NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/unix/misc/4672

Internet Message Format | 1992-12-22 | 5.6 KB

Path: sparky!uunet!gatech!emory!europa.asd.contel.com!howland.reston.ans.net!zaphod.mps.ohio-state.edu!moe.ksu.ksu.edu!cis.ksu.edu!jfy
From: [e-mail] (Joseph F. Young)
Newsgroups: comp.unix.misc
Subject: Re: Windows NT Report
Date: 22 Dec 92 17:55:14 GMT
Organization: Kansas State University
Lines: 104
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]> <[e-mail]>
NNTP-Posting-Host: crucis.cis.ksu.edu

[e-mail] (Jesse W. Asher) writes:

>In article <[e-mail]>, [e-mail] (Raymond Shwake) wrote the following:
>>
>>    Based on their comments, and comparing anticipated functionality
>>and pricing, I must report with some anxiety that Unix on the Desktop could
>>be in trouble. Two of those colleagues are long-time UNIX practitioners in
>>both the development and the integration/support areas. They're now proposing

>I'm also a Unix bigot who is advocating (with others) moving to Windows
>NT.  I have no problem doing this for several reasons:

>1)  Price.  Unix appllications and the OS itself have long been over
>priced.  Both NT and applications will be very cheap compared to those
>for Unix.  Just the other day I heard about a PPP implementation that
>was being sold for Unix for $795.  What a joke.

This is a general problem of the workstation market; software price
structures for the majority of UNIX products are still stuck in the
old mainframe/mini price structures.  The laws of supply and demand
haven't pushed the prices of software down as far in the workstation
market as they have in the PC market.

>2)  Range of applications.  We will finally be able to run all those
>nice dos applications out on the market that have not run worth a hoot
>under unix.  Not only dos, but also windows apps.  And porting to NT is
>not difficult so a wealth of public domain software will still be
>available.

Microsoft has said that they will make sure the top 100 Windows 3.1
applications work under NT.  This doesn't sound like you are going to
be able put your old DOS and Windows 3.x applications on it and expect
that they will run.  So, in any case, people are going to have to buy
new versions of their software to run under NT.

>3)  Ease of programability.  Windows is much easier to problem in than
>X.  Even though it doesn't have all the advantages of X, we can live
>without them for the benefits we gain.  Not only that, but there are X
>servers coming out for NT for those needing X.  But we finally have a
>choice.

One thing to remember: X is not UNIX.  As far as the programmability
of X versus Windows goes, I've heard complaints about both.  There are
a number of toolkits, such as SUIT, that allow you to write window
system independent code and ignore the details of the windowing system.

>4)  Ease of usability.  Nt should be a lot more easier to use for the
>user that actually has to get some work done.  Sharing disks and
>printers is no longer something the user has to speed inordinate amounts
>of time learning.  Users shouldn't have to spend time learning how to
>run the OS just to get it to do something.  They should be spending
>their time getting their work done.

No matter what GUI you put on it and no matter how many layers of
software you put on top of it, networking is going to give the user
problems.  Is Microsoft going to support all combinations of Novell,
Apple Filesharing, NFS, Decnet, TCP/IP, etc. that a user might wish to
integrate with?

>5) Ease of administration.  This relates to
#4 in that with NT the user >can do many things that the sysadmin had
to previously.  Disk sharing, >printer sharing, and so on are
extremely easy.  You will not believe how >easy until you see it.

I've fought with a number of "user-friendly" setup programs.  They are
nice when they work, but when you want to do something that doesn't
quite fit into what is "expected", they make your life miserable.

>6)  Enhancements in OS.  The OS handles some things better than Unix.
>Better scheduling and able to handle 256kb off a serial port (that right
>- 256kb).  They improved what needed to be improved in Unix long ago.

"Better" scheduling?  In what respect?

The problem with UNIX and serial ports is most vendors hadn't
bothered tuning their drivers to go beyond 9600 bps or so, especially
since most of the common serial hardware couldn't handle anything
faster.

>Basically, you will have all the power of unix without a lot of the
>hassle.  This is how unix should have been long ago.  Now I too am
>predicting the death of Unix for workstations (PC based).  Windows NT is
>just too much better than Unix for workstations.  This is _not_ a
>prediction of the death of Unix.  Only people that don't know anything
>about Unix predict such nonsense.  Unix is still a great multi-user
>system with many advantages in other areas.  But it is in big trouble
>for workstations.

I've seen losts of hype about Windows NT, but I've seen little of
substance.  Where is the committment of software vendors to produce NT
software?   Where is the commitment of hardware vendors to support NT
(I don't mean the lip service DEC and SGI have given to it)?
NT still has a ways to go, if it is going to succeed.

>Besides, NT IS Unix with Windows on top of it.  It is just put out by MS
>and doesn't have an AT+T copyright.

Yeah right.  A Mach-based microkernel and a minimally functional POSIX
interface does not make NT a UNIX OS.
--
Joseph Young, Systems Programmer
KSU Department of Computing and Information Sciences
Manhattan, Kansas 66506  FAX: [phone]  Phone: [phone]
Internet: [e-mail]  UUCP: rutgers!depot!jfy