I first encountered UNIX in 1978 at Caltech, we used it in Ed Stone's lab. Then I went to work as a research scientist at Bell Labs where UNIX was invented. I know Ken Thompson and Dennis Ritchie and Rob Pike and most of the people who invented UNIX, I wrote device drivers, tinkered with the kernel, pretty much knew V6 like the back of my hand. I spent a few years at Princeton University's comp sci department, and then became a researcher at Microsoft Research until 1999. Although I personally am a researcher, I've come to believe that the engineering culture (systems engineering I mean) is really where important invention happens, not in the culture of academia and not in the culture of amateur hackers.
The philosophy of UNIX, back then, was to keep things simple. But simplicity it hard, and most hackers don't actually think like that at all. UNIX was no modular, it was monolithic, the kernel did everything and even installing a device driver meant modifying tables and recompiling the kernel. That was true until probalby the early 1990s or last 1980s, when the UNIX community started to learn about DLLs and how to split things up into updateable moduals with well defined interfaces. They learned that from DOS and the Mac.
NT was written in 1989 and represents a well engineered OS done by an engineering culture (Dave Cutler in particular) who understood requirements. NT made up of a number of modules, Kernel32, User32 and gdi being the major ones. It's API interface was cleaner and more powerful than UNIX, and it had a DDI for linking to hardware devices , which UNIX didn't have at that time.
As for crashing, that is bull****. My latest machine has been running XP and 7 for a couple years now, and it has never crashed. PCs that crash all the time have defective hardware or have been trashed by their users, or people are just saying this because they remember Windows 95 crashing.
UNIX today, Linux particularly, is a disappointment. It is by no means simple anymore. Some talented people work on Linux, but many sloppy people just add things they think are cool with no requirements. It's bloated, poorly documented, and there is no architectural vision.
Incidently, saying that Soviet products were superior is hilarious. They couldn't even make solid state electronics that worked. In 1973, they launched four probes to Mars (Mars-4 - Mars-7) and they all failed because of defective electronics. Even in the 1980s, spacecrafts contained a mix of solid state, vacuum tube and electromechanical logic and control systems. YOu had some clever people, but you had no opportunity to make anything happen unless you were a politician and could convince the State burocracy to support your idea. I'm fascinated by the Soviet space program, but I have no illusions that communism worked.
I believe capitalism provides two important forces. First, it rewards talent and doesn't reward mediocrity (ideally at least). And second, the market judges products and differentiates what people want and what works well from stuff that they don't want or that doesn't work. Communism and movements like open source, replace all this with a system where propaganda mostly determines the success of a product, zealous and often bogus claims about the quality of something that was created by a politically correct means and zealous denounciation of products produced by politically incorrect means. All the while, the denounced product is reverse engineered (open office, gimp, the LInux shells...all based on ideas gleened from Apple and Microsoft and Adobe products).
