An 8086 Bios for S-100 systems to run MS-DOS/FreeDOS

[monahanz]

I have been writing 8086 code for a fairly comprehensive 8086 monitor
(~9000 lines of code) that with a number of new S-100 boards (an 8086
board, our IDE Board, our ZFDC board , our PIC-RTC Board and the
Propeller driven Console IO board) allows one to boot MS-DOS. The
goal is to have hardware and a BIOS so that standard generic MS-DOS
(or FreeDOS) does not know it's not talking to a PC motherboard.
This turned out to be more difficult than I thought. Most major PC
software interrupts (0-20H) have to be emulated in very precise
ways. Anyway delighted to say MS-DOS (V4.01) now boots up on 3.5"
floppies or on a CF-Card with our IDE drive S-100 board.

Thanks to this group BTW for suggesting NASM. Fantastic assembler

If you have time take a look at
http://s100computers.com/Software%20Folder/8086%20Monitor/8086%20Monitor.htm
there are a few short videos there that show the system in action.

Except for the 8086 board the other boards are already done. The 8086
prototype has been working here for months (8MHz) and Andrew expects
the "commercial quality" board with trace optimization to go out
fairly soon.

Next up is to checkout a 80286 prototype board I have here. This
will be followed by a XVGA S-100 board using a Cirrus Logic chip
(GD-542x chip). Then an 80386 or 80486!

[Rod Pemberton]

"monahanz" <mon...@vitasoft.org> wrote in message
news:bde97d80-bfe7-4bed...@v18g2000yqj.googlegroups.com...
>
> [...] Most major PC

> software interrupts (0-20H) have to be emulated in very precise
> ways. Anyway delighted to say MS-DOS (V4.01) now boots up on 3.5"
> floppies or on a CF-Card with our IDE drive S-100 board.

> [...]

If you put up on your website any info on what BIOS interrupts are needed
for specific DOS versions, or DOS or BIOS "secrets", or IDE programming
could drop a post here? Thanks!


Rod Pemberton

[monahanz]

On Sep 19, 1:19 am, "Rod Pemberton" <do_not_h...@noavailemail.cmm>
wrote:
> "monahanz" <mona...@vitasoft.org> wrote in message

I will have to backtrack Rob. Unfortunately i did not keep a record as
I stepped along from V1.0, 2, 3.3 and 4.01, sorry. The current BIOS
code contains the minimum for V4.01. Some of the int 15H interrupts
just return the carry flag set. To get to higher versions of DOS
however it looks like I will have to have much more. I think an 80286,
CMOS clock chip and in general an IBM-AT appearance.. Curious, has
anybody run MSDOS > 4.01 on an IBM-PC XT box.

[Rugxulo]

Hi,

On Sep 20, 12:41 am, monahanz <mona...@vitasoft.org> wrote:
>
> Curious, has anybody run MSDOS [greater than] 4.01 on an IBM-PC XT box.

You'll have to ask one of these guys. IIRC, they ran PC-DOS and DR-
DOS, not vanilla MS-DOS. But I'd blindly guess that 6.22 should run on
any machine.

http://www.oldskool.org/
http://www.rubbermallet.org/
http://www.brutman.com/

[Jim Leonard]

On Sep 19, 6:38 pm, Rugxulo <rugx...@gmail.com> wrote:
> > Curious, has anybody run MSDOS [greater than] 4.01 on an IBM-PC XT box.
>
> You'll have to ask one of these guys. IIRC, they ran PC-DOS and DR-
> DOS, not vanilla MS-DOS. But I'd blindly guess that 6.22 should run on
> any machine.
>
> http://www.oldskool.org/http://www.rubbermallet.org/http://www.brutman.com/

I'm one of those guys. Yes, every single DOS (except default builds
of FreeDOS) runs on old PCs and XTs. Some are better to run than
others; I personally run IBM PC DOS 2000 (IBM's PC DOS 7.0 with Y2K
fixes) on any 808x machine with a hard drive and 640K of RAM. If less
RAM or no hard drive, DOS 3.x is my go-to DOS. For any 286 or higher,
I usually install whatever I feel like installing at the time, like
Caldera 7.03.

