Are you conflating "Unix" with "Horrible windowing toolkits designed to
run on an experimental window system that should have been put out to
pasture a decade and a half ago... at the latest"?
Hmmm. I wonder why it hasn't been put out to pasture.
Gnome Is Not Unix
KDE Is Not Unix
X11 Is Not Unix
SDL Is Not Unix
Motif Is Not Unix
CDE Is Not Unix
Openstep Is Not Unix
Etcetera...
Etcetera...
Etcetera...
Sounds like a god to me. "God is not the trees. God is not sky. Etc."
There are two definitions I know:
* Unix is the kernel
* Unix is what I see when I run a Unix box today (the shells, X Windows,
emacs/vi...)
I like the second definition because the first is useless for me. In particular,
there are no alternatives to X, and in general I don't think it's my job to
build a system around a kernel, just like it isn't my job to assemble
motherboards. When I have to kill all my programs because of something built
into the installation, "the system crashed". Period.
Anyway, I think the most similar thing in Unix and Windows NT is the kernel. The
difference is at the higher layers: Unix blows the hardest when it does GUI, and
Windows blows the hardest when it does networking.
You can always use the first definition to excuse almost any blunder in a Unix
system. You can also use it to dismiss things Unix fans like (such as shell
scripts - the shell is not the kernel just like X isn't).
Anything that exports Windows file system semantics is broken by design.
I don't care about this line of flames too much, but why is "exporting" any
other file system semantics any better?
Not just "no", but "hell no". Jesus. We had PDP-11/70s running for weeks at
a time with 30-40 concurrent users and load averages peaking near 80 during
finals week. Despite the DEC FSE cleaning the swap disk platters with spit
(yes, I watched him do this).
Well, there are different rumors. I have to conclude that it depends on your
hardware & software configuration.
If you pick your Linux machine poorly enough, that may be true. It seems to
me that Windows 2000 is not actually as stable as Windows NT 3.51 was, though,
and neither of them are anywhere near the stability of a good Unix system.
Four digit uptimes (that is, 3+ years) are not remarkable for Unix.
uptime is nothing. I've seen totally wedged Unix systems avoiding reboots. It's
sport for Unix admins. However all useful processes were killed at these times
of crisis. For me it's equivalent to a BSOD.
Most of my Unix systems don't have X installed, and at least half the
Windows systems I've ever installed would have been better off if I could
have left out GDI (which was (theoretically) possible back before NT4 came
along and tossed out one of the few good ideas in NT and turned it into
a pure Windows emulation platform). X is not Unix.
I guess you run servers. AFAIK Unix is the best thing for that today, since it's
implementation of the network protocols is the de-facto standard and is very mature.
As I've said above, at levels higher than the kernel Unix and Windows part ways,
each blowing most at it's own domain. So in practice today Unix should run on
servers and Windows should run on desktops (in theory both, being parts of
lock-in strategies, should run in hell).
I'll flame along with you about C++. What does C++ have to do with anything?
This thread started as a reply to a message in another subthread where I claimed
C# was better than C++ in several aspects. So I'll counter with:
"I'll flame along with you about 'portable system interfaces'. What do they have
to do with anything?"
Not that I mind reading about the hateful subject.
One good thing about hating C++ is the amount of competent people who'll flame
along, and the small amount of them who'll come up with excuses.
