Why Windows will dominate
Generic Account 1670
OS/2 vs. NT. Apples and oranges. SMP, Multiple message queues, etc. And OS/2 for Workplace is not here. No comparison.
History will repeat itself. Remember OS/2 2.0 only supporting Windows 3.0? Windows left it in its wake. OS/2 2.1 will have the same fate unless they can find support for the next Windows (Chicago, Windows 4.0 who cares). Is OS/2 that hot? NO! Even DOS outsells it. Go figure.
OS/2 applications are few and far between. Lotus products are slow and not equal to their Windows counterpart. Wordperfect abruptly dropped Wordperfect 6.0 without even fixing the problems of 5.2! Borland is concentrating its efforts on Windows, leaving their Borland C++/2 bug-ridden. If OS/2 has the potential user base, why is Borland staying in the WIndows market where Microsoft is killing it? Because OS/2 is a waste of time and money.
OS/2 compatible with Windows 3.1? NO! Run full Norton Desktop, PC Tools for Windows, Mathcad, Visual C++, Autocad, DoubleSpace, etc.. Stacker for OS/2 sucks,it can't even handle OS/2s HPFS and if you want to use it in DOS, be prepared to spend more money.
OS/2 running with 4mb. Dream on. Multitasking to the hard disk more like it. The real requirements are 8mb.
5 million OS/2 users? Where? From what planet? So many of you question the sales of Windows, I question the user base of OS/2. I think Wordperfect and Borland found this out in the pocket book. But, Worperfect is working on some new OS/2application? And you believe them?
Chicago too late? Who cares? OS/2 has to make headway against Windows 3.1 which is here! DOS even outsells OS/2.
It's ironic that IBM has spent billions developing OS/2, only to sell 5 million copies (your numbers). A substantial portion of that went to Microsoft. Bad business. Then they waste their time on OS/2 for Windows. Why didn't they spend their time on OS/2 applications. I doubt they can, look at their wonderful OS/2 applets. Moreover, how many OS/2 users buy Windows applications so they can make OS/2 productive? Good business for Microsoft. How many bought OS/2 because it supports Windows? Good busines
s for Microsoft. How many OS/2 users have deleted DOS+Windows, only to install OS/2s version of DOS and Windows? Good business for Microsoft.
IMHO, Microsoft let Windows stagnate while the divorce with IBM became final. Microsoft will bring their future Operating Systems to par with their applications. Now, that IBM is off their back, we'll see MS can really do.
OS/2, from the makers of DOS 4.0. How did MS slip that 640k barrier in DOS without IBM's knowledge? Is Bill Gates that smart?
Not a member of Team OS/2!
Rudynell S Millian
>First of all, Windows is not an operating system,
>
press's Computer Dictionary, first edition, examples of Operating Systems
are "DOS, OS/2, Unix".
>History will repeat itself. Remember OS/2 2.0 only supporting Windows 3.0?
>Windows left it in its wake. OS/2 2.1 will have the same fate unless they
>can find support for the next Windows (Chicago, Windows 4.0 who cares).
>Is OS/2 that hot? NO! Even DOS outsells it. Go figure.
^^^^^^
Huh?! Don't you need DOS to run Windows? Go figure. I think you will
find that the for any given period, number of coppies of DOS sold is more or
less the number of coppies of Windows sold.
>
>OS/2 running with 4mb. Dream on. Multitasking to the hard disk more like it. >The real requirements are 8mb.
>
Steep resource requirement is only a liability when it comes to OS/2,
right? NT's resource requirements are double that of OS/2, yet in the case
of NT this is pat of its alure. Go figure. Then again, according to yourself
I am comparing apples with oranges.
>
>Chicago too late? Who cares? OS/2 has to make headway against Windows 3.1
>which is here! DOS even outsells OS/2.
^^^^
Again?! Whatch this; For any given amount of headway OS/2 makes
against Windows 3.1, the same amount of headway will have been made against
DOS. Go figure.
>
>It's ironic that IBM has spent billions developing OS/2, only to sell 5
>million copies (your numbers).
>
Billions, as in billions of half pennies, right? Now we know where
the bulk of IBM's losses went.
>
>IMHO, Microsoft let Windows stagnate while the divorce with IBM became final. >Microsoft will bring their future Operating Systems to par with their
>applications. Now, that IBM is off their back, we'll see MS can really do.
>
>OS/2, from the makers of DOS 4.0.
Well, it's a good thing OS/2 is not from the makers of DOS 6.0!
> How did MS slip that 640k barrier in DOS without IBM's knowledge?
> Is Bill Gates that smart?
>
You think that the 640K barrier is smart? I guess you'll be happy to
know that it was present in DOS 5.0 and, by popular demand,is still present in
DOS 6.0 and DOS 6.2.
>Not a member of Team OS/2!
Phew! Thank goodness.
Not a membet of Team OS/2 either!
Mutant for Hire
goes a long way towards making a posting look more convincing.
In article <2g1ti0$s...@gondor.sdsu.edu>, masc...@ucssun1.sdsu.edu (Generic Account 1670) writes:
>First of all, Windows is not an operating system, it's the best and
>most productive application ever sold! This is why Windows is the
>solution for the majorityof the people. Quit comparing OS/2 to
>Windows. And Chicago is not here.
Whether Windows is an OS or not is rather tricky. I've seen convincing
arguments towards the "Windows is an OS" statement and the "Windows is
a DOS shell" statement. In neither case can it really be called an
application. In my opinion, its 95% an OS, so I'll just round up and
call it an OS.
>OS/2 vs. NT. Apples and oranges. SMP, Multiple message queues, etc.
>And OS/2 for Workplace is not here. No comparison.
BTW, that should be Workplace OS and not OS/2 for Workplace. As for
that extra functionality that NT has, OS/2 plus the appropriate
network software seems to match most of the functionality except for
SMP and multiple message queues. The latter is a small omission, but
how many people out there actually own SMP machines?
>History will repeat itself. Remember OS/2 2.0 only supporting
>Windows 3.0? Windows left it in its wake. OS/2 2.1 will have the same
>fate unless they can find support for the next Windows (Chicago,
>Windows 4.0 who cares). Is OS/2 that hot? NO! Even DOS outsells it.
>Go figure.
Question is, how many people will go to Windows 4.0, considering the
mass of Windows 3.1 software out there? How many people will go on a
massive uprgrade spree of all their 3.1 software to run under Windows
4.0, spending all that money in the process. There wasn't a huge
installed base of Windows applications that Windows 3.1 broke. That
is no longer true.
>OS/2 running with 4mb. Dream on. Multitasking to the hard disk more
>like it. The real requirements are 8mb.
Indeed, and you are willing to believe that Windows 4.0 will run in
less than 8mb? If not, then how will it take off where OS/2 has not?
>5 million OS/2 users? Where? From what planet? So many of you
>question the sales of Windows, I question the user base of OS/2. I
>think Wordperfect and Borland found this out in the pocket book. But,
>Worperfect is working on some new OS/2application? And you believe
>them?
The problem is with Windows sales is that Windows is bundled with lots
of machines, software and periphials. If your machine is bundled with
Windows and you only run DOS, it still counts as a Windows sale.
Owning multiple copies of Windows while you use only one counts as
multiple Windows sales, while the user base is smaller. So the
question is, not how many Windows sales are there, but how many Window
users compared to the sales.
