Path: sparky!uunet!usc!nic.csu.net!csus.edu!wupost!zaphod.mps.ohio-state.edu!
rpi!sarah!cook!psinntp!psinntp!crynwr!nelson
Newsgroups: comp.binaries.ibm.pc.d,comp.os.msdos.programmer,comp.sys.ibm.pc.misc,
comp.windows.ms.programmer
Subject: DOS is dead.
Message-ID: <703214134snx@crynwr.com>
From: nel...@crynwr.com (Russell Nelson)
Date: Tue, 14 Apr 92 01:15:34 GMT
Followup-To: comp.os.msdos.programmer
Distribution: world
Organization: Crynwr Software
Lines: 29

DOS is dead.  (At most) A year from now no one will be programming for MS-DOS.

What will spur the abandonment of DOS?  Well, people have long
acknowledged that DOS is not a real operating system.  The only
reason to run DOS is because you can't run a real operating system.
We now have the metal to run those operating systems.  Because the
benefits of a real operating system will outweigh the cost, people
will run them on their new machines.

That leaves DOS only running on 8088s and 80286's.  These machines
will be worth very little and so will be purchased by people on tight
budgets (or donated to nonprofits, etc).  There will be not much
money to spend on software.  But of course, we know the cost of
software development is not going down like hardware costs.

So, the only people buying expensive new software will be people
running the new real OS(es).  Programmers interested in making the
big bucks will program to this new audience.  And it *won't* be
Windows, which is as lame an OS as DOS is.

Oh yes, there will be *some* people programming for MS-DOS, just like
there are still new programs being written for the Zenith Z-100,
Tandy Color Computer, Commodore C-64, CP/M, etc.  But there will not
be much money in it.

-russ <nel...@crynwr.com>  I'm proud to be a humble Quaker!
Crynwr Software            Crynwr Software sells packet driver support.
11 Grant St.               315-268-1925 Voice
Potsdam, NY 13676          315-268-9201 FAX

Newsgroups: comp.os.msdos.programmer
Path: sparky!uunet!europa.asd.contel.com!darwin.sura.net!ukma!news
From: kev...@primitives.inslab.uky.edu (Kevin Solie)
Subject: Re: DOS is dead.
References: <703214134snx@crynwr.com>
Nntp-Posting-Host: primitives.inslab.uky.edu
Message-ID: <1992Apr17.211541.11058@ms.uky.edu>
Organization: University Of Kentucky, Dept. of Math Sciences
Sender: n...@ms.uky.edu (USENET News System)
Date: Fri, 17 Apr 1992 21:15:41 GMT
X-Bytes: 
Lines: 13

In article <703214134...@crynwr.com> nel...@crynwr.com (Russell Nelson) writes:
> DOS is dead.  (At most) A year from now no one will be programming for  
MS-DOS.
I beleive this is true but a year is pushing it. The DOS carcass will float
around for a long time.
--
+-----------------------------------------------------------------------------+
|            One more day like today and I'll kill you                        |
|                                    -Robert Smith                            |
|                                                                             |
|                                              Kevin Solie                    |
+-----------------------------------------------------------------------------+

Path: sparky!uunet!polari!rwing!fnx!nazgul!bright
From: bri...@nazgul.UUCP (Walter Bright)
Newsgroups: comp.os.msdos.programmer
Subject: 16 bit DOS is dead.
Message-ID: <671@nazgul.UUCP>
Date: 19 Apr 92 19:45:56 GMT
References: <703214134snx@crynwr.com> <1992Apr17.211541.11058@ms.uky.edu>
Reply-To: bri...@nazgul.UUCP (Walter Bright)
Organization: Zortech, Seattle
Lines: 23

In article <1992Apr17.211541.11...@ms.uky.edu> kev...@primitives.inslab.uky.edu 
(Kevin Solie) writes:
/In article <703214134...@crynwr.com> nel...@crynwr.com (Russell Nelson) writes:
/> DOS is dead.  (At most) A year from now no one will be programming for  
/MS-DOS.
/I beleive this is true but a year is pushing it. The DOS carcass will float
/around for a long time.

I'll be so bold as to predict that by year's end no new projects will be
started that are targeted at 16 bit DOS. Writing 32 bit programs for DOS
is easier and easier, and more and more companies are realizing this and
are designing new code with the assumption of 32 bit processors and
capabilities.

