Re: Why?! [way offtopic]
Hamish Moffatt wrote:
> On Tue, Dec 22, 1998 at 03:29:37PM -0500, Ed Cogburn wrote:
> > Considering what I hear from this group and the 'debates' in c.o.l.a.,
> > and elsewhere on the Net, WinNT isn't that stable as it is now. If they
> > really are going to break compatibility with their predecessor OSs then
> > that will change things, but I'm astonished that they would actually do
> > that. After all, the whole world didn't rush out and upgrade from Win95
> > to Win98, and this upgrade was supposed to be compatible. A lot of
> > people have decided Win95 is enough. If they leave compatibility out,
> > M$ runs a risk of having Win2k not being a big seller. Its the
> > compatibility issue that has allowed them to drag everybody along the
> > upgrade path, a real money maker, all this time.
> Windows NT 4 is reasonably stable in that the whole operating
> system rarely crashes; it's still more often than Linux,
> but a lot less often than 95. The GUI still crashes on occasion.
> To me, the main reason for not upgrading to 98 is that there aren't
> any new features which I want. You can get the new desktop
> by putting IE 4 on 95 (and I don't want it anyway). I don't use USB
> and you can get that on 95 anyway (my notebook has it). So there's no reason
> to upgrade, since I rarely use 95 anyway.
> NT could do with some upgrades, though -- it needs to become more
> mainstream to get better driver support. It needs DirectX 5 or better
> (then I'd never need to use 95 at all on my desktop).
Not having used NT, I don't have any first hand knowledge about the frequency
of crashes. There seems to be a wide diversity of opinion about this. Its
probably depends on a number of things, like hardware, and maybe more
importantly, how many apps are installed and used (and perhaps whether the
apps are MS or not?). With a small number of well behaved apps, with
'approved' hardware you could get NT to appear stable.
> > Finally, one the important things about Linux is its configurability,
> > especially being able to trim the kernel to just what is needed. The
> > idea that they can build an OS that is not only good at being a
> > mainstream desktop OS while at the same time as being a strong
> > enterprise server, all from the same shrink-wrapped box, smells real
> > funny to me.
> Isn't that what we do too?
> And RedHat?
In a sense, you got me here; I didn't explain myself very well. Part of the
problem is my personal tendency to think of Linux as a server-type OS, that
just happens to have a significant number of people using it as a desktop OS.
Sure, Redhat is trying to make Linux easier to install and use for the
non-tech crowd. The difference is what we will and won't do to make Linux
work for the masses.
Nothing is wrong with GUI type programs that make installing and Linux easier
for the uninitiated, and I won't argue against some additions into the kernel
to provide support for graphics like the fb code in the development kernel,
and kernel additions suggested by the GGI project. However, the developers of
Linux are not going to try shoe-horning the Mozilla code into the linux kernel
the way MS shoe-horned IE into Win98 just to destroy a competitor.
Linux is also much more modular than W2K will ever be (IMHO, of course).
With Linux, you can strip both the distribution and the kernel of unnecessary
software leaving only the server software installed, for a server type job.
Linux also doesn't force a particular interface on you. If you like the CLI,
then you aren't forced to live with X. This won't happen with W2K, you'll
have a GUI, and software like IE forced on you. Because these things are
integrated into the OS, the OS will become even more complex and bloated than
either NT or 98 is now.
I think MS is about to hit the point of diminishing returns with W2K. There
is only so much you can stuff into an OS before the OS loses the most
important characteristic an OS should have: rock solid stability. The
growing concern over W2K's delay seems to suggest that they now have an OS so
complex that stability is almost impossible.
> Hamish Moffatt VK3TYD firstname.lastname@example.org, email@example.com
> Latest Debian packages at ftp://ftp.rising.com.au/pub/hamish. PGP#EFA6B9D5
> CCs of replies from mailing lists are welcome. http://hamish.home.ml.org
> Unsubscribe? mail -s unsubscribe firstname.lastname@example.org < /dev/null