The difference with OS/2 is that people have to go out and buy the
thing, thus there is less difference between the number of sales and
the size of the base of people using it. As for software companies
having problems with OS/2 sales, given the complaints you raised
against the applications, maybe they sold poorly because they weren't
well written.
>Chicago too late? Who cares? OS/2 has to make headway against
>Windows 3.1 which is here! DOS even outsells OS/2.
DOS Sales > Windows Sales > OS/2 sales was a given for quite a long
time because you needed DOS to run Windows, and each OS/2 sale
counted as a Windows sale. Only recently has that been altered as OS/2
now has a version that runs along side Windows rather than
incorporating it.
Also bear in mind that Windows 4.0 is going to have to go up against
Windows 3.1 and DOS. Windows is its own worst enemy at this point in
terms of upgrades. Windows 1.X and Windows 2.X weren't big enough to
make problems, but Windows 3.1 is certainly large enough in the
installed base that even Windows 4.0 is going to have problems because
of it.
>It's ironic that IBM has spent billions developing OS/2, only to sell
>5 million copies (your numbers). A substantial portion of that went
>to Microsoft. Bad business. Then they waste their time on OS/2 for
>Windows. Why didn't they spend their time on OS/2 applications. I
>doubt they can, look at their wonderful OS/2 applets. Moreover, how
>many OS/2 users buy Windows applications so they can make OS/2
>productive? Good business for Microsoft. How many bought OS/2 because
>it supports Windows? Good business for Microsoft. How many OS/2
>users have deleted DOS+Windows, only to install OS/2s version of DOS
>and Windows? Good business for Microsoft.
First, you haven't specified an exact number for IBM's investment in
OS/2, and you should give where you got those numbers as well. Also
how much of that money went to Microsoft. There are royalties yes, but
how great a percentage that is, is another issue. Also, IBM really
isn't an applications house the way that Microsoft is, they're more
into hardware, OS, and OS support software like networking. Attacking
IBM for not writing spreadsheets is silly.
>IMHO, Microsoft let Windows stagnate while the divorce with IBM
>became final. Microsoft will bring their future Operating Systems to
>par with their applications. Now, that IBM is off their back, we'll
>see MS can really do.
You mean amazingly brilliant software like Windows NT? Its also
interesting to notice that OS/2 took a significant leap in quality
after Microsoft was no longer writing code for it. Also, that single
message queue that you criticized in OS/2 (well, indirectly, because
you praised NT for not having one) was one of Microsoft's ideas. I
have far more faith in IBM's ability to write OS's than Microsoft's.
Windows NT is a great example of diminished expectations.
--
Martin Terman, Mutant for Hire, Synchronicity Daemon, Priest of Shub-Internet
Disclaimer: Sticks and stones may break my bones, but flames are just ignored
mfte...@phoenix.princeton.edu mfte...@pucc.bitnet anonym...@charcoal.com
"Sig quotes are like bumper stickers, only without the same sense of relevance"
Harry Myhre
Account 1670) wrote:
>IMHO, Microsoft let Windows stagnate while the divorce with IBM became final.
Microsoft will bring their future Operating Systems to par with their
applications.
They'd better do more than that. Most of their apps are pretty mediocre in
my opinion. Serviceable, but mediocre.
---
Harry Myhre <har...@netcom.com>
Ted Holden
>Microsoft will bring their future Operating Systems to par with their
>applications. Now, that IBM is off their back, we'll see MS can really do.
We've SEEN what MS can really do: they've REALLY fucked up. They've
brought out the wrong product at the wrong time; NT not only doesn't
compete with OS/2, it doesn't compete with UNIX and/or Netware (as it is
now being hyped) very well either. NT is apparently meant to compete
with VMS, which says that MS needed NT around 1977 or so.
The rest of MS's products are pure vapor at this point. Chicago in
particular appears to be mostly 16-bit code which will be lucky to make
it out in 94 in its present state. It goes downhill from there.
I know I wouldn't want to be competing against IBM with hot air against
real products for a whole year...
Generic Account 1670
: In article <2g1ti0$s...@gondor.sdsu.edu> masc...@ucssun1.sdsu.edu (Generic Account 1670) writes:
: >History will repeat itself. Remember OS/2 2.0 only supporting Windows 3.0?
: >can find support for the next Windows (Chicago, Windows 4.0 who cares).
: >Is OS/2 that hot? NO! Even DOS outsells it. Go figure.
: ^^^^^^
: Huh?! Don't you need DOS to run Windows? Go figure. I think you will
: find that the for any given period, number of coppies of DOS sold is more or
: less the number of coppies of Windows sold.
Precisely my point, Windows is not an operating system.
: Steep resource requirement is only a liability when it comes to OS/2,
: right? NT's resource requirements are double that of OS/2, yet in the case
: of NT this is pat of its alure. Go figure. Then again, according to yourself
: I am comparing apples with oranges.
As I've said apples and oranges.
: >
: >Chicago too late? Who cares? OS/2 has to make headway against Windows 3.1
: >which is here! DOS even outsells OS/2.
: ^^^^
: Again?! Whatch this; For any given amount of headway OS/2 makes
: against Windows 3.1, the same amount of headway will have been made against
: DOS. Go figure.
Go figure. It's 2 products to 1.
: >
: >It's ironic that IBM has spent billions developing OS/2, only to sell 5
: >million copies (your numbers).
: >
: Billions, as in billions of half pennies, right? Now we know where
: the bulk of IBM's losses went.
I meant over a billion dollars. This is fact!
: Well, it's a good thing OS/2 is not from the makers of DOS 6.0!
Say what you will, DOS 4.0 is the worst DOS ever made.
: > How did MS slip that 640k barrier in DOS without IBM's knowledge?
: > Is Bill Gates that smart?
: >
: You think that the 640K barrier is smart? I guess you'll be happy to
: know that it was present in DOS 5.0 and, by popular demand,is still present in
: DOS 6.0 and DOS 6.2.
I am simply stating that Microsoft did not impose this 640k barrier without
IBM's knowledge.
John H. Kim
Generic Account 1670 <masc...@ucssun1.sdsu.edu> wrote:
>
>It's ironic that IBM has spent billions developing OS/2, only to sell
>5 million copies (your numbers). A substantial portion of that went
>to Microsoft. Bad business.
I don't think it's that simple. IBM also sells hardware, in fact,
I think the Personal Systems division makes most of its money off
hardware sales. When IBM was working on OS/2 1.3 and 2.0, even
though their PC sales were going up in raw numbers, PC market
share was going down compared to Macs and workstations. I think,
and I'm pretty sure IBM thought the same thing, that DOS as *the*
operating system for PCs was showing its age. IBM needed a new
operating system to put the PC on par with competing hardware.
Whether OS/2 did this or not was irrelevant. But by continuing
to push OS technology with OS/2, IBM forced Microsoft to push ahead
with a better DOS, a better Windows, and other companies started
pushing their new OSes. Things began to look upbeat on the PC OS
market. And while it may not be entirely due to this reason, IBM
regained from Apple the title of #1 personal computer seller last
year.
--
John H. Kim | (This space to be filled when I think of
jo...@jarthur.cs.hmc.edu | something clever to use as a disclaimer)
James Antoniou
: I meant over a billion dollars. This is fact!
I am curious where you got this figure. Not that I necessarily don't
believe you, but I am just real curious. Please give publication, month,
page #, etc.
: Say what you will, DOS 4.0 is the worst DOS ever made.