With megabytes of memory and 386s so cheap, who in their right mind would
put a major investment into a new programming project for 8088 and 286
machines? A lot of today's "16 bit" programs essentially require at least
a 386 anyway, since they perform abysmally on a 286. Why not take the final
step and make them true 32 bit programs?

In my experience, a program that fiddles a lot with pointers and data
structures will run about twice as fast compiled for 32 bits as the same
code compiled for 16 bit large memory model. I/O bound and floating point
intensive programs run about the same speed.

Path: sparky!uunet!spool.mu.edu!umn.edu!csus.edu!netcomsv!mork!jsteele
From: jste...@netcom.com (John Steele)
Newsgroups: comp.os.msdos.programmer
Subject: Re: 16 bit DOS is dead.
Message-ID: <lb6j8r.jsteele@netcom.com>
Date: 21 Apr 92 21:49:42 GMT
Article-I.D.: netcom.lb6j8r.jsteele
References: <703214134snx@crynwr.com> <1992Apr17.211541.11058@ms.uky.edu> 
<671@nazgul.UUCP>
Organization: Netcom - Online Communication Services  (408 241-9760 guest)
Lines: 67

In <6...@nazgul.UUCP> bri...@nazgul.UUCP (Walter Bright) writes:

>In article <1992Apr17.211541.11...@ms.uky.edu> kev...@primitives.inslab.uky.edu 
(Kevin Solie) writes:
>/In article <703214134...@crynwr.com> nel...@crynwr.com (Russell Nelson) writes:
>/> DOS is dead.  (At most) A year from now no one will be programming for  
>/MS-DOS.
>/I beleive this is true but a year is pushing it. The DOS carcass will float
>/around for a long time.

>I'll be so bold as to predict that by year's end no new projects will be
>started that are targeted at 16 bit DOS. Writing 32 bit programs for DOS

Chalk up one failed prediction (an inherent problem with predictions, BTW).

>is easier and easier, and more and more companies are realizing this and
>are designing new code with the assumption of 32 bit processors and
>capabilities.

Do you really believe that *users* give a $&!* how easy it is to develop
their applications?

>With megabytes of memory and 386s so cheap, who in their right mind would
>put a major investment into a new programming project for 8088 and 286
>machines? A lot of today's "16 bit" programs essentially require at least
>a 386 anyway, since they perform abysmally on a 286. Why not take the final
>step and make them true 32 bit programs?

I think you are missing a MAJOR driving force behind software development,
namely business.  There are still *very many* small businesses without the
capital needed to upgrade all their machines.  Do you have any idea of the
number of XT's attached to small networks serving as terminals for the
retail arena?  I assure you the numbers are large, and they serve the
purposes of those businesses just fine.

>In my experience, a program that fiddles a lot with pointers and data
>structures will run about twice as fast compiled for 32 bits as the same
>code compiled for 16 bit large memory model. I/O bound and floating point
>intensive programs run about the same speed.

This is all fine and dandy from a developer's point of view, but has very
little to do with the business-owners' bottom line.  If they *require* the
speed improvement, and have the capital available, then they will purchase
faster equipment, but only if.

Now don't get me wrong, I would recommend (for many of the same reasons
that you give above) new purchasers to get the newer machines, but there
is still the current equipment investment, and you won't find many
successful businessmen who will just 'throw that investment away'.

From a developer's standpoint, I see no real reason why portable code
doesn't serve the same purpose that it always had, namely creating
products that are available on multiple platforms (including those old
MS-DOS 2.1 XT's :)

Not to flame, but your arguments sound very much the same as the one's I
heard 10 years ago (mainly from naive CS students) that didn't have the
common sense to take a few business courses, or get actual experience out
there 'in the trenches'.