The only exception to the above are clones; some clones can run plain
MS-DOS but perform better or have less quirks if you use the branded
DOS that came with the machine. For example, AT&T MS-DOS 3.3's date
and time commands can correctly update the AT&T PC 6300's built-in
clock chip, whereas plain MS-DOS's date/time commands do not. IIRC,
Toshiba's DOS 3.2's MODE command can adjust the Toshiba 1200's LCD
screen in ways that plain MS-DOS's MODE.COM cannot; same goes for
Sperry machines. Even if there is no obvious differences, it's
usually best to run a branded DOS on a branded machine just to avoid
the unknown.

[monahanz]

On Sep 20, 7:52 am, Jim Leonard <mobyga...@gmail.com> wrote:
> On Sep 19, 6:38 pm, Rugxulo <rugx...@gmail.com> wrote:
>
> > > Curious, has anybody run MSDOS [greater than] 4.01 on an IBM-PC XT box.
>
> > You'll have to ask one of these guys. IIRC, they ran PC-DOS and DR-
> > DOS, not vanilla MS-DOS. But I'd blindly guess that 6.22 should run on
> > any machine.
>

> >http://www.oldskool.org/http://www.rubbermallet.org/http://www.brutma...

>
> I'm one of those guys.  Yes, every single DOS (except default builds
> of FreeDOS) runs on old PCs and XTs.  Some are better to run than
> others; I personally run IBM PC DOS 2000 (IBM's PC DOS 7.0 with Y2K
> fixes) on any 808x machine with a hard drive and 640K of RAM.  If less
> RAM or no hard drive, DOS 3.x is my go-to DOS.  For any 286 or higher,
> I usually install whatever I feel like installing at the time, like
> Caldera 7.03.
>
> The only exception to the above are clones; some clones can run plain
> MS-DOS but perform better or have less quirks if you use the branded
> DOS that came with the machine.  For example, AT&T MS-DOS 3.3's date
> and time commands can correctly update the AT&T PC 6300's built-in
> clock chip, whereas plain MS-DOS's date/time commands do not.  IIRC,
> Toshiba's DOS 3.2's MODE command can adjust the Toshiba 1200's LCD
> screen in ways that plain MS-DOS's MODE.COM cannot; same goes for
> Sperry machines.  Even if there is no obvious differences, it's
> usually best to run a branded DOS on a branded machine just to avoid
> the unknown.

Very interesting and informative Jim. Thanks. Looks like the best way
for me to go then is try and track down PC DOS 7.0. I will be doing a
286 and 386 board in the future but right now I working on an XVGA
board and an CMOS RTC board.

[Jim Leonard]

On Sep 20, 10:57 am, monahanz <mona...@vitasoft.org> wrote:
> Very interesting and informative Jim. Thanks.  Looks like the best way
> for me to go then is try and track down PC DOS 7.0.  I will be doing a
> 286 and 386 board in the future  but right now I working on an XVGA
> board and an CMOS RTC board.

cough ftp.oldskool.org/pub/misc/DOS/PCDOS2K cough

[Rugxulo]

Hi,

On Sep 20, 10:57 am, monahanz <mona...@vitasoft.org> wrote:

> On Sep 20, 7:52 am, Jim Leonard <mobyga...@gmail.com> wrote:
>
> > On Sep 19, 6:38 pm, Rugxulo <rugx...@gmail.com> wrote:
>
> > > > Curious, has anybody run MSDOS [greater than] 4.01 on an IBM-PC XT box.
>
> > > You'll have to ask one of these guys. IIRC, they ran PC-DOS and DR-
> > > DOS, not vanilla MS-DOS. But I'd blindly guess that 6.22 should run on
> > > any machine.
>
> > >http://www.oldskool.org/http://www.rubbermallet.org/http://www.brutma...
>
> > I'm one of those guys.  Yes, every single DOS (except default builds
> > of FreeDOS) runs on old PCs and XTs.

I think FreeDOS should run on 8086. There were minor errors in the
past, but they've long ago been fixed, AFAICT. If you really can't run
it, it would be best to report it to the mailing lists. IIRC, it was a
bootsector issue years ago, but that's been fixed. I know 8086
versions of the kernel and shell exist. Granted, like you say, DOS 3.3
may have a smaller RAM footprint for low-RAM machines (less than 640k)
if you're desperate for more free memory for certain apps.