I don't understand this. You're saying Microsoft had nothing whatsoever
to do with DOS 4.0? Surely not... You must just mean that DOS 4.0 was
terrible in general, a point on which no one would disagree with you.
Generic Account 1670
: >In article <2g1ti0$s...@gondor.sdsu.edu> masc...@ucssun1.sdsu.edu
: (Generic Account 1670) writes:
: brought out the wrong product at the wrong time; NT not only doesn't
: compete with OS/2, it doesn't compete with UNIX and/or Netware (as it is
: now being hyped) very well either. NT is apparently meant to compete
: with VMS, which says that MS needed NT around 1977 or so.
: The rest of MS's products are pure vapor at this point. Chicago in
: particular appears to be mostly 16-bit code which will be lucky to make
: it out in 94 in its present state. It goes downhill from there.
: I know I wouldn't want to be competing against IBM with hot air against
: real products for a whole year...
Let's turn this around.
We've SEEN what IBM can really do: they've REALLY fucked up. They've
brought out the wrong product at the wrong time. ^^ can we use this word?
Generic Account 1670
: In Article <2g1ti0$s...@gondor.sdsu.edu>, masc...@ucssun1.sdsu.edu (Generic
: ---
: Harry Myhre <har...@netcom.com>
Mediocre? No comment.
Generic Account 1670
: Generic Account 1670 (masc...@ucssun1.sdsu.edu) wrote:
: : Say what you will, DOS 4.0 is the worst DOS ever made.
: I don't understand this. You're saying Microsoft had nothing whatsoever
: to do with DOS 4.0? Surely not... You must just mean that DOS 4.0 was
: terrible in general, a point on which no one would disagree with you.
IBM headed the development of DOS 4.0 separate from Microsoft. My point is
IBM has screwed up also. Windows NT receives much criticism, but this is
a stable operating system considering it is first generation. Many argue
that the underlying foundation of NT is flawed, but this remains to be seen.
I have found NT much more stable for development than both OS/2 and DOS+Windows.
I will be the first to admit NT is not suitable for the end user, but this does
not make it inferior to OS/2.
I should not have taken the cheap shot at IBM. I got caught up in some of the
stupid arguments on Windows vs OS/2. Enough said.
nightmare
I've only used DOS 4.01, so what is the scoop on DOS 4.0? As I
remember 4.01 was only a memory hog and the ony improvement was
larger than 32M partitions. And the *gulp* DOSSHELL.EXE!!
Adrian
51g...@cua.edu
What's your point? M$ has done a lot of things "with IBM's knowledge,"
that doesn't mean IBM encouraged it, wanted it, or would have done it the same
way. That 640k barrier will stand forever as testimony to M$'s
short-sightedness and IBM's "don't ask and it'll work out OK" attitude. IBM
seems to be changing. Can you say the same of M$'s OS division?
Todd Louis Green
Team OS/2
merlin
>> : > Is Bill Gates that smart?
The 640K limit was a hardware limitation in the original IBM PC design
created by IBM designers -- not a terrible secret limitation of DOS --
and at the time the IBM PC was initially put on the market most of the
systems were sold with 64K to 256K memory -- few people could afford a
full 640K system. It was inconceivable at the time that anyone using
a personal computer could possibly need more than 640K for any purpose.
51g...@cua.edu
Issue of vulgarities aside, what kind of dumb-ass statement is that?
IBM's OS/2 uses the 386-based processor for what it was designed for, and does
it well. Windoze uses it as a "real fast" 286, and NT uses it as a "real slow"
686. Sorry. IBM's product is *exactly* right for the time. The only reason
it isn't more popular is because M$ spends as much money bad-mouthing OS/2
through it's press slaves as it does hyping its own products.
51g...@cua.edu
As much as it pains me, I've to agree with you here. WinWord 6.0
really is great. Now if Demos only worked . . .
James Antoniou
: IBM headed the development of DOS 4.0 separate from Microsoft. My point is
: IBM has screwed up also. Windows NT receives much criticism, but this is
OK. I can deal with that. I realize that IBM was much more significant
than MS was in the development of DOS 4.0. The implication, however,
that Microsoft had produced no more than 0.0% (i.e., none) of the source
for DOS 4.0 seemed implausible to me. I guess I'm wrong, but it seemed
quite implausible to me that the two companies (when they chose to
co-exist with one another - oh so way back) shared virtually no existing
and new source code for that decidedly shoddy product, in spite of how
separate the development of DOS 4.0 was. But you learn something new
(about the old) every day.
Erik van Linstee
: >> : > How did MS slip that 640k barrier in DOS without IBM's knowledge?
That is no excuse. Granted, it was a hardware limit, but that was
no need for the software to hardcode it too, then, and certainly
no reason why that still is in there 12 years later. If the system
was designed properly to begin with (the software system) there
wouldn't be such dumb limitations.
Don't say, changing the limit later would break applications,
because 1) then those applications are just as ill designed
and 2) the max partition size could be changed too, possibly
breaking a few dumb apps, but adding lots of value with it.
--
ir. E. van Linstee | Delft University of Technology
Man Machine Consultancy | I'll be back ...
----
We are god, 'cause only we can create the idea of his existence
in our holy brains... (Yello)
----
Gentlemen! You can't fight in here. This is the war-room! (P.Sellers)
Wolfgang Strobl
>merlin (mer...@neuro.usc.edu) wrote:
>: >> : > How did MS slip that 640k barrier in DOS without IBM's knowledge?
>: >> : > Is Bill Gates that smart?
>: The 640K limit was a hardware limitation in the original IBM PC design
>: created by IBM designers -- not a terrible secret limitation of DOS --
>: and at the time the IBM PC was initially put on the market most of the
>: systems were sold with 64K to 256K memory -- few people could afford a
>: full 640K system. It was inconceivable at the time that anyone using
>: a personal computer could possibly need more than 640K for any purpose.
>That is no excuse. Granted, it was a hardware limit, but that was
>no need for the software to hardcode it too, then, and certainly
>no reason why that still is in there 12 years later.
The 640K limit isn't hardcoded into MSDOS, and never was. Between '81
and today various vendors produced 8086 and MSDOS based computers
which didn't suffer from the infamous 640K barrier. I used one
myself (a nice fast PC with a very ergonomic and flicker free high
resulution display), which gave almost all of the 1M address space to
the application program, because the display memory wasn't part of
the 1M address space.
Unfortunately, people demanded full IBM hardware compatibility, which forced
systems like this out of the market.
>If the system
>was designed properly to begin with (the software system) there
>wouldn't be such dumb limitations.
If you are talking about MSDOS, then your statement is wrong, because
it doesn't have this limitation.
>Don't say, changing the limit later would break applications,
>because 1) then those applications are just as ill designed
That's true. MSDOS isn't a protected OS, so applications are free
to twiddle with ill designed hardware. The result obviously
is a mess. It's easy to blame the software for it, but it's
even easier to blame the hardware. I believe the whole question
is moot.
>and 2) the max partition size could be changed too, possibly
>breaking a few dumb apps, but adding lots of value with it.
Sure. But just breaking the 640K limit breaks a lot more than
just a few apps, and doesn't add a lot of value (i.e. less than
half of what you already have). Not a very good deal, IMHO.
