Windows NT?

6 views
Skip to first unread message

J Eric Bracken

unread,
Jan 10, 1994, 3:47:26 PM1/10/94
to

Just to get a little dialogue going:

What does Windows NT have to offer that (most) UNIX's don't?

How can UNIX "beat" NT in the marketplace (remembering that this
includes a lot of naive PC users...)?

--Eric Bracken (bra...@accord.ece.cmu.edu)

Bret Orsburn

unread,
Jan 10, 1994, 5:18:13 PM1/10/94
to
In article <BRACKEN.94...@accord.ece.cmu.edu> bra...@accord.ece.cmu.edu (J Eric Bracken) writes:
>
>
>Just to get a little dialogue going:
>
> What does Windows NT have to offer that (most) UNIX's don't?

Windows NT was conceived in an era when a graphical user interface is assumed.

Much of what UNIX has contributed to the state of the art evolved in an era
when a very clever tty-based interface was important: pipes, redirection,
symbolic device independence, powerful shells--even gnu emacs.

Take a look at the old classic "The UNIX Programming Environment" by
Kernighan and Pike, and ask yourself: would you still recommend that
book to someone starting out today? (Somebody swiped my copy several
years ago, and I couldn't bring myself to replace it--too much of it
now is just historical curiosity.)

Even if UNIX survives (and it undoubtedly will, in some form), much of
what old-timers would recognize as the UNIX computing paradigm will have
lapsed into history.

BTW: I just bought a brand new UNIX box, so don't flame me too much.
I did make it a 486, though, because things are changing fast, and my
486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
have I left any out?

---

Bret Orsburn Disclaimer: Speaking for no one but myself.
bors...@mcs.kent.edu

Steve DuChene

unread,
Jan 10, 1994, 8:11:19 PM1/10/94
to
Bret Orsburn (bors...@mcs.kent.edu) wrote:

....stuff deleted .....

: BTW: I just bought a brand new UNIX box, so don't flame me too much.


: I did make it a 486, though, because things are changing fast, and my
: 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
: have I left any out?

Yes, the best PC Unix of them all!!! Linux!!!

: ---

: Bret Orsburn Disclaimer: Speaking for no one but myself.
: bors...@mcs.kent.edu

--
| sduc...@cis.ysu.edu or | Life is short, --- __ o __~o __ o
| s001...@cc.ysu.edu | ride like ---- _`\<, _`\<, _`\<,
| Steven A. DuChene -| the wind. --- ( )/( ) ( )/( ) ( )/( )
| Youngstown State University | Computer Science / Math / Mech. Eng.

Dave Jacobowitz

unread,
Jan 10, 1994, 8:47:20 PM1/10/94
to
In article <2gsk75$8...@usenet.mcs.kent.edu>,

Bret Orsburn <bors...@mcs.kent.edu> wrote:
>In article <BRACKEN.94...@accord.ece.cmu.edu> bra...@accord.ece.cmu.edu (J Eric Bracken) writes:
>>
>>
>>Just to get a little dialogue going:
>>
>> What does Windows NT have to offer that (most) UNIX's don't?
>

Well, for one thing, NT does not seem to have much of a
"heritage" and it comes from Microsoft, the same companythat peddled
MSDOS for x number of years for y many millions of dollars, but
realistically typical computer users don't care. The market has shown
that pretty clearly, and really, that's not not a big surprise.
Moreover, NT can do what Unix can do - pure and simple. It's
does all the OS stuff Unix ever had and it also has
got a C compiler and we all know that's all it takes to port every
typical 'unix' utility including our favorite shells and emacs. If you
want to see Unix every day when you turn on your machine, NT can
oblige. That much is unfortunately pretty obvious.
>
So what can Unixes do to compete?

They can get a whole lot cheaper, and keep adding
'personalities' a la WABI and DOS support and MAC emulation and
whatever else. Unix still runs on the most powerful machines and it
shold be able to host all these things, and if the people who code
these features can work leanly ( something the NT coders seem unable
to do ) they can implement this stuff to run faster and smaller than
NT ever will.


my $.000000002

dave jacobowitz
dg...@virginia.edu
2


Erik Naggum

unread,
Jan 10, 1994, 9:40:31 PM1/10/94
to
[Bret Orsburn]

| Much of what UNIX has contributed to the state of the art evolved in an
| era when a very clever tty-based interface was important: pipes,
| redirection, symbolic device independence, powerful shells--even gnu
| emacs.

The shallow misunderstanding that UNIX is a user interface shall probably
be among us till the end of history. UNIX is a programming environment.
And since user interface code should always be separated from the code
implementing the functionality, anyway, those who only know about event-
driven programming will write user-interface-specific systems. If there is
a curse upon the computer industry, it is that Windows programmers think
they can discard history.

| Take a look at the old classic "The UNIX Programming Environment" by
| Kernighan and Pike, and ask yourself: would you still recommend that
| book to someone starting out today?

_Especially_ today. I also recommend Don Knuth: "The Art of Computer
Programming". Not just for history, but for understanding and appreciation
of computers per se, apart from the glitzy GUI.

Best regards,
</Erik>
--
Erik Naggum <er...@naggum.no> <SG...@ifi.uio.no> | Memento, terrigena.
ISO 8879 SGML, ISO 10744 HyTime, ISO 10646 UCS | Memento, vita brevis.

Peter da Silva

unread,
Jan 10, 1994, 9:54:28 PM1/10/94
to
In article <2gsk75$8...@usenet.mcs.kent.edu>,
Bret Orsburn <bors...@mcs.kent.edu> wrote:
> Windows NT was conceived in an era when a graphical user interface is assumed.

UNIX fits well into a graphical user interface model. The pipeline metaphor
works a LOT better than the flowcharts of most "visual" languages. It's a
pity nobody's written the visual pipeline editor.

> Take a look at the old classic "The UNIX Programming Environment" by
> Kernighan and Pike, and ask yourself: would you still recommend that
> book to someone starting out today?

I don't know... I wasn't terribly impressed by it to begin with.

I *would* recommend Software Tools by Kernighan and Plauger.
--
Peter da Silva. <pe...@sugar.neosoft.com>.
`-_-' Ja' abracas-te o teu lobo, hoje?
'U`
Looks like UNIX, Feels like UNIX, works like MVS -- IBM advertisement.

Jon Krueger

unread,
Jan 10, 1994, 9:39:22 PM1/10/94
to
Bret Orsburn writes:

>> What does Windows NT have to offer that (most) UNIX's don't?

> Windows NT was conceived in an era when a graphical user interface
> is assumed.

Yep.

Now, what does WNT do for GUI that other operating systems don't?

And do we want an operating system to do something special for
a particular kind of user interface?

If so, what user interface services/interfaces do you want
on only one operating system?

-- Jon
--
Jon Krueger j...@ingres.com
echo '[q]sa[ln0=aln256%Pln256/snlbx]sb3135071790101768542287578439snlbxq'|dc

Mike Morgan

unread,
Jan 10, 1994, 7:17:51 PM1/10/94
to
In article <BRACKEN.94...@accord.ece.cmu.edu> bra...@accord.ece.cmu.edu (J Eric Bracken) writes:
>

Well, you could just use Linux and have Unix for free. :-}
--
Michael L. Morgan | This is a Dangerous Place!
mor...@spectra.com | -King Crimson
Finger mmo...@nunic.nu.edu for PGP! |

Diego Martin Zamboni

unread,
Jan 10, 1994, 10:17:15 PM1/10/94
to
On 10 Jan 1994 22:18:13 GMT, in article <2gsk75$8...@usenet.mcs.kent.edu>, Bret Orsburn (bors...@mcs.kent.edu) wrote:
> BTW: I just bought a brand new UNIX box, so don't flame me too much.
> I did make it a 486, though, because things are changing fast, and my
> 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
> have I left any out?

Don't know if there's any other Mach-based Unix for PCs available, but just
in case: make sure it's NEXTSTEP.

--
Diego Martin Zamboni | Depto. de Administracion de Supercomputo
di...@ds5000.dgsca.unam.mx | DGSCA, UNAM, Mexico. Tel. (5)622-85-29

Daryl Biberdorf

unread,
Jan 10, 1994, 10:29:56 PM1/10/94
to
IMHO, the only hope that current variants of UNIX have versus
Windows NT or OS/2 is for the UNIX vendors to rally around a
standardized version of UNIX largely provided by a single
company. Of course, Novell recently tried to be this provider
and received nothing but sneers and jeers from the rest of the
UNIX industry (who are fools).

Right now, Windows NT is faltering (and, yes, I like Microsoft to
have to endure embarrassment once in a while). However, it has all
of the following in its favor:
1. it provides a user interface that is readily usable by a large
segment of the computer-using population (in the penetration wars,
you can't divorce OS from user interface).
2. it looks to be built as an industrial strength OS once the kinks and
bugs are out. (I still haven't seen a UNIX system that *isn't*
flakey.)
3. it will most likely not suffer the incompatibilities between
different offerings of the same "standard", largely because
it will be offered by a single provider. This is the old
proprietary versus open systems debate, but it's laughable when
different flavors of UNIX vary widely in system administration,
standard tools, etc.
4. it is offered by a company that is generally respected by
the computer using public (this excludes people like me, however)
5. NT gets to draw from everyone else's OS experience. This includes
OS/2, OpenVMS, the Big Blue mainframe OS's, *and* UNIX. UNIX is
showing its age by comparison.
6. apps, apps, apps.

Once the first year drought goes by, the apps that *need* NT will appear
quickly. Those that need a lesser OS/user environment will go for
plain Windows.

I personally think the UNIX vendors cut their own throats by refusing to
at least acknowledge the need for a leader like Novell (a company with
enough clout to have 70% of the PC network market -- a share you
CANNOT ignore).

Daryl

--
dar...@sugar.neosoft.com

Diego Martin Zamboni

unread,
Jan 10, 1994, 10:29:27 PM1/10/94
to
On 10 Jan 1994 20:47:26 GMT, in article <BRACKEN.94...@accord.ece.cmu.edu>, J Eric Bracken (bra...@accord.ece.cmu.edu) wrote:

> How can UNIX "beat" NT in the marketplace (remembering that this
> includes a lot of naive PC users...)?