And BTW, not every DOS runs on 8086, of course. Some require 386. And
PTS-DOS (and its PTS-DOS32 counterpart) requires something like 186
and 286, maybe higher, can't remember. But at least the popular
variants (MS, IBM, FD) should run on 8086.

> Very interesting and informative Jim. Thanks.  Looks like the best way
> for me to go then is try and track down PC DOS 7.0.  I will be doing a
> 286 and 386 board in the future  but right now I working on an XVGA
> board and an CMOS RTC board.

Er, I'm not saying PC-DOS is bad, it's not. But it might be hard to
find or expensive. At least one quick look on eBay a few months ago
didn't get my hopes up. I don't see any huge advantage with it. If you
want to use it personally / locally, that's fine. But obviously
FreeDOS is better for hacking / tweaking or even redistributing.
Sources are much easier to read than raw disassembly.

But a minimal PC-DOS setup (w/ FAT32 support) is supposedly in the IBM
Server Scripting Toolkit. At least I *think* it's legal, but of course
I'm no lawyer.

http://www-03.ibm.com/systems/management/sgstk/

[Jim Leonard]

On Sep 20, 5:21 pm, Rugxulo <rugx...@gmail.com> wrote:
> I think FreeDOS should run on 8086. There were minor errors in the
> past, but they've long ago been fixed, AFAICT. If you really can't run
> it, it would be best to report it to the mailing lists. IIRC, it was a
> bootsector issue years ago, but that's been fixed. I know 8086
> versions of the kernel and shell exist. Granted, like you say, DOS 3.3
> may have a smaller RAM footprint for low-RAM machines (less than 640k)
> if you're desperate for more free memory for certain apps.

The main issue, even with the 8086 kernel, was that doing a DIR took
nearly two minutes for the first access (don't remember how long
subsequent accesses were). Vanilla MS-DOS 4.x or before doesn't delay
at all, and 5.x or later delays only about 15-17 seconds on first
access. The next time I need to install a new DOS, I will revisit the
issue and see if its improved.

I'm hesitant to make a request via the mailing lists because I highly
doubt the FreeDOS developers are concerned with performance on a <5MHz
computer.

> And BTW, not every DOS runs on 8086, of course. Some require 386.

Which ones?

> PTS-DOS (and its PTS-DOS32 counterpart) requires something like 186

Yes, but why would you use PTS-DOS? There's no advantage to using PTS-
DOS that I can see (in fact, there are just bugs to deal with).

> FreeDOS is better for hacking / tweaking or even redistributing.
> Sources are much easier to read than raw disassembly.

Agreed, and the OP should give it a try.

> But a minimal PC-DOS setup (w/ FAT32 support) is supposedly in the IBM
> Server Scripting Toolkit. At least I *think* it's legal, but of course
> I'm no lawyer.
>
> http://www-03.ibm.com/systems/management/sgstk/

It is indeed with FAT32 support, and it is indeed a *minimal*
distribution (just enough common things to boot a server and perform
some automation, hence its inclusion in the Server Scripting
Toolkit). It lacks some of the common niceties you would get with a
full distribution. http://www-03.ibm.com/systems/management/sgstk/
has more info and download links.

[monahanz]

Thanks guys for suggestions. Thanks Jim for the URL. I am stuck with a
minor hardware issue (unrelated) at the moment. So it will be a few
days before I get back to software. Will update then.

[Rugxulo]

Hi, :-)

On Sep 21, 9:33 am, Jim Leonard <mobyga...@gmail.com> wrote:
> On Sep 20, 5:21 pm, Rugxulo <rugx...@gmail.com> wrote:
>
> > I think FreeDOS should run on 8086. There were minor errors in the
> > past, but they've long ago been fixed, AFAICT. If you really can't run
> > it, it would be best to report it to the mailing lists. IIRC, it was a
> > bootsector issue years ago, but that's been fixed. I know 8086
> > versions of the kernel and shell exist. Granted, like you say, DOS 3.3
> > may have a smaller RAM footprint for low-RAM machines (less than 640k)
> > if you're desperate for more free memory for certain apps.
>
> The main issue, even with the 8086 kernel, was that doing a DIR took
> nearly two minutes for the first access (don't remember how long
> subsequent accesses were).  Vanilla MS-DOS 4.x or before doesn't delay
> at all, and 5.x or later delays only about 15-17 seconds on first
> access.  The next time I need to install a new DOS, I will revisit the
> issue and see if its improved.