--
o ( Wolfgang Strobl Wolfgan...@gmd.de (+49 2241) 14-2394
/\ * GMD mbH
_`\ `_<=== Schloss Birlinghoven, #include <std.disclaimer.hpp>
__(_)/_(_)___.-._ 53754 Sankt Augustin, Germany _____________________________
rls...@actcom.co.il
>James Antoniou (ez03...@othello.ucdavis.edu) wrote:
>: Generic Account 1670 (masc...@ucssun1.sdsu.edu) wrote:
>
>I have found NT much more stable for development than both OS/2 and DOS+Windows.
>I will be the first to admit NT is not suitable for the end user, but this does
>not make it inferior to OS/2.
>
I have found OS/2 better and more stable for win16 development, when I
dereference an invalid pointer, during debug and testing, all I have to do
is re-start the VDM in which my program was executing, while my other tools
(All of them win-16 applications, non of them available as win32) continue
to function. This is not the case with NT, one GPF and all my win16
applications are down.
- Ron.
Peter Dilley
> >> : > How did MS slip that 640k barrier in DOS without IBM's knowledge?
> >> : > Is Bill Gates that smart?
>
> The 640K limit was a hardware limitation in the original IBM PC design
Hardware Ohhhhh no my friend. Software!
Both the mac and amiga used offsets to have contigous ram space. It takes
more brains to do it this way. M$loth was on a deadline, since the
original sought after os was cpm wasn't it? but that guy was off sailing
the seas with his new bimbo wife.
------------------------------
a...@qedbbs.com (Peter Dilley) or qed!aga
The QED BBS -- (310)420-9327
Elizabeth Henderson
I have only one thing to say to this argument.
16 povray compiles at once and not a single crc lost on
a download at the same time. Windows 3.1 can barely handle one
povray compile and loses crc when I try to print or open a dos
session (using procomm4win, which will download in the
background pretty well )
Z. Elston
51g...@cua.edu
Gee, I don't know abuot the rest of you, but I've got 16M of ram, and
I'm pretty sure that 640+(0.5*640)<16M. But I was never very good with math.
merlin
>> >> : > How did MS slip that 640k barrier in DOS without IBM's knowledge?
>> >> : > Is Bill Gates that smart?
>>
>> The 640K limit was a hardware limitation in the original IBM PC design
>
>Hardware Ohhhhh no my friend. Software!
>
>Both the mac and amiga used offsets to have contigous ram space. It takes
>more brains to do it this way. M$loth was on a deadline, since the
>original sought after os was cpm wasn't it? but that guy was off sailing
>the seas with his new bimbo wife.
The IBM PC 8088 processor chip can reference a maximum of 1024KB address
space broken into 64KB segments using the segment offsets. The 640K mem
limit was imposed by an IBM decision to reserve all space above the 640K
watermark for the ROM BIOS, ROM BASIC, adaptor ROMs and memory mapped IO.
Gary Kildall, President of Digital Research (makers of CP/M) was said to
be 'out flying' when IBM representatives tried to meet with his company
with the intention of negotiating to purchase the rights to run CP/M on
the original IBM PC/XT. Instead, Gates purchased SCP-DOS from Seattle
Computer Products (Tim Patterson's company) for something like $50,000
total unlimited rights. Tim Patterson was then hired on to help MS make
SCP-DOS into the original DOS 1.0 product released with the IBM PC.
bdu...@tis.kelly.af.mil
I just have to jump in here. The so-called 640K barrier is (and was) NOT a DOS
limitation. It was an IBM-PC limitation. DOS was written with the Intel 8086
in mind, which only has 1Mb of address space. In 1980, when IBM designed the
original PC, fully populated mainframes only had 16Mb of memory, so 1Mb for a
PC (with no real applications at the time) was considered huge. IBM decided to
use some of the address space for other things such as video, BASIC in ROM, BIOS,
etc. With hindsight, it may have not been the best decision, but at the time
it wasn't really that bad either.
When the IBM-PC first came out in 1981, it used PC-DOS and was limited to 640 Mb
of usable RAM for a single program. At about the same time as the IBM-PC Zenith
brought out the Z-100 which had both an 8088 and an 8080 for your choice of Z-DOS
(MS-DOS or Zenith) or CP/M. Z-DOS (or a generic MS-DOS) had 768K of memory available.
-- Bruce
R. Bruce Dubbs, Jr., Ph.D. bdu...@tis.kelly.af.mil
Software Engineering Section Surface Mail: SA-ALC/TISCE
Mission Critical Software Division Kelly AFB, Tx, 78241
Joseph Coughlan
>
>> >> : > How did MS slip that 640k barrier in DOS without IBM's knowledge?
>> >> : > Is Bill Gates that smart?
>>
>> The 640K limit was a hardware limitation in the original IBM PC design
>
>Hardware Ohhhhh no my friend. Software!
>
>Both the mac and amiga used offsets to have contigous ram space. It takes
>more brains to do it this way. M$loth was on a deadline, since the
>original sought after os was cpm wasn't it? but that guy was off sailing
>the seas with his new bimbo wife.
Gosh, why get personal about a guy and blame him for everything?
I recall, perhaps it is documented in Accidental Empires, that IBM
wanted a PC with the capacity for software cartridges, like a nintendo
system. The orginal IBM specification was to use 512k for programs
and the higher 512 for cartridges and etc.
I'm 80% sure that MS convinced IBM to relocated 128k from the 512k
cartridge to program memory when defining the OS specification.
2nd, the IBM PC project was to use current, off the self components to
build a business PC. To increase system memory by an order of
magnitude with off the self technology seems reasonable. Like going
from 4mb to 40mb as a base configuration.
joseph
--
Joseph Coughlan, jo...@gaia.arc.nasa.gov.
John H. Kim
Erik van Linstee <lin...@dutecaj.et.tudelft.nl> wrote:
>merlin (mer...@neuro.usc.edu) wrote:
>: >> : > How did MS slip that 640k barrier in DOS without IBM's knowledge?
>: >> : > Is Bill Gates that smart?
>
>: The 640K limit was a hardware limitation in the original IBM PC design
>: created by IBM designers -- not a terrible secret limitation of DOS --
>
>no need for the software to hardcode it too, then,
It's not really hardcoded into the software. IBM reserved the memory
between 640k and 1MB for hardware-related things like video, BIOS, etc.
This fragments the memory address at 640k. Most programs need to run
in contiguous memory so they can't exceed this 640k barrier unless you
mess with memory expanders and such. Just as bad as hardcoding it into
the software.
>and certainly no reason why that still is in there 12 years later.
That's my major gripe with Microsoft. Ever since the introduction of
the 80286, the hardware was capable of running in protected mode w/o
this limitation. But Microsoft sat on DOS, I suspect because there
was no real competition. Windows and OS/2 are testimony to how you
can run in protected mode and switch back to real mode for backwards
compatibility. Strangely, Microsoft decided to base Windows on DOS.
The Blue Beetle
: mer...@neuro.usc.edu (merlin) writes:
: > >> : > How did MS slip that 640k barrier in DOS without IBM's knowledge?
: > >> : > Is Bill Gates that smart?
: >
: > The 640K limit was a hardware limitation in the original IBM PC design
: Hardware Ohhhhh no my friend. Software!
: Both the mac and amiga used offsets to have contigous ram space. It takes
: more brains to do it this way. M$loth was on a deadline, since the
: original sought after os was cpm wasn't it? but that guy was off sailing
: the seas with his new bimbo wife.
I thought he was off flying his new plane??
