On Mon, Apr 12, 2010 at 9:10 AM, Michael C. Robinson
<plu...@robinson-west.com> wrote:
> DOS is not a good candidate for multitasking.  Why DOS can run in less
> memory and with older hardware than NT style systems and Linux style
> systems can, others who are more knowledgeable can comment on that.

I'm not saying that DOS could be a replacement for Linux or NT-based
Windows, but if one were looking for where to expand its
functionality, it seems to me that multitasking is the obvious next
step, because it was done, successfully, in the 1980s. It's nothing
very new.

> I am not impressed with Freedos network support or printer support for
> that matter.

In the days of DOS, these were the jobs of add-on 3rd party systems,
they weren't part of the OS.

> If someone could port CUPS to Freedos, that would be
> really nice.

Not much use, since no DOS app knows how to print to it. DOS apps
contain & manage their own printer drivers.

> As far as networking, someone should make a GPL clone of
> Microsoft Client and possibly others should start making DOS drivers for
> modern network cards and release those under a GPL license.

Well, it would be very cool, sure, but since the "real thing" is there
and freeware, why not use it? Quite a lot of NICs have NDIS2 drivers
that work with DOS, anyway. There is already an existing standard.

There's also Novell's ODI standard and DOS client, which I think was
also freeware, but it was designed for connecting to Netware servers,
which is effectively a dead system now.

>  Linux has
> Freedos beat hands down for hardware support including sound card
> support, but I suppose porting Linux drivers to Freedos could be very
> difficult because Freedos doesn't protect the hardware the way Linux
> does and because application programs typically try to run the hardware
> directly.

Which drivers do you need? Again, a lot of this is app-dependent
stuff, in DOS terms.

My notional graphical X.11 GUI would need drivers, but that's why I
suggested using X.org. It already has lots of drivers.

> In the same sense that DOS is not a system that one wants to multitask
> on, DOS is also not a system that one wants to support multiple users on
> because there is zero as in no file protection.  DOS systems do not
> support the concept of this file belongs to this user and that file
> belongs to that user and so on.

No, although CDOS and so on resolved that. But I am not suggesting DOS
as a fileserver. I installed lots of such systems in the 1980s -
mostly 3Com 3+Share - and it's a nightmare I never want to go back to.

But for a low-end single-user machine, I don't think we'd need it.

> Dosbox is a nice way to go, but Linux probably isn't the fastest host
> system in the world.  Perhaps Freedos 32 needs to be revived and a gui
> developed for it that can run Dosbox.

Why? DOSbox already runs on modern OSs that fully support modern
hardware. What benefit would there be in fitting DOS with a DOS
emulator? That seems a bizarre idea to me.

> I think the future for Freedos is seeing how much hardware power you
> really need to emulate the typical PC of days gone by.  Is a first
> generation Pentium that is too slow now for Linux fast enough to
> emulate the typical IBM PC and run Freedos?  I say virtualization
> and emulation are the future because it may be difficult or impossible
> to drive modern hardware that DOS applications were not written for
> in Freedos running natively.  An emulator can translate calls for an
> old sound blaster style card to drive a newer Intel integrated sound
> system.  Otherwise, I encourage people to check out games like Dirk
> Dashing Secret Agent to get a taste of what can be done natively on a
> Linux system.

It's already been done, on Linux and Windows. There is absolutely no
reason to do it on DOS itself.

> I think a discussion of why Microsoft abandoned DOS for NT is in order
> before people go crazy about enhancing Freedos.  DOS was never intended
> to support multiple users and multiple processes let alone contain badly
> written software.

No, true, it wasn't, but these problems were solved. I think a
multitasking shell on DOS would be a logical next step in enhancing
the OS for users of old hardware with limited resources.

> At a certain point, one has to limit their expectations for Freedos or
> else the system will disappoint.  The point of Freedos is to be able to
> use very old computers and run old software that predates Windows NT and
> Linux.  I would like to see a simple gui developed for Freedos that can
> run Firefox, but even that is probably getting too far away from what
> DOS and Freedos in particular are for.

That's exactly the sort of thing I'm proposing: memory management,
multitasking and a standard X.11 GUI. Exactly the sort of thing you'd
need to port, say, Linux Firefox over. Although I suspect current
Firefox might be a bit "heavy" - but Firefox 2, say, would be OK.

> A true DOS system has zero intelligence about the software being run.
> Viruses etcetera are a serious problem in a DOS environment.  Emulating
> DOS with an emulator that catches bad code before that code has a chance
> to execute is probably a doable thing.


There were tons of antivirus programs for DOS. I am sure we can find
something and update its signatures from those in ClamAV or something.

> I want to see a new release of Freedos with fewer bugs as much as the
> next guy, but before that can happen I think a few things have to be
> considered:
> 1) Interest in freedos will drop even further if a new release doesn't
> run on 386 and older computers.

Older than 386? Really? How many more-than-23-year-old PCs are still
in existence and working, do you think?

I submit that anyone with such an antique is probably a collector and
would therefore run an historically-accurate OS on it, such as MS-DOS

> 2) Freedos is not a multitasking OS or a true networking OS where trying
> to make it one while ignoring other OSes that fit that description
> better is going to be unpopular.  The single task nature of Freedos can
> be a strength.

How, exactly?

> 3) Freedos needs to become Windows 9x compatible at some point or else a
> different gui running on top of Freedos needs to be popularized.  This
> Windows 9x replacement needs to run on hardware as old as the late model
> 486.  This gui needs to make more network cards and printers work on top
> of Freedos than would typically work in a program called from the
> command prompt.

Absolutely ludicrous. Not only a collossal, near-impossible task, but
also one that would violate hundreds of Microsoft patents and IP

If you want a FOSS Windows clone, which to me is a bizarre and foolish
idea, just waiting to be sued into oblivion by one of the most
malicious and litigious companies on the planet, then there is
ReactOS. Brilliant, insane, pointless and doomed, but it's there.

> 4) As we get further and further away from the "good old days" where
> everyone ran a commercial version of DOS, open source DOS software is
> going to become more important and be legal to use.  Popular closed
> source software, even Wordperfect for Dos, won't be legal to use.

I am not sure about this. Whereas there is a lot of dodgy
"abandonware" about, there is some that is kosher, and there was a ton
of shareware and PD software that is 100% legal.

> Freedos doesn't seem to work well on multiprocessor systems

And it never will, without multitasking, and even then, I think that's
over-reaching dramatically.

> nor does
> Freedos support a lot of modern hardware well.

True, but then, DOS never did have much hardware support. That's the
apps' problem.

>   The support for modern
> hardware issue can be solved by reviving Freedos 32

Precisely the thing I was arguing /against./ :¬)

>  and emulating the
> older hardware that the modern stuff is replacing.

Frankly, I think that's a ridiculous idea.

>   Memory protection
> and memory management are doable on x86 computers where the IA32
> architecture especially has hardware to assist with this.  Computers are
> becoming fast enough that hardware calls can be captured and translated
> on the fly.

Which was one of my core points.

Liam Proven • Profile: http://www.linkedin.com/in/liamproven
Email: lpro...@cix.co.uk • GMail/GoogleTalk/Orkut: lpro...@gmail.com
Tel: +44 20-8685-0498 • Cell: +44 7939-087884 • Fax: + 44 870-9151419
AOL/AIM/iChat/Yahoo/Skype: liamproven • LiveJournal/Twitter: lproven
MSN: lpro...@hotmail.com • ICQ: 73187508

Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
Freedos-user mailing list

Reply via email to