In my opinion BeOS was technically superior in the sense that it was far more pervasively multithreaded, which meant UIs never locked up just because they were waiting on some slow computation or network resource. I still believe it's the case that BeOS on circa-2000 hardware was more responsive than any UI I ever used before or since, including MacOS X on 2017 hardware. Even just resizing a Finder window in 2017 on my MacBook lags the cursor in a way that never happened on BeOS.
Using Objective-C is just like C++ in regards to shooting yourself with the C subset common to both languages.
Foundation did not have that much support for heavy multi-processing. At least not at the same level as OS X.
It was all about UNIX programming model + Mach ports and tasks.
If you want to read about it, this is one of the few places where the documentation is still available online.
I grant that NeXT would have been more difficult for develop for when you did require parallelism for running compute intensive tasks on multiprocessor systems, but this was not the NeXT target market. BeOS got the ability to do parallelism by choosing the more fragile concurrency model, and apps paid for it in all cases. NeXT didn't do parallelism in the GUI apps, and was easier to develop for as a result.
Oh, it did. The system was the same as Unix, only difference is you got logged in as root, but it was certainly there.
(Developed BeOS applications in the late 1990s.)
This might in part be because the compositor which syncs screen updates at 60Hz (or whatever your monitor uses) while the mouse cursor is drawn on top of that via the GPU.
It is one of the reasons i really dislike compositors, they make everything feel sluggish. The only exception i know of is XFCE's compositor which has an option to not do any form of vsync so you get immediate updates.
(although TBH i do not use compositors at all since i do not see the point, considering they are mainly used for frilly stuff like translucent windows with shadows that i personally avoid anyway)
Not true at all, the BeOS ui locked up plenty when using any real apps that is. GoBE office suite would slow BeOS and its ui to a crawl, Mozilla would too, a few IDEs that were around did too. Trying to do anything beyond dail-up networking was a joke. There are a few other apps that did slow BeOS as well but they escape me as I haven't really used BeOS in 15 years.
BeOS was no more than a toy OS, I lost interest in it because Windows 2000, FreeBSD and Linux were more stable and just as fast on 2000-2004 era hardware.
That certainly doesn't match my experience: some apps like Mozilla could block their own window contents (but not window manager operations like moving or resizing) but the overall UI never blocked for anything except a video driver crash – and then only for a couple seconds as the kernel restarted that process.
As an example, on a dual P-90 I could simultaneously stream video to disk from a FireWire camcorder (which was actually a real-time challenge on that minimally-buffered hardware which Win98 & OS/2 couldn't meet except with a completely idle system), compile Mozilla (which, IIRC, took about an hour), and not even notice the impact on my email and web browsing activity. Aggregate disk transfer performance would be slower but a good real-time scheduler meant that you didn't have huge lags while reading every small file.
Windows and macOS still fail that test unless run with SSDs, and then only when e.g. Windows Update or the Xcode updates aren't running in the background.
On the other hand, networking lagged far behind and while you could get into arguments about respective API quality there's no question NextStep was more comprehensive.
Also, according to Jean-Louis Gassée, an Apple engineer tried to buy BeOS from Palm for the then-unnamed smartphone project, as a third candidate between Fadell's vision of a souped-up Pixo--the iPod OS--and Forstall's vision of a cut-down OS X. Palm turned the engineer down. (Gassée didn't name the engineer, but it's hard not to suspect it was Dominic Giampaolo, who created BeFS and went on to create Apple's Spotlight and then, unsurprisingly, work on APFS.)
Be seemed to implode with that failure and by the time Palm bought them I doubt they had many good people left. Palm certainly didn't have the vision to invest significantly and I doubt many companies would have bet on a partnership with two failures.