Anyway, his company can back with a vengence with DRDOS 5.0 then DRDOS
6.0. Both of which made old Microsucks wake up in a flurry, and now we
have the DOS wars :)
Soon, we shall have Novell DOS 7, and it back the drawing board again for
Microscum.
: ------------------------------
: a...@qedbbs.com (Peter Dilley) or qed!aga
: The QED BBS -- (310)420-9327
*The Blue Beetle*
Harry Myhre
>I just have to jump in here. The so-called 640K barrier is (and was) NOT a DOS
>limitation. It was an IBM-PC limitation. DOS was written with the Intel 8086
>in mind, which only has 1Mb of address space. In 1980, when IBM designed the
>original PC, fully populated mainframes only had 16Mb of memory, so 1Mb for a
>PC (with no real applications at the time) was considered huge. IBM decided to
>use some of the address space for other things such as video, BASIC in ROM, BIOS,
>etc. With hindsight, it may have not been the best decision, but at the time
>it wasn't really that bad either.
Right, I read once that thr 640k memory barrier kludge was not Microsoft's
fault - it was IBM's fault.
--
|_| |\/| Harry Myhre har...@netcom.com
| | | | Los Angeles, California
Corey Thompson
> bdu...@tis.kelly.af.mil writes:
>I just have to jump in here. The so-called 640K barrier is (and was) NOT a DOS
>>limitation. It was an IBM-PC limitation. DOS was written with the Intel 8086
>>in mind, which only has 1Mb of address space. In 1980, when IBM designed the
>>original PC, fully populated mainframes only had 16Mb of memory, so 1Mb for a
>PC (with no real applications at the time) was considered huge. IBM decided to
>>use some of the address space for other things such as video, BASIC in ROM,
>>time it wasn't really that bad either.
>
> Right, I read once that thr 640k memory barrier kludge was not Microsoft's
> fault - it was IBM's fault.
It was nobody's fault, back when DOS was made, 640k was ALOT of memory.
It *IS* Microsoft's (tm) fault that we still have it. Think about it.
Since 4.01, what has really changed? Nothing. Maybe a memory manger, or
a disk cache, or a bunch of 3rd party stuff bundeled along with it.
The actual DOS code hasn't really undergone much of a change at all.
--
Internet: cor...@mankato.msus.edu | Mankato State University
Compu$erve: 71044,1260 |
Michael Klug
using WABI.
Mike
Marc Lewis
One more little note about this......It _is_ Microsquish's fault that we
are still using it since after the 286 was in production for a while, OS/2
v1.x was released which used the protected mode of that chip and eliminated
that barrier, but (speculation here) since it wasn't MS's project completely,
rather IBM/MS joint venture MS wasn't making enough money off of it, and
continued to push DOS, subsequently licencing the code to IBM for $$$.
Just a speculation, there may or may not be any truth as to why, but the
truth is that that is what did happen.....OS/2 went by the wayside until v2.x
and DOS kept selling....Even with the blunderous v4.xx.......Oh, well......
--
Send _lots_ of small, | "Beware my friend as you pass by,
unmarked bills to: | as you are now so once was I,
| As I am now so you must be,
ma...@ctshp.celtech.com | prepare my friend to follow me." - D. Mustaine
follow the warms! fool!
>
>Say what you will, DOS 4.0 is the worst DOS ever made.
>
>: > Is Bill Gates that smart?
>: >
>: know that it was present in DOS 5.0 and, by popular demand,is still present
>: in DOS 6.0 and DOS 6.2.
>
>I am simply stating that Microsoft did not impose this 640k barrier without
>IBM's knowledge.
I have read that the 640k limit is not the fault of the OS, but the fact
that this is were _IBM decided_ to put some or all of the BIOS stuff.
If this is removed then real mode can get close to the full meg barrier.
I have seen systems that have had text only and had something like 920k
free base memory. But the point is that that 640K limit is IBM's addition
to the PC, NOT MS's!
Matt Austern
> I have read that the 640k limit is not the fault of the OS, but the fact
> that this is were _IBM decided_ to put some or all of the BIOS stuff.
> If this is removed then real mode can get close to the full meg barrier.
> I have seen systems that have had text only and had something like 920k
> free base memory. But the point is that that 640K limit is IBM's addition
> to the PC, NOT MS's!
Probably true (only "probably" because none of us know the full
politics between IBM and Microsoft back in 1980), but also not very
interesting. OK, so a different hardware design might have made it
an 830k barrier (or whatever), instead of 640k. So what?
The real problem isn't that the number is 640 instead of 720; it's
that people are still using a memory addressing mode that limits
memory access to 1mb, in 64kb chunks. If all of memory, including ROM
and DMA, is forced to reside below 1mb, then of course memory is
going to be cramped. The details are unimportant.
The solution isn't to worry about how to squeeze another few bytes of
RAM out of that precious first megabyte, but to use a memory accessing
scheme that can address more than 1MB. Which, of course, is what OS/2
does.
[Incidentally, when I wrote MS-DOS programs, I found the 64k barrier
more annoying than the 640k barrier.]
--
Matthew Austern Never express yourself more clearly
ma...@physics.berkeley.edu than you think. ---N. Bohr
Rudynell S Millian
the fact that a dozen odd years later the kludge is STILL in there. What is
it that kept MS from removing the barrier? Compatibility isn't an excuse,
otherwise Windows would NOT have spread like it did.
R S Rodgers
Sounds honest.
> Now who do you blame for
>the fact that a dozen odd years later the kludge is STILL in there. What is
>it that kept MS from removing the barrier?
People who wanted their copies of WordStar and WordPerfect, not
to mention 123, to continue working. That is, the people who were
using the machines who _should_ have input on such matters.
Oh, and you can certainly point the finger at programmers who, in
the day of 386 machines and 32 bit extenders, stil require 580k
free conventional memory.
> Compatibility isn't an excuse,
>otherwise Windows would NOT have spread like it did.
That's false. Windows spread for a lot of other reasons, and
although compatibility was one of the biggies (and explains why
OS/2 1.3, which could run those business apps one at a time, like
DOS, was a flop), Windows succeeded for a lot of other reasons,
not the least of which is that it took the huge weight of having
to write a print driver library and a graphical interface themselves
(something MS has done _right_--look at the crummy nonsense WordPerfect
put out with WP6.0 for DOS and think about how well MS _did_ when
putting Windows together), not to mention the practical impossibility
of supporting cross application cut and paste of any utility, off
of their shoulders, while giving users a cheap visual thrill, great
cut and paste capabilities, easy configuration and less to worry
about (==learn).
--
The big mistake that men make is that when they turn thirteen or fourteen and
all of a sudden they've reached puberty, they believe that they like women.
Actually, you're just horny. It doesn't mean you like women any more at
twenty-one than you did at ten. --Jules Feiffer (cartoonist)
R S Rodgers
follow the warms! fool! <wa...@ac.dal.ca> wrote:
>I have read that the 640k limit is not the fault of the OS, but the fact
>that this is were _IBM decided_ to put some or all of the BIOS stuff.
Yup. And then, as with other things (e.g., Why can't DOS programs
have lonf filenames? Because every program under the sun is hard
coded to use 8.3 filenames and LFN would be of scattershot
utility under DOS at best because of this), the programs that were
written for DOS that no one could do without.
>If this is removed then real mode can get close to the full meg barrier.
>I have seen systems that have had text only and had something like 920k
>free base memory. But the point is that that 640K limit is IBM's addition
>to the PC, NOT MS's!
