“Unix is a junk operating system designed by a committee of PhDs.” -Dave Cutler (2004)
blog.codinghorror.com
blog.codinghorror.com
Is that so? Yet, no version of Windows has run on anything big endian.
Linus Torvalds started Linux in a way that was totally geared toward 386. He used the 80368 memory segments of all things: the descriptor tables and all!
Some of the 386-isms survived in some of the naming, like get_fs (named after the FS register):
https://unix.stackexchange.com/questions/182631/what-is-fs-s...
None of that stopped Linux from running on all sorts of stuff.
> The widely popular Windows NT 4.0 release carried initial support for Alpha, MIPS and PowerPC in addition to the Intel x86 platform
Dec Alpha is little-endian (little-endian only, I believe).
MIPS and PowerPC are bi-endian (possibly depending on exact model? Maybe some can be switched, whereas some are fixed?)
Windows NT ran on PowerPC in little endian mode. Hey look, A Reymond Chen article from 2018 mentions this: https://devblogs.microsoft.com/oldnewthing/20180806-00/?p=99...
The Pocket PC (later Windows CE, later Windows Mobile) were used on MIPS quite a bit. I did some Windows CE work on MIPS units. It was all little endian. I don't think NT on MIPS was any different.
By the way, I visited the Microsoft Exchange building in Redmond some time in 1999 or 2000. I was shown a room full of Alpha servers. :)
- They were still very logically (and somewhat elegantly) designed, well architected and portable.
- They were amazingly stable (compared to the rest of the Windows family, that is).
That's not saying much.
"Thompson received a Bachelor of Science in 1965 and a master's degree in 1966, both in electrical engineering and computer science, from the University of California, Berkeley, where his master's thesis advisor was Elwyn Berlekamp."
"[Dennis Ritchie] graduated from Harvard University with degrees in physics and applied mathematics."
I think that an operating system where you can't rename or delete a file that is open by some process is idiotic beyond words, even without MS-DOS drive letter names thrown in.
To work around the problem by providing a function to schedule a file rename at the next reboot is ironically cynical: you expect the system to crash often enough that this is viable.
This is not an inherent limitation of the VMS or NT architecture; I'd rather argue it's for compatibility.
> even without MS-DOS drive letter names thrown in.
This is also legacy carried over from DOS and OS/2.
If you're incorporating uncool quirks from crappy previous operating systems into your new one, don't criticize Unix.
Oh, and don't use its implementation language, and related concepts and techniques.
The link gives HTTP 302, but doesn't load from the new location I guess?
We can only wish, alas!