John ("Oh yeah, the 370 won't last another 6 months", call me skeptical)
-- 
#include <disclaimer.h>			/* route all flames to /dev/null  */
/**************************************************************************/
/*   Systems Analyst      | John Steele     email: jste...@netcom.com     */
/* Video Business Systems |                                               */
/*   1-800-255-3088       |    Most people spell COBOLSUX incorrectly...  */
/**************************************************************************/

Path: sparky!uunet!polari!rwing!fnx!nazgul!bright
From: bri...@nazgul.UUCP (Walter Bright)
Newsgroups: comp.os.msdos.programmer
Subject: Re: 16 bit DOS is dead.
Message-ID: <684@nazgul.UUCP>
Date: 22 Apr 92 17:38:35 GMT
References: <703214134snx@crynwr.com> <1992Apr17.211541.11058@ms.uky.edu> 
<671@nazgul.UUCP> <lb6j8r.jsteele@netcom.com>
Reply-To: bri...@nazgul.UUCP (Walter Bright)
Organization: Zortech, Seattle
Lines: 54

In article <lb6j8r.jste...@netcom.com> jste...@netcom.com (John Steele) writes:
/In <6...@nazgul.UUCP> bri...@nazgul.UUCP (Walter Bright) writes:
/>I'll be so bold as to predict that by year's end no new projects will be
/>started that are targeted at 16 bit DOS. Writing 32 bit programs for DOS
/Chalk up one failed prediction (an inherent problem with predictions, BTW).

	Don't count it dead yet, we still got a few months to go!

/>is easier and easier, and more and more companies are realizing this and
/>are designing new code with the assumption of 32 bit processors and
/>capabilities.
/Do you really believe that *users* give a $&!* how easy it is to develop
/their applications?

	Indirectly, yes. The easier it is to develop applications, the
	quicker the user gets them. The user does care a lot about time
	to market, for evidence, look at all the complaints about how
	software lags hardware by *years*.

/I think you are missing a MAJOR driving force behind software development,
/namely business.  There are still *very many* small businesses without the
/capital needed to upgrade all their machines.  Do you have any idea of the
/number of XT's attached to small networks serving as terminals for the
/retail arena?  I assure you the numbers are large, and they serve the
/purposes of those businesses just fine.

	I don't doubt that there are lots of XT's serving as terminals.
	I do doubt that they will form a market for new applications.
	Note that I also said "starting" a new project. When you fire
	up a new project, you should be predicting what your market will
	be when the project is finished, not what it is when you start.

/From a developer's standpoint, I see no real reason why portable code
/doesn't serve the same purpose that it always had, namely creating
/products that are available on multiple platforms (including those old
/MS-DOS 2.1 XT's :)

	Because applications get bigger and bigger due to customer demand
	for more and more features, it is becoming less and less practical
	to expend the development effort to make it compatible with an XT.
	If doing so adds 6 months to your schedule, that's 6 months of lost
	sales. And I suggest that people using XT's will not spend $200-$900
	on new software.

/Not to flame, but your arguments sound very much the same as the one's I
/heard 10 years ago (mainly from naive CS students) that didn't have the
/common sense to take a few business courses, or get actual experience out
/there 'in the trenches'.

	I've been in the PC software business for 10 years.
	I watched the 8 bit CPM business disappear in a remarkably
	short time. Everyone was surprised at how fast it failed.
	Of course, hindsight being 20/20, everyone says how they
	predicted it :-)

Newsgroups: comp.os.msdos.programmer
Path: sparky!uunet!usc!rpi!scott.skidmore.edu!pvonk
From: pv...@scott.skidmore.edu (Pierre VonKaenel)
Subject: Re: DOS is dead.
Message-ID: <1992Apr22.200741.18251@scott.skidmore.edu>
Organization: Skidmore College, Saratoga Springs NY, 12866
References: <1992Apr22.173215.26221@apollo.hp.com>
Distribution: usa
Date: Wed, 22 Apr 1992 20:07:41 GMT
Lines: 34


Someone posted that they thought Russ Nelson asked about DOS' demise
as a means to get reaction, not to predict.  Since he's at an
institution of higher learning, I wonder if Clarkson is currently
running Windows or OS/2 on their PCs.  Being an engineering school,
they just might have the funds to sink into newer computer technology.
But at many liberal arts schools, I wonder how many with DOS machines
have gone to Windows.  In our case, we do have most of the hardware
(55-SX computers for the most part), but we don't have mice!  The one
or two we did have are mysteriously gone (it was just reported
yesterday that someone walked off with a Sun mouse and pad!, I'd love
to see their face when they try to hook it to a Mac or PC..).  We have
tried security kluges, but even then, students will cut the mouse wire
to remove the rodent.