It was either a Neat or C&T chipset that supported huge DOS
conventional memory--something like 860k.
MSDOS isn't limited to 640k. Try it on other machines that
ran it, back in '83, and many of them--in addition to being all
around superior to IBM's lousy design (with its crummy CGA and
MDA), could have more than 640k. DOS 2.11 on a Z100 could have
over 730k free at the command prompt.
51g...@cua.edu
> Holy shit. Windows is no OS, it is a GUI. You can run Windows applications on Unix,
> using WABI.
>
While I tend to fall into the "Win is not an OS camp," your logic is a
bit flawed.
You can run DOS apps under OS/2 without WABI or anything else. Does
that mean DOS (or the combination of DOS+Windows) isn't an OS? I doubt it,
unless a *lot* of people are using computers without any operating system!
merlin
>that people are still using a memory addressing mode that limits
>memory access to 1mb, in 64kb chunks.
The 64KB code/data segment size limitation is really a far more serious
problem than the initial 640KB or 1MB total system memory limitation --
the 640KB or 1MB limitation was easily overcome by using 64KB segmented
windows into whatever extended/expanded memory is installedon a system.
The real difference between 16 bit addressing (inherent in the 8088, the
80286, and most Microsoft inspired software products) and 32 bit memory
addressing (inherent in most UNIX and all UNIX derived operating systems)
is that 16 bit addressing requires you to break up your code / data into
an enormous number of 64KB pieces -- 32 bit addressing allows you to put
all code and data into one memory segment thereby avoiding the overhead
associated with cutting today's memory hungry programs into a series of
incredibly tiny pieces.
IBM did not make much of a mistake in designing a 640KB limitation in
their original IBM PC because programs of that day -- even on largest
mainframes tended to fall in that 640KB limitation. The most serious
limitation -- the 16 bit memory segmented addressing scheme requiring
code and data to be broken up into 64KB segments was a limitation of
the Intel 8088 and 80286 chips. The 8088 real mode limitation was
even carried over to 80386/80486/Pentium (when operated in real mode)
-- thereby leading to continuation of 16 bit/64KB limitation in the
basic design of the Windows 3.0/3.1 and OS/2 1.x operating systems.
It was Intel and not IBM which ultimately designed the 16bit / 64KB
segmented memory architecture -- an architecture which was entirely
appropriate for personal computers of the 1970's and early 1980's --
an architecture and addressing scheme which has no appropriate place
in personal computer systems of the 1990's. The only reason we tend
to use 16bit / 64KB addressing today is inertia from decisions which
incorporate this underpowered architecture in systems of the 1970's.
The argument at the time of the August 12, 1981 introduction of the
original IBM PC was that no one would need more memory than 640KB in
a 'home computer' (most such systems were lucky to have 64KB), the
cost of memory was prohibitively high, and if any market existed for
a higher power system that market should be filled by some other IBM
commercial computer systems division. Remember that as of 1981 few
of us could envision even the underpowered applications which were
ultimately developed for the 8088 processor -- things like a 64 MHz
processor, 256KB cache memory, 8MB main memory, 1024x768 pixel x 24
bit 72Hz noninterlaced display systems, 1 Gigabyte hard drives, 660
MB CDROM drives, 150/250MB tape drives, 10MBPS ethernet, 14.4 KBPS
dial up modems, and laserprinters bundled into a home system for a
little over $2,000.00 which the ordinary man could afford were all
but inconceivable in 1981. Indeed, I felt quite lucky to be able to
acquire one of the first few IBM PC/XT's with crippled motherboard,
10MB hard disk drive, 320KB floppies, and a monochrome text graphics
display system for a bit more than double what a high end computing
system would cost today.
Don't blame IBM for the hardware -- don't be too hard on Microsoft
for their software (at least DOS within the limitations of a 1981
personal computer system) -- and don't be too hard on Intel - they
made the personal computer possible by taking the risk to develop
a relatively inexpensive 16 bit segmented processor (8086) and the
even less costly 8088 processor which could simulate the 8086 while
cutting motherboard costs by using an 8 bit in place of most costly
16 bit external interface bus. IBM selected the 8088 because it
cut their motherboard and interface adaptor costs by permitting use
of 8 bit interfce chips throughout their design rather than having
to increase the cost of their system by buying costly 16 bit parts.
The 8088 based IBM PC was an appropriate design on August 12, 1981
-- but continuing to maintain compatibility with such an outdated
design (16 bit/64KB segmented memory) in operating systems of the
mid-1990's may be inappropriate. Certainly our stock of software
has turned over several generations since 1981. If someone needs
compatibility with 16 bit programs maybe they should buy some $300
high speed 8088/80286 emulator card; otherwise the operating system
vendors and application developers should get on w/ 32 bit programs
R S Rodgers
>In article <CJ5u4...@ztivax.zfe.siemens.de>, mi...@km21.zfe.siemens.de (Michael Klug) writes:
>> Holy shit. Windows is no OS, it is a GUI. You can run Windows applications on Unix,
>> using WABI.
I wonder, if Windows 3.1 isn;t an OS because it relies on DOS for
one thing (the filesystem), are Windows 3.11 and Widnows for Workgroups
3.11 "real" OSs since they handle the disk access themselves?
> While I tend to fall into the "Win is not an OS camp," your logic is a
>bit flawed.
>
> You can run DOS apps under OS/2 without WABI or anything else. Does
>that mean DOS (or the combination of DOS+Windows) isn't an OS? I doubt it,
>unless a *lot* of people are using computers without any operating system!
Or better yet, you can run Os/2 apps under NT. OS/2 an OS? Nah.
Jenkins, David F.
> In article <1994Jan6...@cua.edu>, <51g...@cua.edu> wrote:
> >In article <CJ5u4...@ztivax.zfe.siemens.de>, mi...@km21.zfe.siemens.de (Michael Klug) writes:
{Whole bunch munched]
> Or better yet, you can run Os/2 apps under NT. OS/2 an OS? Nah.
You can run MVS under (on top of, really) VM - does that mean MVS is not
an OS? Nah.
Paul Monday
|> In article <1994Jan6...@cua.edu>, <51g...@cua.edu> wrote:
|> >In article <CJ5u4...@ztivax.zfe.siemens.de>, mi...@km21.zfe.siemens.de (Michael Klug) writes:
|> >> Holy shit. Windows is no OS, it is a GUI. You can run Windows applications on Unix,
|> >> using WABI.
|>
|>
|> I wonder, if Windows 3.1 isn;t an OS because it relies on DOS for
|> one thing (the filesystem), are Windows 3.11 and Widnows for Workgroups
|> 3.11 "real" OSs since they handle the disk access themselves?
|>
Nice comeback, an OS is more than a filesystem though, it's a memory
manager also...does WW3.11 provide it's own memory management?
Good point though...many apps in DOS access devices directly...but that doesn't
make them Operating Systems 8)
Could I boot WW 3.11 without DOS? No....?
|>
|> > While I tend to fall into the "Win is not an OS camp," your logic is a
|> >bit flawed.
|> >
|> > You can run DOS apps under OS/2 without WABI or anything else. Does
|> >that mean DOS (or the combination of DOS+Windows) isn't an OS? I doubt it,
|> >unless a *lot* of people are using computers without any operating system!
|>
|>
|> Or better yet, you can run Os/2 apps under NT. OS/2 an OS? Nah.
This can be done since OS/2 gives you the Virtual Machines. OS/2 still
owns the DOS app. NTs 'emulation' of OS/2 does not go very far ... text mode.
