform to rival NeXT (which was always about ultra-precision, and even used PostScript to draw its GUI).

[gecko]

This is very true. I like to think that, in a lot of ways, Be was what Mac OS 8 (the original Mac OS 8) should have been: a massive, massive improvement on the design of System 7, while retaining the general feel of the operating system. NEXTSTEP and OPENSTEP were always more about being a vast improvement on Unix. Yes, Be had some POSIX emulation, but there's a huge difference between that, and actually integrating properly with a complex multiuser Unix network with shared directories, site-wide logins, and so on--something that NEXTSTEP had had time to hone to perfection due to its heavy use in academia and government.

BeOS is still interesting to me, but mostly in a historical sense. I have to agree that purchasing NeXT was a vastly wiser decision.

EDIT: As a quick addendum to my post: although this has changed in the intervening years, the one thing that Be definitely did have going for it back then was that its kernel was technically superior, delivering mostly the same benefits of a true microkernel design, but with vastly superior performance to Mach. I have no idea what the current state of things would be on that front; between Apple merging BSD into the kernel space and replacing DriverKit (Objective-C) with IOKit (eC++), the differences are probably minimal.

[wmf]

BeOS delivered no benefits of microkernel design; in fact its userspace network stack was clearly worse than competing kernel network stacks.