Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Linux is NOT a real alternative for DOS in it's present state !!!

0 views
Skip to first unread message

Arie Maaskant

unread,
Jan 24, 1995, 11:10:30 AM1/24/95
to

I use a Linux-system at home and I am very statisfied with it. However I am
used to work on Unix-systems for more then 10 years now.

However I do not think, that Linux is a real alternative for DOS users, that
are totally dependent of their system in a production environment. The most
important commercial software can be run on older DOS-versions then the latest
ones. Company's that make software should be stupid otherwise.

The last time I bought a CD with Linux on it, was Oktober 94 ( so only 3 months
ago ). It had the kernelversion 1.0.9 on it. However I had also an MSDOS
partition on it ( doublespaced ) and wanted to be able to read this partition
too. Well, one needs a kernel higher then app. 1.1.18. If you want to use
DOS under Linux, you'd better use a higher version then app. 1.1.60.
These are only two examples.

As a matter of fact one needs to use the latest kernel, to be sure to be
able to run the last applications. For an average DOS user, it is NOT
realistic to upgrade the kernel every day.

My 2 advises:
1) There should be a stable kernel for approximately one year. It is not
necessary that it supports all the latest hardware. The average user
is only interested in newer/fancier/faster etc. hardware when there
is a stable and bugfree kernel for it. So if you want a stable Linux
system then take a selection out of the hardwarelist, that belongs
to the last stable kernel.

2) Now the more important thing I think. All the applications, that are made
for Linux, should run and should be able to compile on this last stable
Linux version and not only on some experimental version, that changes
every day. If there are applications made for newer versions, they
should be called "testversions" or "betaversions" or whatever. The
average DOS users, that want to be converted to Linux could avoid these
versions or jump on the bandwagon of the everchanging kernel.:{

These betaversions can only be called "official" versions, when there
arrives a new "stable" kernel.

To convince a DOS user to try Linux, one should offer him a ( minimal )
and STABLE system ( including applications ), or else Linux stays a
system for hobbiests. ( AND A VERY GOOD SYSTEM, I DON'T WANT TO BE
MISUNDERSTOOD :-)
--
Arie Maaskant University of Amsterdam
Email: a...@fwi.uva.nl Faculty of mathmatics and computer sciences
Phone: +31 20 525 6438 Plantage Muidergracht 24
Fax: +31 20 525 5101 1018 TV Amsterdam

Pixelated!

unread,
Jan 24, 1995, 4:44:56 PM1/24/95
to
In article <3g38pm$h...@bit.fwi.uva.nl>, Arie Maaskant <a...@fwi.uva.nl> wrote:
> However I do not think, that Linux is a real alternative for DOS users, that
>are totally dependent of their system in a production environment. The most
>important commercial software can be run on older DOS-versions then the latest
>ones. Company's that make software should be stupid otherwise.

Try running Windows 3.11 on Dos 2.1. See how far you get. :)

>The last time I bought a CD with Linux on it, was Oktober 94 ( so only 3
> months
>ago ). It had the kernelversion 1.0.9 on it.

Yes, and it had source up to 1.1.59 or so, if you bought the InfoMagic one.
It couldn't have been more than 1 or 2 patchlevels when it came out. Don't
blame the CD vendor for not being able to keep up.

> However I had also an MSDOS
>partition on it ( doublespaced ) and wanted to be able to read this partition
>too. Well, one needs a kernel higher then app. 1.1.18.

That's because they hadn't written the thsfs (doublespace filesystem) yet.
Anyone who wants to use it needs to upgrade. If you want a guaranteed
stable kernel, stay with 1.0.9. That's what it's there for. Besides, I don't
really trust DoubleSpace, even if it didn't trash my HD for the six months I
had it installed. I'd never advise a commercial production environment to
use disk compression--I'd suggest more hard drive space.

> If you want to use
>DOS under Linux, you'd better use a higher version then app. 1.1.60.
>These are only two examples.

DOSEMU's been around a *lot* longer than that. Try 1.1.20s or so.

> As a matter of fact one needs to use the latest kernel, to be sure to be
>able to run the last applications. For an average DOS user, it is NOT
>realistic to upgrade the kernel every day.

If a 486 is available, it'll only take a half-hour to upgrade once a week
or so. Or maybe every 2 weeks. I know the last job I had (full time, even)
left me enough free time to be able to.

> My 2 advises:
>1) There should be a stable kernel for approximately one year. It is not
>necessary that it supports all the latest hardware. The average user
>is only interested in newer/fancier/faster etc. hardware when there
>is a stable and bugfree kernel for it. So if you want a stable Linux
>system then take a selection out of the hardwarelist, that belongs
>to the last stable kernel.

That's 1.0.9. It' been around since at least July. I think it was even
around last January. If so, that's been one year. I'm sorry, I don't
see your point here.