I THINK the OS/2 DOS machines are 'closer' to a real OS than the OS/2 and DOS
emulation under NT since OS/2 allows DOS to do direct hardware manipulation
in some cases. (Correct me if I'm wrong ... I probably am) There are more
compatibility problems with NT since NT prevents apps from doing OS type
operations. OS/2 handles memory mgmt, disk access, hardware mgmt a little
looser when in a DOS box.
(I can't believe I just added to this ongoing debate ... I think I just broke
a New Year's resolution)
--
pb monday
ibm rochester
oo infrastructure
R S Rodgers
Paul Monday <pmo...@rchland.vnet.ibm.com> wrote:
>In article <2gha3d$n...@cville-srv.wam.umd.edu>, rsro...@wam.umd.edu (R S Rodgers) writes:
>|> In article <1994Jan6...@cua.edu>, <51g...@cua.edu> wrote:
>|> >In article <CJ5u4...@ztivax.zfe.siemens.de>, mi...@km21.zfe.siemens.de (Michael Klug) writes:
>|> >> Holy shit. Windows is no OS, it is a GUI. You can run Windows applications on Unix,
>|> >> using WABI.
>|>
>|>
>|> I wonder, if Windows 3.1 isn;t an OS because it relies on DOS for
>|> one thing (the filesystem), are Windows 3.11 and Widnows for Workgroups
>|> 3.11 "real" OSs since they handle the disk access themselves?
>|>
>
>Nice comeback, an OS is more than a filesystem though, it's a memory
>manager also...does WW3.11 provide it's own memory management?
Here we go again. Windows 3.*1* provides its own memory
management. What's your point?
>Good point though...many apps in DOS access devices directly...but that doesn't
>make them Operating Systems 8)
Why not? Suppose you write a 386 extended DOS program that handles
everything itself, IO and so on, and doesn't use DOS for anything
(or, optionally, uses DOS as a filesystem)--is that not an OS?
There was a copy of SCO UNIX a long time ago that actually launched
from the DOS command line. Not an OS? You can't boot into Oberon,
either, on an Intel PC without first booting DOS.
>Could I boot WW 3.11 without DOS? No....?
Could you boot Os/2 without a BIOS chip in your PC?
>|> Or better yet, you can run Os/2 apps under NT. OS/2 an OS? Nah.
[This was sarcasm, incidentally.]
>This can be done since OS/2 gives you the Virtual Machines. OS/2 still
>owns the DOS app. NTs 'emulation' of OS/2 does not go very far ... text mode.
Yes. OK, then are you saying that a copy of OS/2 2.0 that
boots up with CMD.EXE as the shell and doesn't load the WPS is
no longer an operating system?
>I THINK the OS/2 DOS machines are 'closer' to a real OS than the OS/2 and DOS
>emulation under NT since OS/2 allows DOS to do direct hardware manipulation
>in some cases. (Correct me if I'm wrong ... I probably am)
Um, VCPI programs don't run under OS/2 for the same reason...
> There are more
>compatibility problems with NT since NT prevents apps from doing OS type
>operations. OS/2 handles memory mgmt, disk access, hardware mgmt a little
>looser when in a DOS box.
Huh?
>(I can't believe I just added to this ongoing debate ... I think I just broke
> a New Year's resolution)
Same here.
David Tholen
> 5 million OS/2 users? Where? From what planet? So many of you question the
> sales of Windows, I question the user base of OS/2.
The generally accepted number of sales stood at 4 million around the end of
last November. Who is claiming 5 million users?
> It's ironic that IBM has spent billions developing OS/2
We've seen this number before, but never substantiated. Does it even make
sense? Let's say IBM has been working on OS/2 for ten years (a exaggeration,
but a nice round number). Furthermore, let's say IBM pays its programmers
$100,000 per year (another nice round number). Then each programmer was
paid $1,000,000 over the ten years. To have spent "billions" (let's say
2 billion), then IBM must have had at least 2,000 programmers working on
OS/2 full time. Cut the development time to a more reasonable six years
(meaning that it includes version 1.X) and you're up to 3,333 programmers
to have spent 2 billion. The majority of the programmers probably make
more like $50,000 per year, so now we're up to 6,667 programmers working
full time on OS/2. And if "billions" meant 3 billion, then we're up to
10,000 programmers. Yes, this little exercise ignores things like buying
hardware for each programmer, secretarial support, office space, and so on,
but the overhead isn't a factor of ten. Just how many man-years do you
think were spent developing OS/2?
By the way, the Wall Street Journal reported that Microsoft spent $150
million to develop Windows NT. Is IBM really more than ten times less
efficient than Microsoft? Seriously?
> A substantial portion of that went to Microsoft. Bad business.
How much is "substantial"?
> Then they waste their time on OS/2 for Windows.
You flame them for sending money to Microsoft, then you flame them for
releasing a product that eliminates the need to pay royalties to Microsoft!
Make up your mind.
> Why didn't they spend their time on OS/2 applications. I doubt they can,
> look at their wonderful OS/2 applets.
Have you seen their C++ compiler or TCP/IP products, or some of the so-called
"vertical market" stuff like DB/2? Have you seen the speech recognition
demonstrations? Seems to me like they have spent time on OS/2 applications.
> OS/2, from the makers of DOS 4.0.
Windows NT, from the makers of Windows 3.0 and DOS 6.0.
Yuk Lun Chan
> Here we go again. Windows 3.*1* provides its own memory
> management. What's your point?
> Why not? Suppose you write a 386 extended DOS program that handles
> everything itself, IO and so on, and doesn't use DOS for anything
> (or, optionally, uses DOS as a filesystem)--is that not an OS?
Well, in my IMO, it is not an OS.
> There was a copy of SCO UNIX a long time ago that actually launched
> from the DOS command line. Not an OS? You can't boot into Oberon,
> either, on an Intel PC without first booting DOS.
I have no reference to that. But does it replace DOS completely?
>>Could I boot WW 3.11 without DOS? No....?
> Could you boot Os/2 without a BIOS chip in your PC?
Tell me what OS can be boot without BIOS or BIOS like things.
> Yes. OK, then are you saying that a copy of OS/2 2.0 that
> boots up with CMD.EXE as the shell and doesn't load the WPS is
> no longer an operating system?
OS/2 includes a shell called WPS but WPS is not an OS.
--
The above statements do not express the opinion of the Hewlett-Packard Company
-----------------------------------------------------------------------
Yuk Lun Chan, Software Development Engineer
E-mail : y...@hpopd.pwd.HP.COM Desk : Yuk Lun Chan/HP1600
Openmail : Yuk-Lun Chan/pinewood,lab,hpopd
X400: C=GB, ADMD=GOLD 400, PRMD=HP, ORG=HP, OU1=pinewood, SN=Chan, GN=Yuk-Lun
Glenn Amerine
: bdu...@tis.kelly.af.mil writes:
It wasn't a kludge at all. To say it was is like saying everybody hates
Ford Pintos because they had four banger engines in them. Its NOBODY's
fault the original DOS had a 640K limit because the chip could only see
1MB total.