Much of the stuff against Unix comes from the unfriendliness everybody
almost automatically associates with it. Many PC users don't care if
the underlying architecture is the latest object-oriented thing to come
out - they just want something that is powerful enough to allow them to
solve their problems, and do it easily and without having to expend
a year's budget on it. I think Unix makers (PC Unix makers first) should
offer their products (which, in many cases, already satisfy the ease-
of-use requirement) at lower and lower prices. I've seen many Windows
advocates become VERY impressed at seeing something like NEXTSTEP - if
NeXT could give it a mass-market price (something below $200, I think),
many Windows users would inmediately switch over.
I'm talking about NEXTSTEP because that's what I'm in touch
with (very proudly :-). But I think the same could apply to Solaris,
for example, or any other Unix you could think of.

Fritz Ganter

unread,
Jan 10, 1994, 10:45:07 PM1/10/94
to
Bret Orsburn (bors...@mcs.kent.edu) wrote:

You are kidding, aren't you?
Or forget you the nicest Unix for a PC, Linux?

: ---

: Bret Orsburn Disclaimer: Speaking for no one but myself.
: bors...@mcs.kent.edu

Fritz

--

Fritz Ganter Graz University of Technology, Austria
Email: gan...@fvkmapc02.tu-graz.ac.at, gan...@fvkmads02.tu-graz.ac.at
HAM-Radio: OE6...@OE6XYG.AUT.EU, OE6...@OE6FAD.AMPR.ORG
Phone: +43 316 873-7222 (Office), +43 316 663243 (home)
********** Linux... try it, use it, love it. ************

Fritz Ganter

unread,
Jan 10, 1994, 10:46:29 PM1/10/94
to
Steve DuChene (sduc...@gateway.cis.ysu.edu) wrote:
: Bret Orsburn (bors...@mcs.kent.edu) wrote:

: ....stuff deleted .....

: : BTW: I just bought a brand new UNIX box, so don't flame me too much.
: : I did make it a 486, though, because things are changing fast, and my
: : 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
: : have I left any out?

: Yes, the best PC Unix of them all!!! Linux!!!

^^^^^^^^^^ YES, YES!
: : ---

: : Bret Orsburn Disclaimer: Speaking for no one but myself.
: : bors...@mcs.kent.edu

: --
: | sduc...@cis.ysu.edu or | Life is short, --- __ o __~o __ o
: | s001...@cc.ysu.edu | ride like ---- _`\<, _`\<, _`\<,
: | Steven A. DuChene -| the wind. --- ( )/( ) ( )/( ) ( )/( )
: | Youngstown State University | Computer Science / Math / Mech. Eng.

--

Phil Anglin

unread,
Jan 10, 1994, 11:46:56 PM1/10/94
to
In article <2gt6fk$6...@sugar.neosoft.com>,

Daryl Biberdorf <dar...@sugar.NeoSoft.COM> wrote:
> 2. it looks to be built as an industrial strength OS once the kinks and
> bugs are out. (I still haven't seen a UNIX system that *isn't*
> flakey.)

Gee, that's funny, I've never seen a DOS machine that didn't have to be
rebooted at least once a day (though I've been told they exist, somewhere).
OTOH, most Unix systems at work only go down for power failures, hardware
failures, and OS upgrades. 3-6 months is a typical uptime here.

I think I'll wait until *after* the kinks and bugs are out to declare
it an "industrial strength OS". I kinda think the people who wanted
an "industrial strength OS" for the PC have been going to OS/2. With
OS/2 to contend with, where's NT going to find a market?

Speaking of markets, how is NT selling these days? Which hardware
platforms is it available on now? I don't really follow the PC
trade rags...

> 3. it will most likely not suffer the incompatibilities between
> different offerings of the same "standard", largely because
> it will be offered by a single provider. This is the old
> proprietary versus open systems debate, but it's laughable when
> different flavors of UNIX vary widely in system administration,
> standard tools, etc.

Aren't ports to different architectures being done by the hardware vendors
themselves (e.g. SGI, Sun)? If this is the case, how can MS prevent
incompatibilities from cropping up? How can they prevent vendors from
"adding value" to NT?
--
-----------------------------------------------------------------------
Phil Anglin | "Push the button, Frank!"
Millersville, MD | - Dr. Clayton Forrester
-----------------------------------------------------------------------

Jason 'KodaK' Balicki

unread,
Jan 11, 1994, 1:30:57 AM1/11/94
to
bors...@mcs.kent.edu (Bret Orsburn) writes:

>BTW: I just bought a brand new UNIX box, so don't flame me too much.
>I did make it a 486, though, because things are changing fast, and my
>486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
>have I left any out?

Yes, Linux.

David S. Masterson

unread,
Jan 11, 1994, 1:33:35 AM1/11/94
to
>>>>> Regarding Windows NT?; bra...@accord.ece.cmu.edu (J Eric Bracken) adds:

> How can UNIX "beat" NT in the marketplace (remembering that this includes
> a lot of naive PC users...)?

The capabilities of the two operating systems are immaterial. UNIX has the
edge in maturity, but that can quickly vanish if the UNIX market doesn't do
somethings very right over the next two years (2 years is the maximum time I
would give for NT to catch up in *most* people's minds).

The difference that naive PC users will be focusing on (especially in the home
market -- the next big market) is the plug'n'play capability ("Can I take it
out of the box and be running in 5 minutes?" or "System administration -- what
is that?"). The area the developers will (or should) be focusing on
cross-system ABIs ("Do I have to develop a version for Sun, HP, IBM, ...?").

If UNIX doesn't figure this out, Microsoft will win on the simple statement
that most people will be able to make:

Oh, that nice program I bought at the Mall on CD two years
ago will still work on this new machine with the larger
gigadisk and faster microfrotz?!? I'll take it...

--
====================================================================
David Masterson Consilium, Inc.
(415) 691-6311 640 Clyde Ct.
dav...@consilium.com Mtn. View, CA 94043
====================================================================
progasm: the feeling you get when your code works the first time

Dhaliwal Bikram Singh

unread,
Jan 11, 1994, 2:35:47 AM1/11/94
to
In article <2gt6fk$6...@sugar.NeoSoft.COM> dar...@sugar.NeoSoft.COM (Daryl Biberdorf) writes:
>IMHO, the only hope that current variants of UNIX have versus

{Deleted lines}

> 6. apps, apps, apps.
>
>Once the first year drought goes by, the apps that *need* NT will appear

I am sure that the day I need to a 16 meg beheamoth to run "the" app
I am sure will not be soon. There are so many more reliable alternatives
out there.

Everyday I look in the employment adds for programmers I see more and more
and more ads for unix/OSF/motif and OS/2 programmers. I think that this
is an implicit sign of what eyes-on-the-bottom-line business is looking
for.

How many times in recent history has a piece of software faltered out of
the starting gate and has recovered. I am affraid that many are putting
their money behind a dead horse.

>quickly. Those that need a lesser OS/user environment will go for
>plain Windows.
>
>I personally think the UNIX vendors cut their own throats by refusing to
>at least acknowledge the need for a leader like Novell (a company with

There is a phenom occuring at world university in the form of GNU type
standards, all based on free technology. This movement is quietly
coalescing, its only problem is that business is not quick or bold
enough to pick up on what is happening.

>enough clout to have 70% of the PC network market -- a share you
>CANNOT ignore).
>
>Daryl
>
>--
>dar...@sugar.neosoft.com

-bik

Andreas Wolpers

unread,
Jan 11, 1994, 3:55:52 AM1/11/94
to
In article <2gsk75$8...@usenet.mcs.kent.edu>, bors...@mcs.kent.edu (Bret Orsburn) writes:
|> In article <BRACKEN.94...@accord.ece.cmu.edu> bra...@accord.ece.cmu.edu (J Eric Bracken) writes:
|> >
|> >
|> >Just to get a little dialogue going:
|> >
|> > What does Windows NT have to offer that (most) UNIX's don't?
|>
|> Windows NT was conceived in an era when a graphical user interface is assumed.
|> [stuff deleted]

|> BTW: I just bought a brand new UNIX box, so don't flame me too much.
|> I did make it a 486, though, because things are changing fast, and my
|> 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
|> have I left any out?

Of cource! CP/M :-)

--
Andreas Wolpers
Email: wol...@dfki.uni-sb.de
Tel.: +49-681-302-5348
Fax.: +49-681-302-5341

Raymond Nijssen

unread,
Jan 11, 1994, 4:03:31 AM1/11/94
to
In article <2gt6fk$6...@sugar.NeoSoft.COM> dar...@sugar.NeoSoft.COM (Daryl Biberdorf) writes:

> Right now, Windows NT is faltering (and, yes, I like Microsoft to
> have to endure embarrassment once in a while). However, it has all
> of the following in its favor:

> 1. it provides a user interface that is readily usable by a large
> segment of the computer-using population (in the penetration wars,
> you can't divorce OS from user interface).

Wrong. In both cases, there is a clear separation between OS and UI. The
difference is that Windows hides the OS almost completely from Joe User, which
might make him think that the OS is somehow integrated into the UI (as with
Windows 3.x).

> 2. it looks to be built as an industrial strength OS once the kinks and
> bugs are out. (I still haven't seen a UNIX system that *isn't*

when will that be? how long have they been working on Windows already? It's
still one of the most unstable platforms in the marketplace - the difference
here is that PC users really don't mind having to reboot their system several
times a day.

> flakey.)

I know of little professional unix systems that aren't as stable as a rock.

> 3. it will most likely not suffer the incompatibilities between
> different offerings of the same "standard", largely because
> it will be offered by a single provider. This is the old
> proprietary versus open systems debate, but it's laughable when
> different flavors of UNIX vary widely in system administration,
> standard tools, etc.

Limiting freedom of choice makes things less confusing?

This statement is certainly true, but seems a less preferable ideology to me.

> 4. it is offered by a company that is generally respected by
> the computer using public (this excludes people like me, however)

If this were a major asset, what about OS/2 ?

