RE: Winblows

Joe Kiniry (kiniry@cs.caltech.edu)
Sun, 30 Mar 1997 13:48:49 -0800


Joe Barrera writes:

> I don't know whether you object to NT because it's Microsoft or
> because it's not Unix. If the reason is the latter, you might check
> out http://www.softway.com/opennt/home.htm. I've heard good things
> about them. Their stuff is a native NT subsystem, like the
> Microsoft-provided Posix subsystem (but much, much better).

i'm glad to see someone has moved in the open standards direction on
top of the NT kernel. not a bad thing, but i'd rather not have to
fork out $200 a pop to run the dozens and dozens of standards apps
that i use everyday. of course, only _four_ have been ported to
OpenNT...

don't get me wrong, i have not had/used an NT 4.0 system as my desktop
station for more than twenty or thirty hours (NT 3.1 is another
matter...). i'll form a precise opinion after my move to a new system
and learning every in and out of the entire microsoft product line and
system internals in the very near future now.

i do this for several reasons:

o so that i have extremely firm ground on which to stand when
discussing microsoft's newest technologies (rather than several year
old knowledge of internals, "i used to write windows apps", and hiring
other people to integrate systems into my networks for me, etc.).
firsthand guts-level knowledge is always best in my book...

o to increase my global systems knowledge as a user, developer, and a
computer scientist (not necessarily in that order)

o and to increase my value as a professional

in the end, what i object to wrt microsoft, centers more on their
business, marketing, technology _practices_. but i don't think we
need go down that path; it is entirely to exhausting and unproductive
for us all.

> > good luck rohit. just think, NT 5.0 might be decent since it is
> > hitting that magical microsoft third release plateau. microsoft
> > _is_ promising to solve all of the world's distributed computing
> > problems. we're just not getting _anything_ done today and it
> > _is_ the renaissance of DC, didn't you hear?
>
> What's wrong with NT 4.0, or, why does it not qualify as "decent"?

again, my short-term problem is figuring out how to integrate a host
of extremely powerful and useful machines into my existing network
without throwing out the old completely. this means that i have to
integrate file and print services (how much do i have to pay extra for
NFS?, development across platforms (version control, compiler issues,
etc. - how much extra do i have to pay here?), user support (how about
a real emacs port? - how do user identifiers map, if at all? - how
much do i have to pay for X windows? - what about integrating mail
products?), etc.

i have a sinking fear that it's going to turn out that these "free"
systems are as expensive as if i boutht their UNIX counterparts,
especially if one includes the cost of software _and_ support.

per my use of the derogatory "decent"; i'll hold back judgement until
the end of summer once i've covered the bases per my earlier
"personal-growth" moves.

> BTW I'd count NT 4.0 as the magical third release, after NT 3.1 and NT
> 3.51. (Yeah, there was 3.5 but it doesn't count.)

you are right joe, i should have placed higher emphasis on the NT
3.1/3.51 distinction.

but, in terms of semi-revolutionary advances in the product line, as
opposed to (evolutionary) tremendous leaps forward in stability and
usability (which is how i viewed the 3.1->3.5->3.51 at the time that i
had such a system on my desk), you'd have to admit that NT 5.0 is the
third release.

joe