It's probably FreeCOM's fault, though I wouldn't rule out other
causes. Compared to other DOS shells, it's almost entirely written in
C, and by default the binaries were compiled with Turbo C (or C++, I
forget which). You could try recompiling with OpenWatcom (since Bart
semi-recently ported it to that), but I personally am skeptical if
that will help. Your best bet is probably to use a different shell
(4DOS, Gcom, Dog) and/or different "dir" clone.

> I'm hesitant to make a request via the mailing lists because I highly
> doubt the FreeDOS developers are concerned with performance on a <5MHz
> computer.

Well, if anybody can convince them, it's probably you! ;-) But
yeah, they're ridiculously low on manpower, and I don't know, Bart
always seems ultra busy. But hey, it can't hurt to mention it (famous
last words). The only real problem I find is that FreeCOM is
ridiculously overcomplicated, esp. the build process. I mean, the
kernel itself isn't as confusing! :-P

> > And BTW, not every DOS runs on 8086, of course. Some require 386.
>
> Which ones?

I forget, I probably just meant some of the more weird clones like
Real/32 or TSX/32 (neither of which I've tried, I already have enough
to fiddle with, heh).

> > PTS-DOS (and its PTS-DOS32 counterpart) requires something like 186
>
> Yes, but why would you use PTS-DOS?  There's no advantage to using PTS-
> DOS that I can see (in fact, there are just bugs to deal with).

Again, not tried, just mentioning it for completeness. Who knows what
people prefer. I just find it very silly when people think DOS is only
MS-DOS.

> > FreeDOS is better for hacking / tweaking or even redistributing.
> > Sources are much easier to read than raw disassembly.
>
> Agreed, and the OP should give it a try.

I'm not really trying to promote it here, but honestly, there aren't a
lot of other "free" DOSes. Perhaps RxDOS, but that's fairly buggy.

> > But a minimal PC-DOS setup (w/ FAT32 support) is supposedly in the IBM
> > Server Scripting Toolkit. At least I *think* it's legal, but of course
> > I'm no lawyer.
>
> >http://www-03.ibm.com/systems/management/sgstk/
>
> It is indeed with FAT32 support, and it is indeed a *minimal*
> distribution (just enough common things to boot a server and perform
> some automation, hence its inclusion in the Server Scripting
> Toolkit).  It lacks some of the common niceties you would get with a
> full distribution.  http://www-03.ibm.com/systems/management/sgstk/
> has more info and download links.

Well, you could probably use the FreeDOS userland for the utils. I
imagine PC-DOS' kernel is better for things like Win 3.x or Desqview
or other exotic stuff. "Whatever works", right?

[Rod Pemberton]

"Rugxulo" <rug...@gmail.com> wrote in message
news:7924c011-7ec9-46f2...@i21g2000yqk.googlegroups.com...
>
> > > PTS-DOS

>
> I'm not really trying to promote it here, but honestly, there aren't a
> lot of other "free" DOSes. Perhaps RxDOS, but that's fairly buggy.

OpenDOS/DR-DOS ?


Rod Pemberton

[Rugxulo]

Hi,

On Sep 23, 9:36 am, "Rod Pemberton" <do_not_h...@noavailemail.cmm>
wrote:
> "Rugxulo" <rugx...@gmail.com> wrote in message

Honestly, I've not messed with OpenDOS or EDR-DOS a lot. They're kinda
buggy, esp. the former (lacking even a lot of Novell bugfixes). Not
too bad, but anyways. Besides, I think? only the kernel and shell were
(temporarily) free'd, and only for non-commercial use. As you know,
some people would complain. But whatever works is fine, so I have no
problem recommending it or even various others.

[Rod Pemberton]

"Rugxulo" <rug...@gmail.com> wrote in message

news:04371d2b-8aba-4eaa...@h34g2000yqd.googlegroups.com...

> On Sep 23, 9:36 am, "Rod Pemberton" <do_not_h...@noavailemail.cmm>
> wrote:
>
> > OpenDOS/DR-DOS ?
>
> Honestly, I've not messed with OpenDOS or EDR-DOS a lot.

I've not used it. I looked at it a few years ago. Some issues kind of made
me back off, such as the license, binary patch updates, and difficulty
determing what to download, or if I needed to apply numerous patches ...