> 5. NT gets to draw from everyone else's OS experience. This includes

This is by no means automatically true. Certainly, from an OS engineering
point of view, NT offers few things that we haven't seen before.
The point is that it takes a lot more to combine all desirable concepts
developed in the past into one product.

> OS/2, OpenVMS, the Big Blue mainframe OS's, *and* UNIX. UNIX is
> showing its age by comparison.

depends on what you mean with "UNIX". (eg. is MACH UNIX? or is OSF/1 UNIX?)
For most pragmatic people, any system providing all desired calls (as defined
in the X/OPEN standard), a unix-like filesystem interface, pipes and the whole
thing, are to be referred to as UNIX - if it looks like a duck, and if it
sounds like a duck, and if it walks like a duck...

So it's an evolving definition (mind that the definition of DOS has evolved
too)

> 6. apps, apps, apps.

granted. so what do you think about WABI? Or even advanced DOS emulators?

-Raymond

--
Raymond Nijssen | LOSEdoze is a more applicable name for MS-Windows
|
Eindhoven University of Techology, Design Automation (ES) | phone +31-40-473373
EH 7.16, PO. Box 513, 5600 MB Eindhoven, The Netherlands | fax +31-40-464527

Martin v.Loewis

unread,
Jan 11, 1994, 4:43:10 AM1/11/94
to
bors...@mcs.kent.edu (Bret Orsburn) writes:

>
> BTW: I just bought a brand new UNIX box, so don't flame me too much.
> I did make it a 486, though, because things are changing fast, and my
> 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
> have I left any out?
>

Solaris, NextStep (sp?)
Chicago, Cairo :-)

Mike Dowling

unread,
Jan 11, 1994, 5:37:35 AM1/11/94
to
>>> Bret Orsburn writes:

>> What does Windows NT have to offer that (most) UNIX's don't?

Bret> Windows NT was conceived in an era when a graphical user interface is
Bret> assumed.

Bret> Much of what UNIX has contributed to the state of the art evolved in an
Bret> era when a very clever tty-based interface was important: pipes,
Bret> redirection, symbolic device independence, powerful shells--even gnu
Bret> emacs.

Please forgive my heresy, but why do so many people seem to think that
graphical interfaces, be they for DOS or UNIX, are a good idea? I for one only
use a graphical interface (X11) when I absolutely have to; i.e., when I have a
graphics application. My dislike for these graphical interfaces stems from the
following criticisms.

1. Mindlessly aiming mice at what are often small dots on the screen is much
more time consuming than issuing a shell command.

2. You cannot cram much into a menu, so the result is that, if you want to do
something reasonably complex, you have to wade through an endless stream of
menus.

3. If you want to something non-trivial, more often than not, there is no
menu/mouse support for it. The result is a considerable loss of flexibility.

4. I'm a mathematician and so like to use my computing resources for my own,
mathematical programs. A few years ago, there was a wide consensus that
operating systems should not waste computing resources. Now people think
nothing of it when X11 slurps 8MB memory! That's 8MB that I cannot use for my
problems, and I don't like it! Unix is nice; I can kill X11 if I want to do
so. To Hell with OS/2 and NT, 'cos you can't kill the graphical interface, no
matter how desperate you are for memory.

5. I find graphical interfaces simply less comfortable to look at for any
length of time. I used to think that the reason was the bad quality of the
monitor I was using. It was not. I now have a decent 17" 80 Hz monitor, and I
still don't like it much. Even with a 19" monitor, you still cannot have
non-overlapping applications, so you are eternally aiming mice to raise
windows. I find it so much faster swap virtual consoles in text mode!

6. In the bad old days when I had OS/2, I did not find the graphical interface
a substitute for competence. In order to get the persentation manager into a
usable state, it cost a great deal of time and effort. There are no man pages,
and the on line help, which IBM claimed was so good, was anything but
intuitive. I remember once trying to install a printer spooler. A colleague
an I sat for over an hour trying to work out how to do it. In the end, it was
simple; you just pull an icon out of a box and onto the screen. I also wasted
a lot of time trying to get a printer spooler up and running for UNIX. In the
end, it turned out that I had repeatedly over looked the "rp", or "remote
printer" attribute in the printcap file. I lost a lot of time with both
systems, but, with OS/2, the reason was the unintuitive help, and with UNIX, it
was my own folly. Normally, UNIX is more complex, and so more powerful than
OS/2 (and I presume NT), and it generally is easier to implement stuff with
UNIX despite its complexity. Why? Simply because UNIX does not proceed on the
hypothesis that I am a incompetant fool, and things have to made simple, i.e.,
unintuitive with lots of inappropriate analogies.

Bret> Take a look at the old classic "The UNIX Programming Environment" by
Bret> Kernighan and Pike, and ask yourself: would you still recommend that book
Bret> to someone starting out today? (Somebody swiped my copy several years
Bret> ago, and I couldn't bring myself to replace it--too much of it now is
Bret> just historical curiosity.)

Well, actually, only last week I recommended this book to a UNIX novice. The
really nice thing about UNIX is the programmable shell. This enables the
computer competent to use their computer very efficiently indeed. Of course,
there are two sides to every coin; to obtain computer competence, one must
invest time to learn. It pays. MS-Windows is aimed at the computer
illiterate, and is all right for "application programmers", who really only
want to start WORD and later on, Lotus, and who otherwise have little
experience with computers. MS-Windows is unusable for programmers.

Bret> Even if UNIX survives (and it undoubtedly will, in some form), much of
Bret> what old-timers would recognize as the UNIX computing paradigm will have
Bret> lapsed into history.

You forget that true domain for UNIX has always been scientific computation,
and not "applications programming", which is the domain of DOS. You cannot
kill UNIX for workstations without first providing the programmer with the
tools he or she needs. DOS does not. OS/2 does not. Of course, you can port
the GNU UNIX utilities, but then, if you find yourself trying to make DOS look
like UNIX, you might as well go the whole hog, and get UNIX.

Bret> BTW: I just bought a brand new UNIX box, so don't flame me too much.

But, surely, flaming is what lists like this are all about? It is so much fun!

It is funny to see just how ernestly people take their operating systems.
People are always talking about one os killing another on the marketplace, and
how much better X is when compared with Y. Personally, I think that it is a
good thing to have a choice, and I don't see a conflict. DOS is all right for
applications programmers, and UNIX is not, because Word Perfect only uses DOS.
I cannot use DOS for my problems in scientific computation, because DOS cannot
find its own memory with both hands.

Mike

--
-----------------------------------------------------------------------------
Dr. Michael L. Dowling (__) moocow.math.nat.tu-bs.de
Abteilung f|r Mathematische Optimierung (oo)
Institut f|r Angewandte Mathematik \/-------\
TU Braunschweig || | \
Pockelsstr. 14 ||---W|| *
38306 Braunschweig ^^ ^^ Ph.: +49 (531) 391-7553
Germany
on.do...@zib-berlin.de
-----------------------------------------------------------------------------

Patrick Divine

unread,
Jan 11, 1994, 6:03:25 AM1/11/94
to
Bret Orsburn (bors...@mcs.kent.edu) wrote:

: BTW: I just bought a brand new UNIX box, so don't flame me too much.


: I did make it a 486, though, because things are changing fast, and my
: 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
: have I left any out?

LINUX


--
o----FISHTANK--Art-Xchange---// TBS (604) 535-6926 (data) || "You said it so
o POB 75115 // pdi...@unixg.ubc.ca || nicely I just
o __ White Rock, BC // Who is the modern Jesus? || had to laugh."
(__>< Canada, V4A 9M4 // Matthew 10:34 || - Anon.

Peter da Silva

unread,
Jan 11, 1994, 6:13:29 AM1/11/94
to
In article <MIKE.94Ja...@moocow.moocow.math.nat.tu-bs.de>,

Mike Dowling <on.do...@zib-berlin.de> wrote:
> 1. Mindlessly aiming mice at what are often small dots on the screen is much
> more time consuming than issuing a shell command.

I use a shell under a GUI, on both UNIX and AmigaOS.

> 2. You cannot cram much into a menu, so the result is that, if you want to do
> something reasonably complex, you have to wade through an endless stream of
> menus.

So you use a shell for that.

> 3. If you want to something non-trivial, more often than not, there is no
> menu/mouse support for it. The result is a considerable loss of flexibility.

Why? You're using a shell for that.

> 4. I'm a mathematician and so like to use my computing resources for my own,
> mathematical programs. A few years ago, there was a wide consensus that
> operating systems should not waste computing resources. Now people think
> nothing of it when X11 slurps 8MB memory! That's 8MB that I cannot use for my
> problems, and I don't like it! Unix is nice; I can kill X11 if I want to do
> so. To Hell with OS/2 and NT, 'cos you can't kill the graphical interface, no
> matter how desperate you are for memory.

Run the window manager and telnet clients in your X Terminal, and you're not
putting any load on the system when you're not running an X app.

> Even with a 19" monitor, you still cannot have
> non-overlapping applications,

I can fit 4 non-overlapping Xterms on my monitor, using the right font, and
have space for a couple of handy menu-driven shortcut programs written in Tk.

> 6. In the bad old days when I had OS/2, I did not find the graphical interface
> a substitute for competence.

There is no substitute for competance. The main advantage for a pure-GUI system
is that it limits the amount of damage the incompetant can do before saying
to themself, "Hey, that's just not right".

Colin Simpson

unread,
Jan 11, 1994, 6:45:59 AM1/11/94
to
In article <2gsk75$8...@usenet.mcs.kent.edu>, bors...@mcs.kent.edu (Bret Orsburn) writes:

Unix is simply an environment to program in. You can replace the shell with a
graphical interface if you wish (and this has been done). If you look at the
processes running on a unix machine you will see that the shell is in fact a user
process and is not in fact 'part' of the operating system.

Personally, I find GUI's useful for doing certain things such as selective file
copies, the visual previewer and selectors in xv3 and reading news is (I find)
easier under the X reader than the command line. I separately find the command
line easier to use for other things. My ideal environment is a GUI with a shell
window(s) so I can have the best of both worlds.

