stak,
IF, NT 5.0 isn't ready for prime time as soon as it is released,
when would it be ready for the enterprise market to really be
considered bulletproof
I can't answer that question. High-quality O/S's take years to
develop & test. Working bugs out is a MAJOR undertaking & MSFT
is not used to supporting enterprise customers. It's out of
their league right now. They would have to rely on CPQ or other
OEM's to take care of the support which opens up another can
of worms.
IA-64 is a qualitatively different architecture & requires a
complete re-write from 32-bits. MSFT could probably do it, but
until their 32-bit OS is on a level to compete with Linux, there's
no point considering it. I was involved in HP's effort to port
HPUX to IA-64 and I can tell you that it is quite an undertaking.
Unix has the advantage of age: it has been tested & used by thousands
of propeller-heads since the early days (80's) so it is very, very
stable. I, personally would take FreeBSD over any MSFT O/S, any
day of the week, if for no other reason than the stability.
My sense is that MSFT is really an applications software company.
They think in terms of features & nice presentation graphics etc...
They have never even come up to the level of Novell in file server
technology, let alone enterprise servers.
Why would a product mgr. allow a NEW O/S project grow to 30+ million
lines of code?? Bulletproof usually means small, tight, and
very powerful at the same time. They're crazy if they think that
they need all that code to be as powerful as Linux, which is about
1/15th that size. What are they doing, anyway?? Maybe marketing
has taken the place of common sense.
MSFT's marketing has always been application-oriented. They're
probably trying to pollute the NT code with hundreds of features
that will sell this O/S to the unwitting, untrained, and not-
real-bright customer. Talking paper clips, cute graphics that
make it "easy-to-use" etal... I can't say I really know exactly
what they have in mind, but if NT 4 is any indicator, it is
going to be too big, too slow, no RAS, poorly scaled, and not
ready for anywhere but the workgroup.
BTW, I wouldn't be surprised if MSFT began announcing that they
are porting some of their applications to Unix (Linux probably)
in the near future. They're smart enough to hedge their bets
in an increasingly competitive market.
cheers,
cherylw |