But, the code is open source, and it is 100% compatible with MS-DOS up to a
certain point because of a lawsuit with Caldera. Apparently, MS was
required to provide whatever information was needed for DR-DOS to be 100%
compatible for a ten year period from 1996 (e.g., to 2006):

http://www.maxframe.com/DR/Info/fullstory/ca_sues_ms.html

Gates declared MS-DOS dead after releasing XP in 2001. If Caldera kept up,
that should cover everything upto the final version of MS-DOS, v8.0 that
came with ME. So, OpenDOS/DR-DOS should be a good for reference info on DOS
internals. Since MS-DOS is dead, other DOS' can improve as they see fit
without needing to be MS, or IBM, or Caldera etc compatible.

IIRC, DR-DOS also had a good .pdf on their DOS with a section on
writing DOS device drivers. I'm not sure anymore if it that was on the
OpenDOS or commercial DR-DOS website. I don't see it on either. Let me
search ... Nope, memory fades, the .pdf is for PC-DOS 7:

http://www.redbooks.ibm.com/redbooks/pdfs/gg244459.pdf


Rod Pemberton

[Rod Pemberton]

"Rod Pemberton" <do_no...@noavailemail.cmm> wrote in message
news:j5ls5s$mpa$1...@speranza.aioe.org...

DR-DOS manuals as html
http://www.drdos.net/documentation/

Rod Pemberton

[Rod Pemberton]

"Rod Pemberton" <do_no...@noavailemail.cmm> wrote in message

news:j5i5gb$eoj$1...@speranza.aioe.org...

RDOS
http://www.rdos.net/rdos/index.htm

PDOS
http://sourceforge.net/projects/pdos/


Rod Pemberton

[Rugxulo]

Hi,

On Sep 24, 7:21 pm, "Rod Pemberton" <do_not_h...@noavailemail.cmm>
wrote:
> "Rugxulo" <rugx...@gmail.com> wrote in message

>
> news:04371d2b-8aba-4eaa...@h34g2000yqd.googlegroups.com...
>
> > On Sep 23, 9:36 am, "Rod Pemberton" <do_not_h...@noavailemail.cmm>
> > wrote:
>
> > > OpenDOS/DR-DOS ?
>
> > Honestly, I've not messed with OpenDOS or EDR-DOS a lot.
>
> I've not used it.  I looked at it a few years ago.  Some issues kind of made
> me back off, such as the license, binary patch updates, and difficulty
> determing what to download, or if I needed to apply numerous patches ...

Admittedly, the license isn't ideal, but most of us don't need
commercial use anyways (right?). It's more of a pain worrying about
how to patch / build / download / install than anything else.

> But, the code is open source,

Not "open source" per se, just opened (temporarily, pretty much only
in 1997, nothing else), but even that won't save you from license
arguments. :-(

> and it is 100% compatible with MS-DOS up to a certain point because of a lawsuit with Caldera.

No, not really. OpenDOS 7.01 wasn't even the final Novell version
since the version control system was (allegedly) archaic and hard to
use. So it lacked a lot of "official" bugfixes. Even later versions
(e.g. 7.03 from early 1999) only claimed IBM PC-DOS 6.0 compatibility
(via version check). My DR-DOS 7.03 Lite copy has no built-in LFN nor
FAT32 nor LBA support. It can run Win 3.1 but nothing beyond that (no
Win9x). So no, it's not really 100% compatible.

(BTW, yeah, FreeDOS is a bit lacking in Win 3.1 support, but
apparently nobody cares, esp. me! Heh. But I guess it wouldn't hurt to
fix that ... in theory.)

> Apparently, MS was required to provide whatever information was needed
> for DR-DOS to be 100% compatible for a ten year period from 1996 (e.g., to 2006):
>
> http://www.maxframe.com/DR/Info/fullstory/ca_sues_ms.html

I would be surprised if that were true. Anyways, it's moot because
"Caldera" [sic] barely ever did jack squat with Novell DOS. Basically,
7.00 was all Novell, 7.01 was the "open" release, 7.02 and 7.03 were
bugfixed a reasonable bit, but nothing further. It all stopped after
1999 (AFAICT), so there haven't been any real releases or improvements
since then. Latest sold to the public is still 7.03, I think.