I also wonder why everyone goes on about Windoze NT, ok so Microsoft make it.
But, it does not seem to do anything that is new and improved, combine that with
the rotten windows environment. I also wonder how many PC users would be that
interested in it, given the machine spec. you need to run it. In addition, Your
average PC user does not appear to be over interested in getting the full 32-bit
performance out of their machine as so many of them have stuck to running DOS for
so long. Given these two reasons why should they switch from Windows 3.X.

I just hope NT flops 'cause I hate Microsoft for being so arrogant as to make
lots of money from such a rotten OS as DOS.

Colin Simpson

Guest of AA4

unread,
Jan 11, 1994, 9:34:11 AM1/11/94
to

Well, I just have to add my $0.02:

Some years ago, XENIX was the only stable multiuser operating system for
PC's (which we needed because of low cost). XENIX was (and is) the
implementation base of our communications software. I was very pleased
with the forming of the Open Software Foundation and expected UNIX to
'unfify' in a couple of years. I was somewhat sceptical of the participation
of the big-iron-salesman IBM, but figured it would eventually have to go
along with the other big companies and user demand.

Now, I find that almost *all* members of all UNIX unifying groups do *not*
want a single standard for UNIX. Understandibly, they want to bind customers
to their software / hardware by introducing incompatibilities (extra features).
The arrival of NT and Chicago, later this year, has brought about the most
serious UNIX unification effort, but I think it is too late. I expect that
Windows, and its compatibles and derivatives will 'rule the world' in 5 - 10
years from now.

Well, because Windows has the folowing advantages:

- OS
- Standard interface, advanced fonts, sound, video, music, display
- Less incompatibility
- Easy installation and maintenance (no tons of text files to edit)
- Cheap
- Huge installed base
- Low resources Windows (for mass), comparative resources for NT vs UNIX

- Software
- More variety
- Cheap
- Affordable productive programming languguages (C++, VisualBasic, dBASEs)
- Better quality (mostly)
- Vendor independance

- Hardware
- Cheap
- More diverse
- Vendor independant

Why not UNIX then? It has WABI, SoftPC and installed base? For UNIX to be
useful, it will have to have other advantages besides being able to run
Windows / DOS programs. If everything can be done in Windows, the UNIX below a
WABI is just overhead and complexity. Advantages of UNIX over Windows NT are:

- Multi user, but
I suspect that a WABI type product, that maps Windows calls to RPCs will
appear from third party vendors shortly. (Maybe already there, there are
some for modems). BTW, it puzzles me why Microsoft didn't standardize this
in NT.

- Installed base in business environments, but
- Smaller users will want Windows because of lower software /
hardware price, reduced complexity and personell availability.

- Mainframe users that want to switch, will switch to Windows in favour
of UNIX more and more because the the advantages mentioned above. (Even
a large Dutch bank, ING morgage department, to my surprise)

- The current number of UNIX users that are happy with their system may
not be large enough to sustain it. These users probably already have
Windows for standard office task like wordprocessing, spreadsheed, small
databases. Getting rid of the extra UNIX compexity will be appealing.

Don't get me wrong, I would like a standard OS that is vendor independant, but
economic realities give Windows the nod IMHO.

Richard

--
___________________________________________________
Fred Kwakkel | Fred.K...@cwi.nl (kwa...@cwi.nl)
| Centrum voor Wiskunde en Informatica
<>< | Kruislaan 413, Amsterdam, Nederland

Bjorn Hell Larsen

unread,
Jan 11, 1994, 8:48:53 AM1/11/94
to

In article <RAYMOND.94...@krait.es.ele.tue.nl> ray...@krait.es.ele.tue.nl (Raymond Nijssen) writes:
>
> when will that be? how long have they been working on Windows already? It's
> still one of the most unstable platforms in the marketplace - the difference
> here is that PC users really don't mind having to reboot their system several
> times a day.
>

I don't particularly *like* having to reboot my system several times a day.

But is having a stable system (ie. replacing Windows with Unix) worth
the price? In order to run Unix with some kind of performance I'd have
to double my RAM and disk capasities, and I would have to pay about
five times as much as I do for my software today.

bjørn

--
Bjorn Hell Larsen bla...@statoil.no
Corporate Information Technology
STATOIL I have trouble enough
Stavanger, Norway speaking for myself.


Juergen Nickelsen

unread,
Jan 11, 1994, 9:55:54 AM1/11/94
to
In article <2gt4d4$5...@sugar.NeoSoft.COM> pe...@sugar.NeoSoft.COM
(Peter da Silva) writes:

> UNIX fits well into a graphical user interface model. The pipeline metaphor
> works a LOT better than the flowcharts of most "visual" languages. It's a
> pity nobody's written the visual pipeline editor.

Doesn't the Khoros user interface implement exactly this? I know next
to nothing about it, but once I saw someone using it and heard a short
description. What I vaguely remember is that it has indeed a visual
pipeline editor which lets you combine filters etc. on picture data.

Of course the same thing would have to be different for use with Unix,
but I can imagine such a beast.

--
Juergen Nickelsen

Simon Patience

unread,
Jan 11, 1994, 10:08:34 AM1/11/94
to
In article <2gsk75$8...@usenet.mcs.kent.edu>, bors...@mcs.kent.edu (Bret Orsburn) writes:
> BTW: I just bought a brand new UNIX box, so don't flame me too much.
> I did make it a 486, though, because things are changing fast, and my
> 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
> have I left any out?

Yes, OSF/1.

Simon.

--
Simon Patience Phone: (617) 621-7376
Open Software Foundation FAX: (617) 621-8696
1 Cambridge Center Email: s...@osf.org
Cambridge, MA 02142 uunet!osf!sp

Eric P. Armstrong

unread,
Jan 11, 1994, 10:36:35 AM1/11/94
to
mi...@MooCow.MooCow.math.nat.tu-bs.de (Mike Dowling) writes:

>Please forgive my heresy, but why do so many people seem to think that
>graphical interfaces, be they for DOS or UNIX, are a good idea? I for one only
>use a graphical interface (X11) when I absolutely have to; i.e., when I have a

X11 is NOT a graphical interface, it is a windowing system. There is
a difference.

>graphics application. My dislike for these graphical interfaces stems from the
>following criticisms.

>1. Mindlessly aiming mice at what are often small dots on the screen is much
>more time consuming than issuing a shell command.

Bring up lots of shells in lots of windows at once. I usually have 4
shells on screen at all times, with space left over on the screen.

>2. You cannot cram much into a menu, so the result is that, if you want to do
>something reasonably complex, you have to wade through an endless stream of
>menus.

So don't use them, X does not force you to do much of anything.

>3. If you want to something non-trivial, more often than not, there is no
>menu/mouse support for it. The result is a considerable loss of flexibility.

If you want to do something non-trivial, you usually end up writing
your own program too. Most programs are very general and not good for
specifics.

>4. I'm a mathematician and so like to use my computing resources for my own,
>mathematical programs. A few years ago, there was a wide consensus that
>operating systems should not waste computing resources. Now people think
>nothing of it when X11 slurps 8MB memory! That's 8MB that I cannot use for my
>problems, and I don't like it! Unix is nice; I can kill X11 if I want to do
>so. To Hell with OS/2 and NT, 'cos you can't kill the graphical interface, no
>matter how desperate you are for memory.

So increase your swap space, if it isn't being used, out the door it goes.
There are lots of ways to get X to use less memory. Recompile it
without the extensions that you do not need. Use full optimization on
a good compiler, don't use complex window managers, use something
simple like twm instead of mwm or olwm. Don't bring up needless
windows, such as your favorite xload, xclock, xbiff, xdaliclock(ugh).

>5. I find graphical interfaces simply less comfortable to look at for any
>length of time. I used to think that the reason was the bad quality of the
>monitor I was using. It was not. I now have a decent 17" 80 Hz monitor, and I
>still don't like it much. Even with a 19" monitor, you still cannot have
>non-overlapping applications, so you are eternally aiming mice to raise

Monitor size has very little to do with screen resolution. I have
machines that run 1024x768 on both 14" and 19" monitors. You can use
a nice small readable font instead of a fancy font.
"-schumacher-clean-medium-r-normal--10-100-75-75-c-50-iso8859-1" is a
nice one. If you can't read it, see your optometrist, you may need
(new) glasses. :-)

>windows. I find it so much faster swap virtual consoles in text
mode!

What's text mode? :-)

>UNIX despite its complexity. Why? Simply because UNIX does not proceed on the
>hypothesis that I am a incompetant fool, and things have to made simple, i.e.,
>unintuitive with lots of inappropriate analogies.

It expects you to know what you are doing, pretty much everything
does in one way or another. Heck, the programmer knows how to do it,
who else needs to know? :-)

>Well, actually, only last week I recommended this book to a UNIX novice. The
>really nice thing about UNIX is the programmable shell. This enables the
>computer competent to use their computer very efficiently indeed. Of course,

shell programming efficient? sorta. :-)

>there are two sides to every coin; to obtain computer competence, one must
>invest time to learn. It pays. MS-Windows is aimed at the computer
>illiterate, and is all right for "application programmers", who really only
>want to start WORD and later on, Lotus, and who otherwise have little
>experience with computers. MS-Windows is unusable for programmers.

MS-Windows is more of a resource pig than unix for what it does too.
NT is worse. (NT is New Title)

>You forget that true domain for UNIX has always been scientific computation,
>and not "applications programming", which is the domain of DOS. You cannot

DOS is nothing but a dumb program loader, and isn't even good for
application programming, since you have to write an OS to go with your
program from scratch every time.

Also it sounds to me like you are running unix on a cheezy pc clone,
which is a horrid architecture to begin with (registers are keen).
Maybe you should buy a real unix workstation that has some power
behind it.

----
Eric P. Armstrong e...@phys.ksu.edu

Carl Kumaradas

unread,
Jan 11, 1994, 11:06:11 AM1/11/94
to
In article <MIKE.94Ja...@MooCow.MooCow.math.nat.tu-bs.de> on.do...@zib-berlin.de writes:
>so. To Hell with OS/2 and NT, 'cos you can't kill the graphical interface, no
>matter how desperate you are for memory.
>
>--

You definitely CAN kill the graphical user interface in OS/2. Actually you
choose to run OS/2 without the GUI, with the GIU but no desktop-type manager,
or with the GUI AND the desktop-type manager (workplace shell).