>2) Now the more important thing I think. All the applications, that are made
>for Linux, should run and should be able to compile on this last stable
>Linux version and not only on some experimental version, that changes
>every day. If there are applications made for newer versions, they
>should be called "testversions" or "betaversions" or whatever. The
>average DOS users, that want to be converted to Linux could avoid these
>versions or jump on the bandwagon of the everchanging kernel.:{

That defeats the purpose of a code freeze although I suppose you could
try patching EIDE drivers (for example) into 1.0.9. Sticking with
old kernels means you limit yourself in what hardware support you have
available. I think a better idea would be a list of "stable" 1.1.x kernels.
F.e., 1.1.59 seemed o.k. to me; it never even hiccupped for me. ditto 1.1.72.

>These betaversions can only be called "official" versions, when there
>arrives a new "stable" kernel.

1.x.y series, wheree x is odd, *are* beta versions. when 1.2.0 comes out,
it'll be silly to call 1.1.x versions "official" as they'll be obsolete.

>To convince a DOS user to try Linux, one should offer him a ( minimal )
>and STABLE system ( including applications ), or else Linux stays a
>system for hobbiests.

Again, if they can live without newer hardware support and things like
VESA powesaver options, then they can use 1.0.9. But I'd rather use 1.1.72,
which was (in my opinion) rock solid, and had a *lot* of support for stuff.
I don't want to muck around with manually patching stuff like support
for my Panasonic CD-ROM, or the ide performance patches, into 1.0.9, though,
which is why I use newer versions. And you should read the article about
Virginia Power in Linux Journal #7. They use linux PCs to gather millions
of points of data per day. The article's author claims to have not lost
a single data point.

> ( AND A VERY GOOD SYSTEM, I DON'T WANT TO BE
>MISUNDERSTOOD :-)

I should hope not. :) But this comment seems strained given the rest of
your article, kind of like you're saying "well don't flame me, I was
just joking." Maybe you're not, I don't want to say you are, but it just
feels funny.


--
Richard Cooley Extraordinaire "Yeah. Arrgh."
pi...@gnu.ai.mit.edu These are my opinions, not MIT's.
rcoo...@dgl.ssc.mass.edu Linux Linux Linux Linux Linux

Christopher B. Browne

unread,
Jan 25, 1995, 12:07:27 AM1/25/95
to
In article <3g38pm$h...@bit.fwi.uva.nl>, Arie Maaskant <a...@fwi.uva.nl> wrote:
>2) Now the more important thing I think. All the applications, that are made
>for Linux, should run and should be able to compile on this last stable
>Linux version and not only on some experimental version, that changes
>every day. If there are applications made for newer versions, they
>should be called "testversions" or "betaversions" or whatever. The
>average DOS users, that want to be converted to Linux could avoid these
>versions or jump on the bandwagon of the everchanging kernel.:{

I would agree; I think that this problem is not a long term problem, as the
set of functionality is starting to converge towards the 1.2 release.

There's been a further problem with the proliferation of versions; not only
are there applications that "only work for version > 1.1.x," there are also
applications that "only work for version < 1.1.x." I haven't had the
patience (or the interest, or much need) to figure out why DOSEMU stopped
working for newer patch levels than 1.0.9. (It may be configuration; it may
be that I should upgrade DOSEMU; I use it so little that I really don't care
if I need to reboot with 1.0.9.)

Hopefully things will stabilize soon towards 1.2. From what I can tell, the
recent changes have related to trying to have a common code base for all
hardware architectures. That's not particularly useful to *me,* but I'm
willing to wait a little for it to happen. It appears that it's shaking out
bugs that happened to get counteracted by the vagaries of architecture.
Which means that multiple platforms is actually *good* for everyone.

But to those that are out there hacking the kernel, we *do* fairly
desparately need to have a "version 1.2.0" that can be considered pretty
stable.
--
Christopher Browne - cbb...@io.org
TeX or Word: Whose product would you rather to be using:
That of a Stanford Professor of Computer Science --OR-- that of a
Harvard freshman dropout?

Arie Maaskant

unread,
Jan 25, 1995, 4:09:45 AM1/25/95
to
newc...@aa.csc.peachnet.edu (Dan Newcombe) writes:

>In article <3g38pm$h...@bit.fwi.uva.nl> a...@fwi.uva.nl (Arie Maaskant) writes:

>>The last time I bought a CD with Linux on it, was Oktober 94 ( so only 3 months
>>ago ). It had the kernelversion 1.0.9 on it. However I had also an MSDOS
>>partition on it ( doublespaced ) and wanted to be able to read this partition
>>too. Well, one needs a kernel higher then app. 1.1.18. If you want to use
>>DOS under Linux, you'd better use a higher version then app. 1.1.60.
>>These are only two examples.

>That's bullshit. True: One can not read a doublespace drive under Linux
>without having the dos emulator running. (By that I mean: You can't mount

No, I do mean, there was a double-spaced filesystem reader made by somebody
in Germany, it was possible to mount a Double-spaced filesystem in read-
only mode. One only needed kernel version 1.1.18 ( I think ) with
enhancements ( loadable modules ). Overall it was OK, but certainly not
installable for the "average DOS-user".

> Sounds like a personal problem to me :)

No , it isn't. I do not think, that the "average DOS-user" is capable
of using "loadable modules".

>> As a matter of fact one needs to use the latest kernel, to be sure to be
>>able to run the last applications. For an average DOS user, it is NOT
>>realistic to upgrade the kernel every day.

>Then why do so? