NeXTSTEP on Pa-RISC
openpa.net
openpa.net
On the contrary, macOS is NeXTSTEP plus several years of development. It's what the NS means in NSLog.
The point that the article makes is about opening up NeXT to other hardware platforms. So while from one perspective, you might argue it lives on inside Apple, you could also argue that's where nextstep died.
In the early 2000s I worked for a company that went all in on NeXTSTEP a decade earlier. The product was developed in a "4 GL" called 4D or 4th dimension.
We had to do a painful migration to windows nt/xp because NeXTSTEP was discontinued and apple actively fought to kill attempts to fork or open source the code base.
NeXTSTEP drivers were written in Objective-C, originally OS X used C++ subset based on COM (IO Kit), now moved into userspace and called Driver Kit, in homage to the NeXTSTEP DriverKit name.
NeXTSTEP was focused on OpenGL and Renderman, OS X used OpenGL, macOS is now using Metal.
NeXTSTEP drivers were on kernel space, now everything is moving into userspace.
NeXTSTEP used Display Postscript, OS X moved into PDF subset, nowadays that is only part of the rendering stack.
NeXTSTEP had a X Windows Server as well, on macOS that is now gone.
macOS Finder is nothing like the NeXTSTEP file application.
NeXTSTEP supported a concept similar to OLE, it is nowhere to be seen on macOS.
https://fahrplan.events.ccc.de/congress/2007/Fahrplan/attach...
https://www.amazon.de/-/en/Mac-OS-Internals-Approach-paperba...
If anything it is going more into that direction, after Apple announced removing all kernel extensions, and having userspace counterparts to them.
https://developer.apple.com/documentation/SystemExtensions
https://developer.apple.com/documentation/systemextensions/i...
Usually it takes one OS version between introducing new userspace APIs, and removing the old way on the following version.
Additionally, if there is a userspace extension to an existing kernel extension, the userspace one will take precedence.
Turns out using a managed userspace is viable, if management is on board to support the development all way through.
An educational copy of OPENSTEP 4.2 was the last thing I purchased for myself from Apple since they discontinued the Newton MessagePad.... and I'm sad my Cube quit booting, and that I never got it running on my ThinkPad.
Let's hope projects like https://github.com/trunkmaster/nextspace succeed.
Jobs made a huge mistake by going with the 68K in the first place. DEC would prove just a few months after NeXT's October 1988 launch the viability of a MIPS-powered workstation.
Even better, in the long term, would have been to go with the 80386.
The difference is that 68k was ubiquitous, reasonably cheap 32bit capable platform with MMU that had huge availability of parts and made porting software easy. Sun was working with 68k partially because they chose it in 1980, a year after it was made available, and by 1986 they published SPARC ISA and shipped first systems a year later
Sun launched its first SPARC-based system more than a year before the NeXT launch in October 1988.
Sun came out of Stanford and was aware of the Stanford and Berkeley RISC architectures (the latter of which led to SPARC). NeXT had academia heritage, too, via Mach from CMU, but I guess it wasn't enough to persuade Jobs to go for a more exotic architecture than the one he was familiar with from Apple, or the "enemy" in Intel.
One can see a world in which NeXT goes with 80386 from the beginning, eventually pivots much earlier to software-only, and becomes a real rival to Microsoft and IBM in the early 1990s to provide a multitasking successor to DOS. Or, for that matter, IBM goes with NeXTSTEP (or just buys NeXT) instead of the AIM Alliance.
But the alternate computer history is interesting to imagine, could have been via x86 or PowerPC with IBM or something else. Same with Be.