OS/2 wuthout the GUI uses about 1.5-2M of RAM, and 2.5M with the second
option. I've used it in all three configurations.
--
J. Carl Kumaradas Ontario Cancer Institute
ckum...@pip.oci.utoronto.ca University of Toronto

Spencer Dawkins

unread,
Jan 11, 1994, 11:37:46 AM1/11/94
to
In article <ENAG.94Ja...@gyda.ifi.uio.no>,
Erik Naggum <en...@ifi.uio.no> wrote:
>[Bret Orsburn]

>
>| Take a look at the old classic "The UNIX Programming Environment" by
>| Kernighan and Pike, and ask yourself: would you still recommend that
>| book to someone starting out today?
>
>_Especially_ today. I also recommend Don Knuth: "The Art of Computer
>Programming". Not just for history, but for understanding and appreciation
>of computers per se, apart from the glitzy GUI.

Oh, come ON! I'm as big a fan of Brian Kernighan's as there is, but "The
UNIX Programming Environment" is so old it ducks the vi vs EMACS question
by using "ed", and not only that, they give reasons why this is the right
thing to do! "The UNIX Programming Environment" was a terrific book when
it came out, but almost every tool it discusses has been overtaken by events:

was: cc now: c++
was: ed now: vi or EMACS
was: sh now: csh or ksh
was: awk, sed now: PERL
was: sccs now: either sccs or RCS

The discussion of lex and yacc is still quite good, but that's the only
area where "The UNIX Programming Environment" is still competitive, IMHO.

Nigel Horspool's "The Berkeley UNIX Environment" (Prentice-Hall, 1992) is
actually a second edition that has moved into the 1990s with some grace.
"The UNIX Programming Environment" badly needs a second edition -- I'd be
quite irate if someone recommended it to me without SIGNIFICANT qualifications.

I still owe Brian my hide, but that's because he still shows up as a major
contributor to the best UNIX books today ("Advanced Programming in the UNIX
Environment" by W. Richard Stevens being the most recent I've stumbled over).
Read the prefaces!

Spencer
--
------------------------------------------------------------------------------
- Spencer Dawkins Bell-Northern Research -
- (214) 684-4827 P.O. Box 833871 -
- Internet: daw...@bnr.ca Richardson TX 75083-3871 -

Science Research Lab

unread,
Jan 11, 1994, 4:29:09 AM1/11/94
to
bors...@mcs.kent.edu (Bret Orsburn) writes:

>BTW: I just bought a brand new UNIX box, so don't flame me too much.
>I did make it a 486, though, because things are changing fast, and my
>486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
>have I left any out?

Linux.


--
Science Research Laboratory Email: sre...@world.std.com
15 Ward Street Voice: 617-547-1122
Somerville MA 02143 Fax: 617-547-4104

Sean Brunnock

unread,
Jan 11, 1994, 9:57:26 AM1/11/94
to
In article <BRACKEN.94...@accord.ece.cmu.edu> bra...@accord.ece.cmu.edu (J Eric Bracken) writes:

> What does Windows NT have to offer that (most) UNIX's don't?

> How can UNIX "beat" NT in the marketplace (remembering that this
> includes a lot of naive PC users...)?

Unix has dominated the scientific/engineering market for the past
10 years or so. Microsoft has no intention of entering that market.

The market that Unix and NT will be battling for is the corporate
database market which IBM has dominated for so long. Microprocessor
based systems now rival mainframes in terms of speed, but micros offer
a much greater value than mainframes. Corporate America is slowly
downsizing their mainframe-based information systems to client-server
networks.

The clients will be DOS/Windows-based PCs. There is no compelling
reason to install NT or UNIX on these 50 million+ systems. NT and
Unix will be battling for the server end of the market. We're not
talking about millions of systems here. There are only 50,000 IBM
mainframes in current use around the world.

In my mind, Unix has several advantages over NT as far as conquering
the corporate database market goes:

1.) Unix has been around for 20 years. It has very good name recognition
(in a CIO magazine poll asking subscribers what technologies they
would be incorporating, Unix was the most popular choice) and it has
already made in-roads in the financial services market (Sun).

2.) Microsofts biggest competitor will be Novell which already dominates
the PC LAN market and will be pushing UNIX as its server OS.

3.) There are already lots of products which allow PCs and Macs connect
to UNIX systems. NT is playing catch-up.

Sean Brunnock

Message has been deleted

Bret Orsburn

unread,
Jan 11, 1994, 12:53:06 PM1/11/94
to
In article <2gtpio...@sbusol.rz.uni-sb.de> wol...@cs.uni-sb.de (Andreas Wolpers) writes:
>In article <2gsk75$8...@usenet.mcs.kent.edu>, bors...@mcs.kent.edu (Bret Orsburn) writes:
>|> BTW: I just bought a brand new UNIX box, so don't flame me too much.
>|> I did make it a 486, though, because things are changing fast, and my
>|> 486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
>|> have I left any out?
>
>Of cource! CP/M :-)

Not a problem. I still have a Z80/S-100 machine from Quasar Data Products,
where I got my first systems programming job.

It has 256KB RAM, bank-switched, that supports a banked CP/M BIOS (to increase
the TPA) and lots of disk cache (for my dual 1.2MB 8" floppies).

BTW: To all the Linux fans: I listed flavors of UNIX, not specific versions.
As I understand it, Linux is SV-ish, and my intent was to subsume it in that
category. My personal preference is for BSD-ish.

---

Bret Orsburn
bors...@mcs.kent.edu

Bambang Nurcahyo Prastowo

unread,
Jan 11, 1994, 12:12:24 PM1/11/94
to
In article <2guh23$c...@newserv.ksu.ksu.edu> e...@phys.ksu.edu (Eric P. Armstrong) writes:
>a good compiler, don't use complex window managers, use something
>simple like twm instead of mwm or olwm. Don't bring up needless

Use fvwm, it is as nice as mwm or olwm, it lets you define keyboard
replacement of mouse actions, and take 1/3 less memory than twm.

Bambang
--
Bambang N. Prastowo (pras...@qucis.queensu.ca, grad. student)
Software Technology Laboratory, Dept. Computing & Info. Sci.
Queen's Univ., Kingston, Ontario, Canada K7L 3N6

Tim Llewellyn

unread,
Jan 11, 1994, 1:02:58 PM1/11/94
to

In article <DAVIDM.94J...@malibu.consilium.com>, dav...@consilium.com
(David S. Masterson) writes:

[snip...]


|>The capabilities of the two operating systems are immaterial. UNIX has the
|>edge in maturity,


aw comeon. I can't let this one pass. Are u trying to persuade up that each
of the maze of slightly different unix variants includes the same code? Of
course not. They may be implemented to the same or similar specs, but to me
that doesn't necessarily imply a mature system. Especially if the spec is out
of date and flawed like Unix so obviously is to anyone whos used any other
sort of semi-serious comercial machine/OS :-)

[snip...]

|>The difference that naive PC users will be focusing on (especially in the
|>home
|>market -- the next big market) is the plug'n'play capability ("Can I take it
|>out of the box and be running in 5 minutes?" or "System administration --
|>what
|>is that?"). The area the developers will (or should) be focusing on
|>cross-system ABIs ("Do I have to develop a version for Sun, HP, IBM, ...?").
|>

Interesting point. We are having great fun this week with a couple of
DECStations
running Ultrix. Basically we had to change system boxes and a few controllers.
Ultrix wouldn't even BOOT without totally reinstalling.

Apparently some Unices do configure hardware dynamically :-) Apparently ...
Hmmm

[snip...]

|>--
|>====================================================================
|>David Masterson Consilium, Inc.
|>(415) 691-6311 640 Clyde Ct.
|>dav...@consilium.com Mtn. View, CA 94043
|>====================================================================
|>progasm: the feeling you get when your code works the first time
|>

--
-----------------------------------------------------+---------------+
Tim Llewellyn - OpenVMS, Soukous and Cricket Addict | Read at your |
Physicist Programmer, Bristol Uni Particle Physics. | own risk. |
HEPNET/SPAN 19716::TJL Internet t...@siva.bris.ac.uk | Std disclaimer|
Pet Hates: Case Sensitivity! Unix. Tremolo systems. | implicit |
-----------------------------------------------------+---------------+

Bruce Barnett

unread,
Jan 11, 1994, 2:25:01 PM1/11/94
to
In article <2gsk75$8...@usenet.mcs.kent.edu> bors...@mcs.kent.edu (Bret Orsburn) writes:

> Take a look at the old classic "The UNIX Programming Environment" by
> Kernighan and Pike, and ask yourself: would you still recommend that
> book to someone starting out today?


Yes I do. It is a good example of ...

Of course you could write a program from scratch, but why do
this when you can interface to several other programs that already does the
hard part? Use your program to call another program, and bootstrap
your code onto something that is already working. This sames you time
and energy, and it is perfectly acceptable to do this.

and

Build a prototype that has the correct function, and
add and correct the functionality before you add speed, bells and whistles.

Since the same principles hold true with or without a GUI, I don't see
this as obsolete.

Which PC/Mac environments allow you to combine *programs* with a
scripting environment? (Note plural). I am not criticizing
PC/Mac/Windows NT. Just ignorant.

--
Bruce Barnett <bar...@crd.ge.com> uunet!crdras!barnett

Brian Powell

unread,
Jan 11, 1994, 2:47:26 PM1/11/94
to
Sayeth Daryl Biberdorf:
:
: Right now, Windows NT is faltering (and, yes, I like Microsoft to

: have to endure embarrassment once in a while). However, it has all
: of the following in its favor:
...
: 6. apps, apps, apps.

:
: Once the first year drought goes by, the apps that *need* NT will appear
: quickly.

Do you realize just how many shrink wrapped apps are available for a
popular UNIX version like Sun Solaris? There are a LOT. If NT is
*really* lucky it may reach that number of NT-specific apps in several
years. Also, with WABI or SoftPC, Solaris can run most Win-31 apps...