In other words, DOS was long dead to them, they didn't develop it
further. Funny (?) that they sued MS at all, esp. since they weren't
exactly interested in DOS anymore. But I think some (most?) of them
used it to fund their Linux business, believe it or not.

(I don't really know the details, and maybe I'm misunderstanding here,
so just FYI, it's a confusing mess.)

BTW, they did fake some internal structures to make Win9x run, but
naturally they never released that patch. I think it's fairly common
knowledge that Win9x ran atop DOS similar to Win 3.1 in most ways,
just now was all a single bundled package instead of separate (and
used some undocumented calls).

> Gates declared MS-DOS dead after releasing XP in 2001.

No surprise. MS-DOS and Win9x were always considered unstable as hell.
Also, even as far back as 1985, text mode was considered obsolete by
MS, which is why they started developing Windows so heavily. I just
wish NTVDM didn't suck as bad. I don't know why they never bothered
fixing that.

> If Caldera kept up, that should cover everything upto the final version of MS-DOS,
> v8.0 that came with ME.

They didn't keep it up. 1999 was the last "real" release, AFAICT.
WinME has LBA, LFNs, FAT32, etc, but Caldera had none of that
(officially). There were some hacks they made for FAT32 and VFAT, but
I blindly guess (??) they abandoned that for patent reasons. Dunno,
it's weird.

> So, OpenDOS/DR-DOS should be a good for reference info on DOS internals.

Probably not worth it.

> Since MS-DOS is dead, other DOS' can improve as they see fit
> without needing to be MS, or IBM, or Caldera etc compatible.

Yes, but it hasn't really happened. Most effort is on compatibility,
which is probably a wise idea. (Besides, not a lot of DOS developers
exist, for whatever reason.)

[Jim Leonard]

On Sep 26, 10:58 pm, Rugxulo <rugx...@gmail.com> wrote:
> > and it is 100% compatible with MS-DOS up to a certain point because of a lawsuit with Caldera.
>
> No, not really. OpenDOS 7.01 wasn't even the final Novell version
> since the version control system was (allegedly) archaic and hard to
> use. So it lacked a lot of "official" bugfixes. Even later versions
> (e.g. 7.03 from early 1999) only claimed IBM PC-DOS 6.0 compatibility
> (via version check). My DR-DOS 7.03 Lite copy has no built-in LFN nor
> FAT32 nor LBA support. It can run Win 3.1 but nothing beyond that (no
> Win9x). So no, it's not really 100% compatible.

Wait a second, hold on. You need to explicitly specify what
"compatible" is because you're painting an incorrect picture about
what DOS compatibility is. Plain jane DOS typically means 4.0 or
later when support for 2G partitions was added. In the above
paragraph, you're holding DOS up to a standard that only the Windows
95 back-end could provide like LFN, FAT32, the ability to run Windows
95, etc. -- DOS could never do that and never claimed to. So it is
quite disingenuous to state that OpenDOS isn't really "100%
compatible" because you're not stating what you're comparing it to.
Is it not compatible with the back-end DOS "7" that came with Windows
95? No, it's not, but I don't consider that DOS. If you do, then you
need to state that up front.

Up to MS-DOS 6, all three major DOSes (MS-DOS, PC-DOS, and DR-DOS)
kept their version numbers in sync with each other. After that, they
stopped trying. Windows 95's "MS-DOS 7" (which was never available as
a standalone product and does not contain the full suite of utilities
that MS/PC/DR-DOS 6 has) has LFN and FAT32 support while PC/DR do
not. So you can't say "OpenDOS 7 isn't compatible with MS-DOS 7"
because there isn't an MS-DOS 7, at least not properly. It's the back-
end provided with Windows 95.

If you want to call it DOS 7, or the ME version DOS 8, then go ahead
but don't consider other DOS versions as being inferior because they
don't have the same features. 95/98/ME DOS versions were never
available outside of a 95/98/ME installation and aren't utility-
complete.

> (BTW, yeah, FreeDOS is a bit lacking in Win 3.1 support, but
> apparently nobody cares, esp. me! Heh. But I guess it wouldn't hurt to
> fix that ... in theory.)

Now that *is* a problem, since any DOS should be able to run Win 3.1.
If it doesn't run, something wrong with the DOS and I would be worried
about running anything else under it.

> No surprise. MS-DOS and Win9x were always considered unstable as hell.
> Also, even as far back as 1985, text mode was considered obsolete by
> MS, which is why they started developing Windows so heavily.