This is only one problem.  There are also software maintenance
problems for PCs in public user rooms.  Etc, etc.  It may take a while
before schools opt for newer OSes that require more hardware/software.
In addition, there are a lot of faculty who write DOS software to
support the curriculum.

As for high schools.... our college students report that computers in
9-12 grades consist of Apple IIs and GS models, along with other
older-generation hardware.

I suspect DOS will continue for a number of years, albeit in narrow
segments of our society.


-- 
Pierre von Kaenel	| Skidmore College	     | pv...@scott.skidmore.edu
Math & CS Dept.		| Saratoga Springs, NY 12866 | (518)584-5000 Ext 2391
"Behold the turtle.  He makes progress only when he sticks his head out." 
 - J. Conant

Path: sparky!uunet!psinntp!crynwr!nelson
From: nel...@crynwr.com (Russell Nelson)
Newsgroups: comp.os.msdos.programmer
Subject: DOS is dead. 
Distribution: world
Message-ID: <704033211snx@crynwr.com>
References: <1992Apr22.200741.18251@scott.skidmore.edu>
Date: Thu, 23 Apr 92 12:46:51 GMT
Organization: Crynwr Software
Lines: 43

In article <1992Apr22.200741.18...@scott.skidmore.edu> pv...@scott.skidmore.edu 
writes:

  > Someone posted that they thought Russ Nelson asked about DOS' demise
  > as a means to get reaction, not to predict.

That was Timo.  No, I really meant to predict, I stand by my prediction, but
I clarified it in another article.

  > Since he's at an
  > institution of higher learning,

Not since last July.

  > I wonder if Clarkson is currently
  > running Windows or OS/2 on their PCs.  Being an engineering school,
  > they just might have the funds to sink into newer computer technology.

Yes, and instead of upgrading the student PCs so that *all* their
students could run a real OS, they continue to purchase antique PCs
(IBM PS/2 Model 30/286) so that IBM would donate 110 RS/6000s.
Attack of the mainframe syndrome.  I told them not to do it, but did
they listen to me?  Noooooooo!

Undoubtedly they're going to purchase *more* machines that are
obsolete the day they're shipped, much less four years from now when
the students graduate.  And, being IBM, they won't be upgradable.  :(

  > We have tried security kluges, but even then, students will cut
  > the mouse wire to remove the rodent.

That's pretty sad, considering that you can get a mouse for $10.
Perhaps you should sell them in your bookstore for that price.  Throw
in a free sixpack and you'll sell thousands.  :)

  > I suspect DOS will continue for a number of years, albeit in narrow
  > segments of our society.

Um, that's approximately what I said in my original posting.

-russ <nel...@crynwr.com>  I'm proud to be a humble Quaker!
Crynwr Software            Crynwr Software sells packet driver support.
11 Grant St.               315-268-1925 Voice
Potsdam, NY 13676          315-268-9201 FAX

Path: sparky!uunet!psinntp!crynwr!nelson
From: nel...@crynwr.com (Russell Nelson)
Newsgroups: comp.os.msdos.programmer
Subject: 16 bit DOS is dead. 
Distribution: world
Message-ID: <704033822snx@crynwr.com>
References: <684@nazgul.UUCP>
Date: Thu, 23 Apr 92 12:57:02 GMT
Organization: Crynwr Software
Lines: 26

In article <6...@nazgul.UUCP> bri...@nazgul.UUCP writes:

  >         Because applications get bigger and bigger due to customer demand
  >         for more and more features,

also customer demand for more and more simplicity.  That's what GUI's
are all about -- spending machine cycles to make the program easier to use.

  >         If doing so adds 6 months to your schedule, that's 6 months of lost
  >         sales.

And six months of extra programmer's salaries and other overhead.
Time to market is *very* important.

  >         And I suggest that people using XT's will not spend $200-$900
  >         on new software.

Right, I made that point in my original article that started this
thread.  The converse of that is also true: If you're going to spend
$500 on PageMaker or Designer or AutoCAD, you're going to first
upgrade or buy the appropriate hardware to run it on.

-russ <nel...@crynwr.com>  I'm proud to be a humble Quaker!
Crynwr Software            Crynwr Software sells packet driver support.
11 Grant St.               315-268-1925 Voice
Potsdam, NY 13676          315-268-9201 FAX