-- Brian

o-----------------------The Ohio Supercomputer Center-----------------------o
| Brian S. Powell bpo...@osc.edu |
o-----------------------"My other computer is a CRAY"-----------------------o

Brian Powell

unread,
Jan 11, 1994, 2:56:09 PM1/11/94
to
Sayeth Raymond Nijssen:
: In article <2gt6fk$6...@sugar.NeoSoft.COM> dar...@sugar.NeoSoft.COM (Daryl Biberdorf) writes:
: >
: > 4. it is offered by a company that is generally respected by
: > the computer using public (this excludes people like me, however)
:
: If this were a major asset, what about OS/2 ?

What about it? It is selling *very* well and is approaching a 4
million installed base. What is UNIX's installed base after 20 years?
Don't get me wrong, I really like UNIX and use in at work, but "The
rumors of OS/2's death have been greatly exaggerated." :-) It's about
time people stopped using OS/2 as an example of a failed OS. Version
2.x is quite successful and has a very bright future. Now that
Microsoft is out of the equation, OS/2 is really succeeding.

Brian Powell

unread,
Jan 11, 1994, 3:03:52 PM1/11/94
to
Sayeth David Barr:
: In article <2gukkq$e...@crchh903.bnr.ca>,
: Spencer Dawkins <daw...@bnr.ca> wrote:
: > was: cc now: c++
:
: Ha! C++ is dying. People mistakenly looked up on C++ as being better
: than C. Now people are really waking up. (The programming language
: community woke up several years ago, - the general programming community
: still is starting to catch on) C++ is no longer considered the
: "successor" to C as it once was.

I disagree.

Look at the help wanted pages of "Computerworld" or "Open Systems
Today" and count the number of C positions vs. C++ positions wanted.
You'll find that they're pretty equal. That's a lot of available jobs
for a "dying langauge". I've noticed more and more C++ positions each
time I look at these magazines...

Felix Sebastian Gallo

unread,
Jan 11, 1994, 3:28:22 PM1/11/94
to
daw...@bnr.ca (Spencer Dawkins) writes:

>"The UNIX Programming Environment" was a terrific book when
>it came out, but almost every tool it discusses has been overtaken by events:

> was: cc now: c++

False. One trillion c programmers disagree.

> was: ed now: vi or EMACS

sure; however, the principles behind ed live on in vi's command
language, as well as in 'sed', which is an extremely important
thing to know. Knowing the ideology behind 'ed' helps a great
deal with the ideology behind Unix.

> was: sh now: csh or ksh

No, it's still sh, and it'll always be sh. If you ever feel
tempted to write a script in csh, please read a few articles
by Tom Christiansen first.

> was: awk, sed now: PERL

perl's great, but it builds on, rather than replaces, knowledge
of sed and awk. It would be sort of difficult to sit down with
a newbie and explain perl in terms of perl rather than perl in
terms of sed, at least to my mind. Also, perl's more of an
internet phenomenon than a new unix standard, as yet.

> was: sccs now: either sccs or RCS

i.e., is still sccs. :)

>"The UNIX Programming Environment" badly needs a second edition -- I'd be
>quite irate if someone recommended it to me without SIGNIFICANT qualifications.

I'd question your irateness. Unix, although it lends itself to
enormous monolithic processor-chowing whores like emacs, c++, and perl,
does not require, and should not require, their use. The basic
Unix toolkit as shipped on the tape is remarkably expressive, and
is often the only thing needed for a given task.

If you learn the basic tools (including ed, sed, lex, yacc, awk,
cat, dd, rm, ls, mv, sh, cp, grep), then you are one with the Tao
of Unix, and you will pass through all barriers, and the wheedlings
and mewlings of your users will be as the first droplets of summer
rain on the broad shoulders of a toadstool to you. Your power will
be spoken of across the land, and you shall not need to lap at the
rotten trough of object orientation nor the feedback of featurism.
Should you need to ride the wild stallions of perl or emacs, you
will be able to do so with a calm, confident rein rather than with
the inchoate, senseless spirit of abandon.

And many will say unto you, "you need emacs," or "it is important
that you install perl upon thy system," and your laughter will
roll as thunder, and the weak and craven will be smitten down to
cower in the rank puddles of their own ignorance. For your hands
will fly over the keyset, yea, even when you have only a bootable
emergency disk, and with cpio and dd you will have the lever to
move the world.

>Spencer

e...@netcom.com

unread,
Jan 11, 1994, 4:02:24 PM1/11/94
to


And C needs more programmers to maintain an equivalent number of
programs because the C programs are more work to maintain. Those ads
for C programmers might be mostly for maintenance of old C code, which
might not even be ANSI C.

Nick Maximov

unread,
Jan 11, 1994, 4:10:20 PM1/11/94
to
Brian Powell (bpo...@osc.edu) in comp.lang.c write:
> Sayeth David Barr:
> : In article <2gukkq$e...@crchh903.bnr.ca>,
> : Spencer Dawkins <daw...@bnr.ca> wrote:
> : > was: cc now: c++
> :
> : Ha! C++ is dying. People mistakenly looked up on C++ as being better
> : than C. Now people are really waking up. (The programming language
> : community woke up several years ago, - the general programming community
> : still is starting to catch on) C++ is no longer considered the
> : "successor" to C as it once was.
>
> I disagree.
>
> Look at the help wanted pages of "Computerworld" or "Open Systems
> Today" and count the number of C positions vs. C++ positions wanted.
> You'll find that they're pretty equal. That's a lot of available jobs
> for a "dying langauge". I've noticed more and more C++ positions each
> time I look at these magazines...

Geat idea. C++ is suitable language for windows and similar object
technology. But if you are the system programmer C++ is getting dull.
-- IMHO --


--
Nick Maximov

PC Centre "Techno" Representative
Internet: ma...@adnserv.kaija.spb.su

Deon Botha

unread,
Jan 11, 1994, 5:34:15 AM1/11/94
to
In <2gsk75$8...@usenet.mcs.kent.edu> bors...@mcs.kent.edu (Bret Orsburn) writes:

>In article <BRACKEN.94...@accord.ece.cmu.edu> bra...@accord.ece.cmu.edu (J Eric Bracken) writes:
>>
>>

>>Just to get a little dialogue going:
>>

>> What does Windows NT have to offer that (most) UNIX's don't?

>Windows NT was conceived in an era when a graphical user interface is assumed.

I think the fact that most applications in the larger commercial environment
are financial/database type system and typically run as back-end servers.
So they don't need fancy graphics. That is left to the Window/Mac/etc etc.
front-end clients. For a server we need a rich, stable mature environment
which has some historic momentum to ensure continuity. NT (Nice Try) is
a misfit IMHO except in environments which are primarily PC LAN based where
it apparently makes a good SQL Server and multi-LAN protocol (brouter). NT
will battle with batteries of online data-capturers capturing invoices and
credit-notes. Well I suppose EDI will come and rescue us eventually.

>BTW: I just bought a brand new UNIX box, so don't flame me too much.
>I did make it a 486, though, because things are changing fast, and my
>486 lets me opt for BSD,SVR4,SVR3,Mach,OS/2,MS-Windows,Windows NT--
>have I left any out?

Linux and NextStep and the biggest of them all - DOS.

In a mature organisation, the choice of OS is not going to be the critical
one. They all interoperate in any case (or are claiming to be almost there).
--
...............................................................
Deon Botha | Tel: +27 (21) 419-2690
Aztec Information Management | Fax: +27 (21) 25-1254
de...@aztec.co.za | Cape Town, South Africa

David S. Masterson

unread,
Jan 11, 1994, 2:25:06 PM1/11/94
to
>>>>> Regarding Re: Windows NT?; a336...@cdf.toronto.edu (Dhaliwal Bikram Singh) adds:

> Everyday I look in the employment adds for programmers I see more and more
> and more ads for unix/OSF/motif and OS/2 programmers. I think that this
> is an implicit sign of what eyes-on-the-bottom-line business is looking
> for.

> How many times in recent history has a piece of software faltered out of
> the starting gate and has recovered. I am affraid that many are putting
> their money behind a dead horse.

"faltered out of the starting gate"...?

Can you say OS/2? I knew you could...


--
====================================================================
David Masterson Consilium, Inc.
(415) 691-6311 640 Clyde Ct.
dav...@consilium.com Mtn. View, CA 94043
====================================================================

"Software is the heart and soul of a computer company."
-- DEC President Ken Olsen

Message has been deleted

Anton Rang

unread,
Jan 11, 1994, 12:43:05 PM1/11/94
to
In article <2guvof$3...@mane.cgrg.ohio-state.edu> bpo...@osc.edu (Brian Powell) writes:
>Do you realize just how many shrink wrapped apps are available for a
>popular UNIX version like Sun Solaris? There are a LOT. If NT is
>*really* lucky it may reach that number of NT-specific apps in several
>years.

Do you have specific numbers? I find it hard to believe that the
number of applications for any OS (like Solaris) with a tiny installed
base (as compared to {Mac/MS-DOS/Win3.x}) is that large. If WNT gets
a large number of users initially, through emulation of Win3 or
whatever method, it will make economic sense to develop for it, and
the applications will be inexpensive enough that ordinary users can
afford them.

Right now, unless you're producing high-end software and can charge
accordingly, it simply doesn't make sense to develop except in the PC
world. The installed base of any particular UNIX version that I know
of -- how big is the Solaris 1.x base? 2.x? -- is small compared to
the 10+ million Macintoshes, or the even larger PC market.

If UNIX were monolithic, I think it would have a better chance of
competing. As it is, even the small utilities I've done -- text-based,
some networking involved -- have hundreds of lines of conditional
compilation to get them to work on SunOS/HPUX/SVR3/SVR4, etc. And each
vendor has different bugs to be worked around. Oh, joy ....
--
Anton Rang (ra...@gnu.ai.mit.edu)

Daryl Biberdorf

unread,
Jan 11, 1994, 11:58:36 PM1/11/94
to
In article <RAYMOND.94...@krait.es.ele.tue.nl>,

Raymond Nijssen <ray...@krait.es.ele.tue.nl> wrote:
>In article <2gt6fk$6...@sugar.NeoSoft.COM> dar...@sugar.NeoSoft.COM (Daryl Biberdorf) writes:
> > 1. it [NT] provides a user interface that is readily usable by a large
> > segment of the computer-using population (in the penetration wars,
> > you can't divorce OS from user interface).
>
>Wrong. In both cases, there is a clear separation between OS and UI. The
>difference is that Windows hides the OS almost completely from Joe User, which
>might make him think that the OS is somehow integrated into the UI (as with
>Windows 3.x).

