On Thu, Jan 3, 2013 at 2:07 AM, Michael Robinson
<plu...@robinson-west.com> wrote:
> Dos makes sense for 8/16 bit computers that can't handle multitasking
> very well.

[MS-based] "DOS" is exclusively 16-bit, it won't work on 8-bit at all
(at least not in "traditional" sense or any reasonable variant that I
know of).

> There are plenty of 8/16 bit computers around still, think
> e-readers probably and other embedded devices that don't need the
> higher functionality a 32/64 bit machine/multi core machine offers.

I would guess that most e-readers (Nook, Kindle ... at least Kindle)
use some variant of Android (modified Linux) or similar. I don't think
you'll have luck convincing many of them to switch to DOS. Even
OpenPandora uses Linux. (But feel free to look up the old PoqetPC
[sic] on eBay.)

> Dos was a quick and dirty operating system that was needed when personal
> computers were far less powerful and something cheap was needed now.

Sure, it was meant for the original IBM PC back when MBs of RAM
weren't even commonly available. Of course, with DOS extenders, we're
not stuck to such low amounts (for good or bad, hence some apps
actually use too much).

> Nowadays, except for special purpose/embedded devices, DOS doesn't make
> much sense.  As we move away from the original BIOS model, I hope dosbox
> gets updated sufficiently.  There is real time Linux, I don't know much
> about it though.

A multitasking DOS could work (and/or networking, USB, etc.), or even
just decent support for the DOS API, as has been proven with modern
Windows' NTVDM (well, more or less) and DOSEMU (which used to also
work on FreeBSD, but ...). Except that a lot of things demand GUI
(esp. Unicode) and refuse to work otherwise, for good or bad.

> About the time that Windows 95 came out, DOS lost official support.

DOS was well documented and widely supported. And while it sounds
trite (and is an unfair generalization), it seems that MS wanted to
"own" the market with their own proprietary 32-bit system and thus
defeat IBM's OS/2. (This was a marketing strategy learned from AT&T
and Xenix, you must "control the standard" or always play catch up /
pay up.) Sadly, instead of making things easier, that transition made
things harder. But they were wildly successful with Win9x. This is
where Bill Gates made most of his billions. The sad part is that all
the DOS support of Win9x was better than NT, so we've regressed.  :-(

> Microsoft should have gone straight to NT, but Microsoft didn't.
> Windows 9x is a nasty quasi dos/partial implementation of Win32.

WinNT needed gobs of RAM that Win9x didn't. You could (in theory) run
Win95 on a 4 MB machine. Not so for NT, not sure how lean, but it
required something like 16 MB of RAM and 80 MB of hard disk space (and
of course such requirements would grow exponentially with every

WinNT is obviously more stable and has a few more features, but sadly
it never fixed various bugs in its DOS subsystem (NTVDM), not even the
obvious ones. Sure, they added some stuff, but it seems like politics
got in the way of "fixing" even the good stuff with XP. Worse is that,
as time went on, more and more DOS apps were available, so it seems
even less wise to worsen (or abandon) support. But alas ... some
people don't care (they throw away 90% for the 10%).

> It isn't DOS and it isn't NT.  Unless your computer is a 16 bit
> 286 or older machine, Linux will run on it.

ELKS (Linux) will run on a 286, not sure about others. And of course
it's not well supported, maybe not even developed anymore. It had a
few niceties, but overall it wasn't fully fleshed out. For my uses, at
least, DOS makes more sense (more or less).

> A 386 won't run a
> modern Linux distribution most likely and it definitely won't
> run Firefox, but chances are good that it will run Freedos
> directly or a pared down Linux system with dosbox.

A 386 definitely won't run DOSBox, or at least nowhere near native
speed. A 1 Ghz machine is recommended for 486 speeds!! Besides, DOSBox
is hardcoded to (default 16) max 64 MB of RAM.

True, you can't really run most any modern Linux (or *BSD) on a real
386 anymore. In fact, I think they just sheared out the last vestiges
of 386 support (XADD, CMPXCHG). Most assume 586 or 686 features these
days, and that's not counting RAM requirements or peripherals or

> The 90's were
> a sad period for diversity of general purpose computers and DOS as well.
> The Tandy Color Computer III for example disappeared in the
> early 90's and the Commodore seems to have gone away as well.

Not sure about Tandy, but Commodore basically went bust in 1994. And
they were far from their heyday in the '80s. (At least Jack Tramiel
was head at Atari from 1985 until it folded in 1996.)

> The DEC Alpha basically failed on the market about 1998 or so.

Dunno the details. DEC got sold to Compaq in 1997, and HP bought
Compaq circa 2002. I don't think the Alpha was truly dead until 2007
or 2008 or such. From what I hear, it's only dead because they would
rather focus on other (almost dead??) systems like Itanium (or the
still vibrant AMD64).

> Sun Microsystems is no more and I wouldn't be surprised to hear that
> Sun stations are no more.

