But 23 years behind OS-9, which I happily ran on my CoCo 9 years before NT was born.
https://en.wikipedia.org/wiki/OS-9
https://colorcomputerarchive.com/repo/Documents/Books/OS-9%2...
The Motorola 6809 was a nice architecture for UNIX-like systems... if we accept the 16-bit memory address space.
Firstly, NeXTStep came out in 1989, and had both preemptive multitasking and threads, and OS X 1.0 didn't really improve on NeXTStep in that area at all.
Secondly, while OS/2 1.0 had threads and preemption, it was text-only and had to run on IBM AT 80286 PCs so most of that technological might was wasted. It would be 1990 before it had a real GUI. And it would be nearly 1992 before OS/2 was 32-bit.
Presentation Manager was available in 1988, by the way, not 1990. Petzold's book on it came out in 1989. First (and only) edition sitting on my bookshelf as I type this.
This is just every OS when a new version comes out... sad!
WinXP needed 64MB RAM and was considered hugely bloated at the time.
EDIT: Actually, I just checked, and NT 3.5 opens in around 8 minutes and uses 20MiB of swap, when booted with 12MB.
It was marketed and sold as a business OS. I agree with the Op (OPs?) that it was revolutionary.
This remark points you don't realize how bloated NT truly was.
In the 2000s I could still run Linux _and XFree86_ in under 4MB of RAM.
NT in 1993 would just BSOD with less than 10MB of RAM.
The parent poster claimed the bloat was because of portability and other metrics, but this is also not true since other very much portable OSes from the next decade still had lesser minimum requirements.
NT was not such a good design as it is often claimed. It was a clone of a esoteric big iron OS that they were trying to retrofit on much smaller devices. They had to backpedal on everything from microkernel to personalities in order to be able to actually market it to "their target market".
I genuinely don’t understand all this looking down on what, in hindsight, clearly worked.
NT was always marketed and priced for the professional & enterprise markets, so the whole “but what about home computers at the time” rings hollow.
NT ushered in a new era of business computing and networking, and more importantly a new era of thinking about and working with technology in the workplace, something its competitors at the time couldn’t manage. Even Novell netware, dearly missed, got steamrolled by NT. You can pooh-pooh NT all you like, and talk about what a shitty clone it was, and how poor the technology was.
NT clearly won, at the time.
No, I am not. I used to do that with at least 2.2 kernels. https://tldp.org/HOWTO/4mb-Laptops.html
But besides, I know 4MB in 2000s was low enough to be awkward. The point is that _it could_. NT in 93 could not even boot with 8MB RAM, and then it was definitely not low.
This seems to confirm my statement.
On the same machine, NT 3.5 BSODs. With 12MB is the first I can get it to boot, albeit it complains that some services failed to start, and generally thrashes like hell, using 20MB of swap...
You can even try muLinux yourself, albeit it's kernel 2.0, not 2.2, and it opens some X11 utils that bump swap usage to 2MB (not very usable).
There’s no way NT “almost killed Microsoft” because Windows 3.x was a money printer. And if that hadn’t happened, Microsoft wouldn’t have exited the OS/2 relationship.
Windows NT was originally meant to be OS/2 3.0. In the worst-case scenario they could have stuck with that plan and played the role IBM had designed for them, which would also have been lucrative (although nothing like the bonanza of the desktop monopoly they had by the late 1990s).
I don't think 2K could do DOS games, as they usually talked to the hardware.
The stability was amazing, but there were still some compatibility issues with games and so on.
[0] https://www.itprotoday.com/compute-engines/death-alpha-nt
[1] https://techmonitor.ai/technology/dec_forced_microsoft_into_...
Using all that money to push PRISM/ALPHA and MICA into workstation and server and large server space would have yields them better results. That said, I'm not sure if DEC could have survived with that strategy either.
What it would have done is make it so that Windows NT ends up not existing or existing in a very, very, very different way, likely in a far worse way. Lets remember that Gates didn't even believe in Virtual Memory.