To quote myself, I clearly stated, "in the pentration wars,..." What
counts in this particular game is market penetration. From my
experience as a consultant, only a few IS shops understand the
sometimes nebulous distinction between user interface and OS.

In the case of Windows 3.x (not NT), Windows blurs that line. The only
OS function that DOS is obviously used for is its file system (and
some interrupt handling).

In the case of Windows NT, DOS isn't even a factor -- it's not even
remotely necessary to have DOS to run NT.
>
> > 2. it looks to be built as an industrial strength OS once the kinks and
> > bugs are out. (I still haven't seen a UNIX system that *isn't*


>
>when will that be? how long have they been working on Windows already? It's
>still one of the most unstable platforms in the marketplace - the difference
>here is that PC users really don't mind having to reboot their system several
>times a day.

Windows 3.x is an unprotected operating system (ala Amiga and Macintosh).
Windows NT is protected -- that adds quite a bit of stability right there.
Please get this point now, Windows NT is *NOT* Windows 3.x. They share
a common user interface and that's it.

>
>I know of little professional unix systems that aren't as stable as a rock.

I've used at least a dozen UNIX systems, and they all get "wedged"
entirely too often ("wedged" is a term used by my local system management
to explain when things get screwed up unexplainably).
>
> > 3. it will most likely not suffer the incompatibilities between
> > different offerings of the same "standard", largely because
> > it will be offered by a single provider. This is the old
> > proprietary versus open systems debate, but it's laughable when
> > different flavors of UNIX vary widely in system administration,
> > standard tools, etc.
>
>Limiting freedom of choice makes things less confusing?
>This statement is certainly true, but seems a less preferable ideology to me.

To you, perhaps. To the people who buy computers in quantities that matter
(companies), being able to get up and running simply and quickly is
a top priority. Personnel costs can easily eat up any hardware/software
price advantage you got by buying an "open" system. Compute how long
it takes to eat up a $10,000 price advantage when your personnel cost
about $30/hour, and consultants cost $50 or more per hour. Not very long.


>
> > 4. it is offered by a company that is generally respected by
> > the computer using public (this excludes people like me, however)
>
>If this were a major asset, what about OS/2 ?

What about it? Once Microsoft's sabotage attempts were out of the picture,
OS/2 2.x has been selling well beyond anyone's expectations.

> > 5. NT gets to draw from everyone else's OS experience. This includes
>This is by no means automatically true. Certainly, from an OS engineering
>point of view, NT offers few things that we haven't seen before.
>The point is that it takes a lot more to combine all desirable concepts
>developed in the past into one product.

Which is exactly what NT is doing.
>
>granted. so what do you think about WABI? Or even advanced DOS emulators?

"Imitation is the sincerest form of flattery." As OS/2 is amply
demonstrating, emulating popular platforms is only a way to minimize
the pain of switching to another platform. If people want a Macintosh,
they will buy a Macintosh, not an Amiga with an emulator. The emulator
will always be playing catch-up.


Daryl

--
dar...@sugar.neosoft.com

Carl L. Gay

unread,
Jan 12, 1994, 12:18:32 AM1/12/94
to

From: mi...@MooCow.MooCow.math.nat.tu-bs.de (Mike Dowling)
Date: 11 Jan 94 02:37:35 PST

Please forgive my heresy, but why do so many people seem to think that
graphical interfaces, be they for DOS or UNIX, are a good idea? I for one only
use a graphical interface (X11) when I absolutely have to; i.e., when I have a

graphics application. My dislike for these graphical interfaces stems from the
following criticisms.

Your criticisms seem to be against specific GUIs as opposed to GUIs in
general.

1. Mindlessly aiming mice at what are often small dots on the screen is much
more time consuming than issuing a shell command.

2. You cannot cram much into a menu, so the result is that, if you want to do


something reasonably complex, you have to wade through an endless stream of
menus.

3. If you want to something non-trivial, more often than not, there is no


menu/mouse support for it. The result is a considerable loss of flexibility.

I think the above three points fall into the same category: Any good
GUI should provide access to all commands via both the keyboard *and*
the mouse (or other input device). Most are pretty bad in this area,
and X (from my limited experience with it) is abysmal. Some commands
are clearly more efficiently entered via the keyboard and some through
menus. The latter especially for novices.

4. I'm a mathematician and so like to use my computing resources for my own,
mathematical programs. A few years ago, there was a wide consensus that
operating systems should not waste computing resources. Now people think
nothing of it when X11 slurps 8MB memory! That's 8MB that I cannot use for my
problems, and I don't like it! Unix is nice; I can kill X11 if I want to do

so. To Hell with OS/2 and NT, 'cos you can't kill the graphical interface, no
matter how desperate you are for memory.

Then X is a hog. I'm guessing that this is because it is a lot more
than a GUI. (Well, it's a lot less too... the window manager you
choose to run on top of X is probably more accurately the GUI.)
Anyway, a simpler GUI that actually makes some things easier for you,
such as on the Mac, doesn't have to be (and isn't) so huge.

5. I find graphical interfaces simply less comfortable to look at for any
length of time. I used to think that the reason was the bad quality of the
monitor I was using. It was not. I now have a decent 17" 80 Hz monitor, and I
still don't like it much. Even with a 19" monitor, you still cannot have
non-overlapping applications, so you are eternally aiming mice to raise

windows. I find it so much faster swap virtual consoles in text mode!

What do you mean you can't have non-overlapping applications? If you
mean there isn't an infinite amount of desktop space, you're right.
It's like that in real life too. If your complaint is that you have
to use the mouse to switch applications, I'm sure that can be fixed.
(If you find a fix, let me know; I'd like it too, for that extremely
rare occasion when I find that X would provide enough value added over
my VT100 emulator to actually go to the lab physically and sit in
front of an X-terminal.)

6. In the bad old days when I had OS/2, I did not find the graphical interface
a substitute for competence. In order to get the persentation manager into a
usable state, it cost a great deal of time and effort. There are no man pages,
and the on line help, which IBM claimed was so good, was anything but
intuitive. I remember once trying to install a printer spooler. A colleague
an I sat for over an hour trying to work out how to do it. In the end, it was
simple; you just pull an icon out of a box and onto the screen. I also wasted
a lot of time trying to get a printer spooler up and running for UNIX. In the
end, it turned out that I had repeatedly over looked the "rp", or "remote
printer" attribute in the printcap file. I lost a lot of time with both
systems, but, with OS/2, the reason was the unintuitive help, and with UNIX, it
was my own folly.
^^^^^^^^^^^^
Normally, UNIX is more complex, and so more powerful than
OS/2 (and I presume NT), and it generally is easier to implement stuff with


UNIX despite its complexity. Why? Simply because UNIX does not proceed on the
hypothesis that I am a incompetant fool, and things have to made simple, i.e.,
unintuitive with lots of inappropriate analogies.

I find it baffling that in the case of OS/2 you blame the OS because
it was unintuitive (though easy) and with Unix you blame yourself even
though it was equally unintuitive (and difficult). Why is making
things easier to do equivalent, in your mind, to assuming you are an
incompetant fool?

Mark Justin Cecil

unread,
Jan 12, 1994, 12:54:37 AM1/12/94
to
bpo...@osc.edu (Brian Powell) writes:
>Look at the help wanted pages of "Computerworld" or "Open Systems
>Today" and count the number of C positions vs. C++ positions wanted.
>You'll find that they're pretty equal. That's a lot of available jobs
>for a "dying langauge". I've noticed more and more C++ positions each
>time I look at these magazines...

It may be a fact that you see more and more C++ ads, but keep in mind that
the industry bought into C++ in A BIG WAY, and they will simply try to
protect their investment by continuing to hire people proficient in the
language. This doesn't necessarily mean that the industry is happy with C++,
just that they spent a lot of money on it and they aren't willing to give it
up just because it isn't what they expected.

Hey, I know shops that still use RPG...

Mark
--
===========================================================================
Mark Justin Cecil || m...@nuance.com
Cray Systems Analyst || (205) 955-4777 (day)
USASSDC Simulation Center, Huntsville, AL || (205) 837-6750 (night)

Erik Naggum

unread,
Jan 12, 1994, 1:32:17 AM1/12/94
to
[Spencer Dawkins]

| Oh, come ON! I'm as big a fan of Brian Kernighan's as there is, but
| "The UNIX Programming Environment" is so old it ducks the vi vs EMACS
| question by using "ed", and not only that, they give reasons why this is
| the right thing to do!

I guess it's a question of what you wish to accomplish. I have saved the
day many a time by knowing ed, and I still use ed to edit essential system
files when it is vitally important that I can look up the screen and see
what the line used to be before I "fixed" it. This "before" and "after"
feature in ed is not found in either vi or emacs.

Moreover, ed is so much more convenient in scripts...

It is no longer a viable alternative when programming or writing text,
though. I write non-linearly, going back to add words, swap sentences,
etc, and I don't think I could handle a "typewriter" like ed, anymore. But
not _knowing_ ed? It's like thinking you don't need to read because you
have cable TV.

| "The UNIX Programming Environment" was a terrific book when it came
| out, but almost every tool it discusses has been overtaken by events:
|
| was: cc now: c++
| was: ed now: vi or EMACS
| was: sh now: csh or ksh
| was: awk, sed now: PERL
| was: sccs now: either sccs or RCS

GNU tools. You forget all the GNU tools.

Another important thing is being able to use _any_ UNIX system you might
come across. Knowing the basic tools, you will never be at a loss for how
to get your job done, and you will be more efficient and professional than
one who complains that he can't find GNU Emacs 19.22 on a system, and
proceeds to complain about the Internet bandwidth, disk space, not having
gcc 2.5.7 to compile emacs, etc, etc.

I also carry a Maglite and the largest Victornix "Swiss Army" pocket knife.
I guess it goes with the character.