Dunno, check with Oracle. You still see various *nix things claim
SPARC support (and other semi-rareties like MIPS or PPC or ARM or
whatever), but it does seem obvious that those are less of a focus
than mainstream x86/x64 desktops.

> General purpose computer diversity has
> diminished, but cell phones and tablets vary widely at least as far
> as hardware and software.  Hopefully, Microsoft's attempt to dominate
> tablets won't lead to another period of Monopolization.

It's too green, too much change over too short a time. I'm not
impressed. Lots of soggy bread over too many waters ....

> I blame the consumer for the Microsoft domination and the decline
> of all of the competitors except Apple.

A lot of it is computer gaming. Windows is the de facto marketplace
for that, no other OS comes close (yet, but who knows about Linux +
Valve's Steam). And a lot of that is tied to fancy GPU hardware and
similar. (At least that's my impression: Windows for gaming, *nix for
servers, Apple for multimedia).

> Apple being 50% owned by Microsoft isn't much of an alternative though.

I heard that Apple long ago bought back all their shares from MS. They
have much more money and popularity too (even if MS still is very
successful and claims 1.3 billion users).

> As long as there are old computers, special purpose computers doing
> real time work, and modern computers that can emulate older ones
> Freedos should have a future.

The problem isn't so much FreeDOS or support as it is that nobody can
standardize on anything. Even somebody like Brian Kernighan says he'd
take C with him to a desert island. (No offense, but reminds me of
DOS!) Yet we have excellent support for C on FreeDOS and still heavily
lack developers. Why? Because it's never enough. Developers (and
users) seem to jump at every feature in the world, and instead of
getting decent support for a few reliable things, they get weak
support for everything. Creeping featurism. They take the pessimistic
approach of only seeing what you can't do instead of what you can. (I
just hate when DOS support is 99% working but abandoned!!)

> Freedos without Win32 though leaves
> out a lot of software that has kinda fallen through the cracks as
> some of this stuff doesn't even work in NT.  I think ReactOS which
> is based on NT is the better way to address that than adding a
> Windows 9x compatible gui to Freedos.

Windows offers a lot to developers, apparently. Not sure why they
bother, though, as it's basically throwing away your effort every few
years. It's not as stable as you'd think. And backwards compatibility
should be a priority, but developers (and MS) seem more intent on
creating new things in hopes of future popularity rather than keeping
old stuff stable. Face it, everything old is frowned upon, it's not
hip or cool or fun, hence people avoid it. It doesn't matter how
unfair that view is, that's just the way it is. When the novelty
aspect isn't there, it won't "sell".

> I question the wisdom of maintaining Freedos long term.  Sooner or
> later, people will have to emulate for Freedos anyways which means
> that they will have to deal with a modern operating system and modern
> hardware.

Modern hardware is a nightmare. Look at all the changes to modern OSes
over the past 5-10 years. It's ridiculous how much bloat is going on,
not to mention the hideous Borg that is HTML5 and similar
technologies. What happened to keeping things simple? Why are people
basically writing a whole new subsystem into web browsers? Why so much
focus on multimedia? And yet that's what sells.

> What will be needed going forward is capable software that
> works on other systems.  If you like Lotus 123 for example, you might
> like LibreOffice or GNU cash.  Syllable may take off and will likely
> support software programs that are unique to it.  Trying to address
> the need for a system compatible to one that was popular in the past
> is a difficult proposition, and there may be little interest if the
> public as a whole is enamored with newer computers that can do more.

True, people want new and fresh, they (mostly) don't stick to classic
and stable software.

But there are still a few systems that try to be backwards compatible:
 ReactOS, Haiku, AROS, FreeBSD, etc. (And we're lucky others even
bother being source code compatible!)

> The wisest thing to do perhaps is to help projects like Syllable and
> ReactOS and pressure legislators to crack down on Microsoft's strangle
> hold on the software market.

Don't waste time with legalities, it's of no use. It's better to just
not stress out over it, just use what works, don't waste too much time
worrying and being miserable.

Though honestly I'd rather blame hardware vendors for breaking
hardware compatibility (and thus software compatibility) than anyone
else, but it's unlikely to influence anyone. Getting working drivers
is one of the biggest problems for OSes.

I guess the only thing really worth doing is learning to write
"strictly conformant" portable code (instead of Win/*nix/Mac only
stuff that will be obsolete in a year or two). But it's quite
difficult and may not pan out (esp. if you're too heavily into
networking or GUI stuff). In this way, I'm kinda glad for a POSIX
utilities base that is more or less ubiquitous (even if a lot of it is
overkill or arcane). At least you can get "something" done portably,

> Freedos is in pretty good shape where
> the developers can probably do more good by working on other systems
> that are positioned to take better advantage of modern hardware.

It's probably wiser to work on hypervisors or emulators than abandon
DOS entirely. Or work on more ports of "portable" compilers (and
interpreters and runtimes) so that code can be exchanged with minimal
effort. But that too is quite difficult (esp. when POSIX is too
heavily (mis)used or things are badly designed).

Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
Freedos-user mailing list

Reply via email to