That's not true at all if you look at their product line. They were
maintaining/upgrading mainstream textmode products well into 1991 (MS
Word, Visual Basic, MS Works)

[Rugxulo]

Hi Jim,
FYI, I'm not really disagreeing with you, but we're probably just
splitting hairs here,

On Sep 27, 9:50 am, Jim Leonard <mobyga...@gmail.com> wrote:
> On Sep 26, 10:58 pm, Rugxulo <rugx...@gmail.com> wrote:
>
> > > and it is 100% compatible with MS-DOS up to a certain point because of a lawsuit with Caldera.
>
> > No, not really. OpenDOS 7.01 wasn't even the final Novell version
> > since the version control system was (allegedly) archaic and hard to
> > use. So it lacked a lot of "official" bugfixes. Even later versions
> > (e.g. 7.03 from early 1999) only claimed IBM PC-DOS 6.0 compatibility
> > (via version check). My DR-DOS 7.03 Lite copy has no built-in LFN nor
> > FAT32 nor LBA support. It can run Win 3.1 but nothing beyond that (no
> > Win9x). So no, it's not really 100% compatible.
>
> Wait a second, hold on.  You need to explicitly specify what
> "compatible" is because you're painting an incorrect picture about
> what DOS compatibility is.  Plain jane DOS typically means 4.0 or
> later when support for 2G partitions was added.  In the above
> paragraph, you're holding DOS up to a standard that only the Windows
> 95 back-end could provide like LFN, FAT32, the ability to run Windows
> 95, etc. -- DOS could never do that and never claimed to.

"Never claimed to" would be a stretch. In particular, DR-DOS did
indeed do all of those things and claim all of them. Unfortunately,
for whatever reason, some of it wasn't made publicly available and was
only used in-house (and for legal reasons).

> So it is
> quite disingenuous to state that OpenDOS isn't really "100%
> compatible" because you're not stating what you're comparing it to.

OpenDOS isn't even fully bugfixed. It doesn't even have all the Novell
patches. There were some ugly bugs left unpatched there, so the public
never got any of the benefits of that. Well, you could argue the
public never got much of anything thanks to the whole "open" deal not
working out. But anyways, I'm just saying, compared to MS-DOS or even
later DR-DOS versions, I would consider OpenDOS "incomplete" in
several ways (sadly).

> Is it not compatible with the back-end DOS "7" that came with Windows
> 95?  No, it's not, but I don't consider that DOS.  If you do, then you
> need to state that up front.

Well, indeed I do consider MS-DOS 7 to exist because the version
number (and check) reports that. And you can easily unbundle it and
run it stand-alone if desired. So yeah, I consider it true DOS,
although admittedly they didn't sell it separately (for purely
marketing, not technical, reasons).

Specifically, I was responding to Rod's opinion, where he says: "If

Caldera kept up, that should cover everything upto the final version
of MS-DOS, v8.0 that came with ME."

No, I personally don't believe you need to support FAT32, LFNs, nor
Win95 support to be "true" DOS. But indeed those are things that MS-
DOS 7.1 did support, and many have tried (or failed) to emulate.

> Up to MS-DOS 6, all three major DOSes (MS-DOS, PC-DOS, and DR-DOS)
> kept their version numbers in sync with each other.  After that, they
> stopped trying.

Not really. They were competitors, and they had very different
codebases (eventually) because they all hated each other, heh, ugh.
Let me quote you a line from the DR-DOS Wikipedia page:

"DR DOS version 5.0 (code-named "Leopard") was released in May 1990,
still reporting itself as "PC DOS 3.31" for compatibility purposes,
but internally indicating a single-user BDOS 6.4 kernel. (Version 4
was skipped to avoid being associated with the relatively unpopular MS-
DOS 4.0.)"

Internally it isn't even consistent either because a lot of that has
to be for (strict) software compatibility only. Externally of course
is just bogus marketing. So really version numbers mean almost nothing
these days.

In short, I think DR-DOS 5 was akin to MS-DOS 3.3, DR 6 was akin to MS
5, and DR 7 was akin to MS 6. (!!!)

> Windows 95's "MS-DOS 7" (which was never available as
> a standalone product and does not contain the full suite of utilities
> that MS/PC/DR-DOS 6 has) has LFN and FAT32 support while
> PC/DR do not.