Best regards,
</Erik>
--
Erik Naggum <er...@naggum.no> <SG...@ifi.uio.no> | Memento, terrigena.
ISO 8879 SGML, ISO 10744 HyTime, ISO 10646 UCS | Memento, vita brevis.

Mike Dowling

unread,
Jan 12, 1994, 2:26:46 AM1/12/94
to
>>> On 11 Jan 1994 19:56:09 GMT, bpo...@osc.edu (Brian Powell) said:

Brian> What about it? It is selling *very* well and is approaching a 4 million
Brian> installed base. What is UNIX's installed base after 20 years? Don't
Brian> get me wrong, I really like UNIX and use in at work, but "The rumors of
Brian> OS/2's death have been greatly exaggerated." :-) It's about time people
Brian> stopped using OS/2 as an example of a failed OS. Version 2.x is quite
Brian> successful and has a very bright future. Now that Microsoft is out of
Brian> the equation, OS/2 is really succeeding.

There is nothing wrong with OS/2, except

(a) With the latest tcp/ip stuff you cannot even log onto an HP
(b) NFS is slower than a wet week, and you cannot write to an NFS mounted disk
(c) If you have the misfortune to telnet to an OS/2 machine while sombody is
using a DOS application, then you have to wait eons even for "DIR"
(d) If you want to use xterms from remote UNIX machines, you need a lot of
patience as it too is as slow as a wet week.
(e) Everything on an OS/2 machine is very significantly slower than the same
thing on the same box running linux.
(f) There is no means of controlling who gets access to what. This is crucial
in a networked environment.
(g) It's a single user system. If you have ever had to share an OS/2 box with
others, then you you will know why that is important.

No, OS/2 is not dead, and nor is DOS, and for the same reasons!

Mike
--
-----------------------------------------------------------------------------
Dr. Michael L. Dowling (__) moocow.math.nat.tu-bs.de
Abteilung f|r Mathematische Optimierung (oo)
Institut f|r Angewandte Mathematik \/-------\
TU Braunschweig || | \
Pockelsstr. 14 ||---W|| *
38306 Braunschweig ^^ ^^ Ph.: +49 (531) 391-7553
Germany
on.do...@zib-berlin.de
-----------------------------------------------------------------------------

Chris Gray

unread,
Jan 12, 1994, 3:39:01 AM1/12/94
to

-In article <2gv0n8$3...@mane.cgrg.ohio-state.edu>, bpo...@osc.edu (Brian Powell) writes:

> Look at the help wanted pages of "Computerworld" or "Open Systems
> Today" and count the number of C positions vs. C++ positions wanted.
> You'll find that they're pretty equal. That's a lot of available jobs
> for a "dying langauge". I've noticed more and more C++ positions each
> time I look at these magazines...

This is the *only* reason I intend to study C++ one day, i.e. as soon as I
get Bjarne's book back from the freind I leant it to because he thought it
would help him in his search for a job... I suppose it is also the reason
why so many poor souls are having C++ thrust upon them as a first language
for teaching purposes.

Of course Real Programmers (tm) will claim to know any language if the job
looks interesting; and by the end of the week they know it damn well... ;)

__________________________________________________________________________
Chris Gray cg...@se.alcbel.be Compu$erve: 100065.2102
Ignore my broken mailer - the addresses above are the only truth
__________________________________________________________________________

F*ck modularity - how about readability and ease of writing and
understanding what you (and others) write?

- Linus Torvalds

Benjamin Z. Goldsteen

unread,
Jan 12, 1994, 1:34:23 AM1/12/94
to
bar...@grymoire.crd.ge.com (Bruce Barnett) writes:

>Which PC/Mac environments allow you to combine *programs* with a
>scripting environment? (Note plural). I am not criticizing
>PC/Mac/Windows NT. Just ignorant.

I used to work with this Mac guy and I have to say what they are
doing now with AOCE, AppleScript, etc is pretty impressive. He said he
can just call up the spreadsheet from Excel for getting input for
example. The language is fairly English like. My biggest problem with
the Mac right now is that it is not protected mode, I don't like the
look&feel, and the 68K compatible macs are slow(GUI response and
numerical work).

[This may have been possible with NextStep for a long time...not
sure...I found NextStep really nice but never got to spend a lot of
time with it]
--
Benjamin Z. Goldsteen

Thomas G. McWilliams

unread,
Jan 12, 1994, 6:20:58 AM1/12/94
to
: BTW: To all the Linux fans: I listed flavors of UNIX, not specific versions.

: As I understand it, Linux is SV-ish, and my intent was to subsume it in that
: category. My personal preference is for BSD-ish.

This is something of a myth. Linux is modeled upon POSIX, and was
heavily influenced by BSD. The SunOS system manuals were often
used as an arbiter for many design decisions. The ubiquitous use
of the GNU utilities by Linux distributions also provides a strong
BSD influence. In 1994 it seems many of the Unixes have converged
upon a similar functional model as Linux. The commercial product by
BSDI BSD/386 spends most of its advertising touting its POSIX
compliance. Other NET-2 derivatives such as the NetBSD are busy
adding POSIX and SysV extensions such as IPC. The lines between a
a classic BSD system and a SysV system are increasingly blurred.
I think this is a good thing.

D'Arcy J.M. Cain

unread,
Jan 12, 1994, 8:39:30 AM1/12/94
to
bpo...@osc.edu (Brian Powell) writes:
>Look at the help wanted pages of "Computerworld" or "Open Systems
>Today" and count the number of C positions vs. C++ positions wanted.
>You'll find that they're pretty equal. That's a lot of available jobs
>for a "dying langauge". I've noticed more and more C++ positions each
>time I look at these magazines...

Yes but how many people have you met who refer to themselves as C++
programmers only to find out that they program C in Borland C++ except
for, perhaps, the use of "//"" for comment delimiters?

Also I bet a lot of people placing the ads don't really understand the
difference and simply ask for the newer, more up to date programming
languages because they think that's what they need.

--
D'Arcy J.M. Cain (da...@druid.com) |
PlanIX, Inc. | There's no government
Toronto, Ontario, Canada | like no government!
+1 416 424 2871 (DoD#0082) (eNTP) |

Larry Pyeatt

unread,
Jan 12, 1994, 1:04:33 PM1/12/94
to

In article <CJGrC...@dcs.ed.ac.uk>, c...@dcs.ed.ac.uk (Colin Simpson) writes:
|>
|> Personally, I find GUI's useful for doing certain things such as selective file
|> copies, the visual previewer and selectors in xv3 and reading news is (I find)
|> easier under the X reader than the command line. I separately find the command
|> line easier to use for other things. My ideal environment is a GUI with a shell
|> window(s) so I can have the best of both worlds.

Amen, brother! That is the major problem I have with Windoze NT. It does
not provide a decent shell. I know, there are $econd party $hell$ for
NT, and that brings me to my second complaint. By the time you add on all
the extras to make NT equal to Unix, you have spent as much money as it
would cost to buy a commercial version of Unix, and you have several vendors
to deal with instead of just one. I have played the "It's not our fault, it's
company Z!" game before. No, thanks.
--
Larry D. Pyeatt All standard disclaimers apply.
pye...@cs.colostate.edu Void where prohibited.

Brian Powell

unread,
Jan 12, 1994, 1:39:35 PM1/12/94
to
Sayeth Chris Gray:
:
: This is the *only* reason I intend to study C++ one day, ...
: I suppose it is also the reason

: why so many poor souls are having C++ thrust upon them as a first language
: for teaching purposes.

I really don't see what's so bad about C++. It has a lot of nice,
compile time continuity checking. It is very flexible, with
overloading, etc. It is very modular (read: object oriented) for easy
code reuse. It is easy to write fairly bullet proof code. I like C
too, but C++ is not the monster that some of you portray it as...

Actually, OSU uses Modula-2 (yuck) as their first teaching language.

Brian Powell

unread,
Jan 12, 1994, 1:31:21 PM1/12/94
to
Sayeth Mark Justin Cecil:
:
: It may be a fact that you see more and more C++ ads, but keep in mind that

: the industry bought into C++ in A BIG WAY, and they will simply try to
: protect their investment by continuing to hire people proficient in the
: language. This doesn't necessarily mean that the industry is happy with C++,
: just that they spent a lot of money on it and they aren't willing to give it
: up just because it isn't what they expected.

Well, I guess the next question that begs to be asked is: Who cares?

As a programmer (employee or prospective employee), I don't really
care *why* an employer is using a certain language, just that they
*are* using it and *will continue* to use it for a while. I also care
that enough employers are using it to justify specializing in it.
Your above statements, while possibly true, are irrelevant to the
situation.

Brian Powell

unread,
Jan 12, 1994, 1:49:09 PM1/12/94
to
Sayeth Anton Rang:

: In article <2guvof$3...@mane.cgrg.ohio-state.edu> bpo...@osc.edu (Brian Powell) writes:
: >Do you realize just how many shrink wrapped apps are available for a
: >popular UNIX version like Sun Solaris?

: Do you have specific numbers? I find it hard to believe that the


: number of applications for any OS (like Solaris) with a tiny installed
: base (as compared to {Mac/MS-DOS/Win3.x}) is that large.

I was waiting for someone to ask for numbers :-) I had a UNIX
software catalog around here with hundreds of apps for Solaris, but I
can't find it now. When I do, I'll post some numbers. In any case,
the number of apps will be far less than those for DOS/Win3. But,
we're talking about NT here, not DOS. Solaris can have just as good
16 bit DOS/Windows (maybe better?) emulation as NT has.

: ... The installed base of any particular UNIX version that I know


: of -- how big is the Solaris 1.x base? 2.x? -- is small compared to
: the 10+ million Macintoshes, or the even larger PC market.

But, how much of the installed PC base can and will actually run NT?
A *very* small slice of it to be sure. The installed base of Solaris
is larger than NT and very well may remain so.

: ... As it is, even the small utilities I've done -- text-based,


: some networking involved -- have hundreds of lines of conditional

: compilation to get them to work...

No argument with this part :-), except that hopefully it is changing
for the better.

Brian Powell

unread,
Jan 12, 1994, 2:10:37 PM1/12/94