Blame also can't be handed out because BOTH Microsoft and IBM tried to push
OS/2 once there were enough 286+ machines out there, but thats a different
thread :-}
--
Glenn Amerine Inet: gl...@pie.mhsc.org
Computer Systems Analyst Ma Bell: (614)224-1336
Metropolitan Human Services Commission Fax: (614)224-6472
Todd Walk
The 640K barrier is a problem of software, design architechture, and
the intel chips. The 8086 boots from the top 4 bytes of memory, so
IBM decided to put all the adapter/bios stuff up there (they had to
put something because the top 4 bytes couldn't be used anyway).
This fragmented memory below and above 1M. Then since DOS can't
handle fragmented memory itself, you only have 640K accessable
by DOS. The mac had this same problem at the 8M area, but
got around it by using the memory remapping features of the 68020+.
--
Todd Walk
wa...@mrcnext.cso.uiuc.edu
Wolfgang Strobl
You have found a way to run real mode applications in protected mode?
Plain old DOS apps written to be run on a 8088, that is; I'm not talking
about Windows 2.x or family mode apps).
Please tell us how you did it. I'm pretty sure that I'm not the only
one who would like to have his old DOS apps be able to use plenty
of RAM. I've got 32M, btw.
--
o ( Wolfgang Strobl Wolfgan...@gmd.de (+49 2241) 14-2394
/\ * GMD mbH
_`\ `_<=== Schloss Birlinghoven, #include <std.disclaimer.hpp>
__(_)/_(_)___.-._ 53754 Sankt Augustin, Germany _____________________________
Wolfgang Strobl
Why should it? MSDOS doesn't have a 640 K memory barrier, and never had.
Run it on a (non IBM compatible) machine where both the BIOS and the
display buffers are located outside the 1M real mode address space,
and you'll get almost the full 1024K for your applications use.
Kai Uwe Rommel
>> It was nobody's fault, back when DOS was made, 640k was ALOT of memory.
>>It *IS* Microsoft's (tm) fault that we still have it. Think about it.
>>Since 4.01, what has really changed? Nothing. Maybe a memory manger, or
>>a disk cache, or a bunch of 3rd party stuff bundeled along with it.
>>The actual DOS code hasn't really undergone much of a change at all.
>
>Why should it? MSDOS doesn't have a 640 K memory barrier, and never had.
>
>Run it on a (non IBM compatible) machine where both the BIOS and the
>display buffers are located outside the 1M real mode address space,
>and you'll get almost the full 1024K for your applications use.
Wolfgang, we are not talking abbout peanuts, about 1M vs. 640k. We are
talking about orders of magnitude. That should have been obvious.
And the DOS API *is* tied to 16-bitness (pointers passed to DOS in API
calls and many other things) and real mode (it is common to base
address calculations in DOS progams on the values of segment
registers).
Kai Uwe Rommel
--
/* Kai Uwe Rommel Muenchen, Germany *
* rom...@ars.muc.de CompuServe 100265,2651 *
* rom...@informatik.tu-muenchen.de Fax +49 89 324 4524 */
DOS ... is still a real mode only non-reentrant interrupt
handler, and always will be. -Russell Williams
David Charlap
>
>Why should it? MSDOS doesn't have a 640 K memory barrier, and never had.
>
>Run it on a (non IBM compatible) machine where both the BIOS and the
>display buffers are located outside the 1M real mode address space,
>and you'll get almost the full 1024K for your applications use.
Yes, but almost all applications assume that you are on a PC compatible
computer. If you have a system where the BIOS and display buffers are
anywhere other than where they are on IBM-PC machines, you can be sure
that it won't run most PC software.
Every good DOS programmer knows that relying on DOS/BIOS calls for all
I/O makes for a very slow program. And only those programs will run on
your computer.
--
David Charlap | Clouds on your shoulder
dic...@hertz.njit.edu | Aren't they grazed by the afterglow
-----------------+-----+ How quiet they gather
TEAM-OS/2 member | When the storm is about to blow - Duran Duran
Wolfgang Strobl
>str...@gmd.de (Wolfgang Strobl) writes in article <strobl.7...@gmd.de>:
>>> It was nobody's fault, back when DOS was made, 640k was ALOT of memory.
>>>It *IS* Microsoft's (tm) fault that we still have it. Think about it.
>>>Since 4.01, what has really changed? Nothing. Maybe a memory manger, or
>>>a disk cache, or a bunch of 3rd party stuff bundeled along with it.
>>>The actual DOS code hasn't really undergone much of a change at all.
>>
>>Why should it? MSDOS doesn't have a 640 K memory barrier, and never had.
>>
>>Run it on a (non IBM compatible) machine where both the BIOS and the
>>display buffers are located outside the 1M real mode address space,
>>and you'll get almost the full 1024K for your applications use.
>Wolfgang, we are not talking abbout peanuts, about 1M vs. 640k. We are
>talking about orders of magnitude. That should have been obvious.
Well, I understood the paragraph starting with "Since 4.01 .." I
responded to as "why didn't Microsoft change MSDOS somewhere between
version 1 and version 6 so that it can give an app more linear address
space than just that tiny 640K fraction of my computers main memory?"
>And the DOS API *is* tied to 16-bitness (pointers passed to DOS in API
>calls and many other things) and real mode (it is common to base
>address calculations in DOS progams on the values of segment
>registers).
And exactly because of that the whole complaint "why didn't MS change
MSDOS to 32 bit - the've had enough time, after all" simply can be answered
by "because it isn't possible without breaking 99,8% of all applications ".
--
o ( Wolfgan...@gmd.de (+49 2241) 14-2394
/\ * GMD mbH #include
_`\ `_<=== Schloss Birlinghoven, <std.disclaimer>
Mark Gazda
I guess you are just another person with no knowlege of the facts... let
me fix that..
Windows 3.x was the child of OS/2 1.3... Yes that is correct.. not only
is the interface the same, but so are sections of the code...
Second... OS/2 2.1 & 2.1 for Windows were designed to be an operating
system, not an aplication for DOS... Windows can never or will never
become equal in potential to OS/2 without a 32bit operating system.. It
sold well... because it was a good product... for it's time. That was
because it had a long .. and very pothetic history ... (look at
MS-Windows 2.0 or 5.0 DOS-Shell should I say) ... It only works well
because it is built on top of DOS... I don't think MS will be able to
pull off a well written operating system that wasn't stolen from someone
else (ie: DOS)...
Now that I mad my view point clear... have fun with NT... and think of
upgrading to OS/2!
Mark Gazda
Attainable Solutions
mark....@onlinesys.com
Joseph Coughlan
>Hi there!
>
>I guess you are just another person with no knowlege of the facts... let
>me fix that..
>
>Windows 3.x was the child of OS/2 1.3... Yes that is correct.. not only
>is the interface the same, but so are sections of the code...
Nope, OS/2 1.3 was an IBM effort, MS stopped with OS/2 1.2.
>
>Second... OS/2 2.1 & 2.1 for Windows were designed to be an operating
>system, not an aplication for DOS... Windows can never or will never
>become equal in potential to OS/2 without a 32bit operating system.. It
>sold well... because it was a good product... for it's time. That was
>because it had a long .. and very pothetic history ... (look at
>MS-Windows 2.0 or 5.0 DOS-Shell should I say) ... It only works well
>because it is built on top of DOS... I don't think MS will be able to
>pull off a well written operating system that wasn't stolen from someone
>else (ie: DOS)...
>
>Now that I mad my view point clear... have fun with NT... and think of
>upgrading to OS/2!
>
>Mark Gazda
>Attainable Solutions
>mark....@onlinesys.com
--
Joseph Coughlan, jo...@gaia.arc.nasa.gov.