Original Win95 didn't have FAT32, it was added later in OSR2 or
whatever. Last Win95 was OSR2.5 or such. I think even later Win95 and
Win98 both reported 7.1 while (crippled?) WinME reported 8. (And as we
know, none of them natively supported LFNs outside of running Windows
proper, sadly.)

Anyways, the only reason MS bundled DOS with Win95 was to avoid
competition. They wanted the market all for themselves, so they saw no
reason to bother with anyone else. It wasn't for technical reasons (as
they admitted for the court case). So MS-DOS 7 could've been sold
standalone, it just wasn't, for marketing only.

> So you can't say "OpenDOS 7 isn't compatible with MS-DOS 7"
> because there isn't an MS-DOS 7, at least not properly.  It's the back- end provided with Windows 95.

Win95 used some undocumented calls and data structures, which DR-DOS
never had access to. So out of the box it could never run Win95. (You
could ask why you would want to if you already have MS-DOS and it
works, but that's a different matter.) That wasn't their fault. They
did hack up a quick patch prototype (WinBolt?) but only tested it in
court and never released it for public consumption.

> If you want to call it DOS 7, or the ME version DOS 8, then go ahead
> but don't consider other DOS versions as being inferior because they
> don't have the same features.  95/98/ME DOS versions were never
> available outside of a 95/98/ME installation and aren't utility-
> complete.

I don't blame anybody for not supporting undocumented stuff like
Windows. Frankly, I (almost?) don't see the use, but it would still be
nice to have for completeness.

The FAT32 stuff is a whole ball of wax and probably not easy to add to
an existing kernel. LFN is patented since 1997, so it's not "free"
until 2017 or such. I'm blindly guessing that's why DRLFN was recalled
(so either that or it was buggy).

It's already been said that Caldera gave up on DOS, and clearly they
were only interested in other things (embedded Linux?). So they didn't
put in any effort after the 1990s. So it's effectively dead (or frozen
in a spin-off, DeviceLogics, which still sells 7.03, dated from early
1999).

> > (BTW, yeah, FreeDOS is a bit lacking in Win 3.1 support, but
> > apparently nobody cares, esp. me! Heh. But I guess it wouldn't hurt to
> > fix that ... in theory.)
>
> Now that *is* a problem, since any DOS should be able to run Win 3.1.

Well, except it's a two-way street. You can't call Windows DOS and DOS
Windows, can you? No, I don't like incompatibilities, but at the same
time, how do you expect to run something that's old, unsupported,
buggy, undocumented, hard to find, expensive, etc. (Okay, that's a
stretch, but you get the idea.)

> If it doesn't run, something wrong with the DOS and I would be worried
> about running anything else under it.

No, again, I think it's undocumented stuff. Win 3.1 runs under FreeDOS
but only in standard mode, not with 386 Enhanced shelling out all the
DOS boxes. It probably has some low-level minor bug. Search the
mailing list, it probably tells somewhere.

Seriously, though, it could be fixed (of course), esp. since both
DOSEMU and DOSBox support it (!), but there just aren't enough
volunteers. Honestly, everybody seems to want to do different things,
so it's hard to organize. (Not that I've tried, I'm not really much
help, sadly.)

If you want to rally support, feel free! But even hardcore retro folks
like us would probably find it hard to care about Win 3.1 these days.
You thought DOS was hard to run on new (multi-core / Ghz / GB RAM)
machines, try Win 3.x!!! :-)

> > No surprise. MS-DOS and Win9x were always considered unstable as hell.
> > Also, even as far back as 1985, text mode was considered obsolete by
> > MS, which is why they started developing Windows so heavily.
>
> That's not true at all if you look at their product line.  They were
> maintaining/upgrading mainstream textmode products well into 1991 (MS
> Word, Visual Basic, MS Works)

They long ago professed to make everything GUI only, which was "the
future". They laughed at obsolete things like (TUI) Topview and
Desqview. Maybe it was Mac envy, who knows? Or maybe Xerox really was
that impressive. I'm just pretty sure they never thought of textmode
as ideal. Only with Win8 Server will they finally have a non-GUI
version (optional, of course). It's just not interesting to them. But
they probably keep the console around to avoid the classic Mac (pre OS
X) mistake of not having one.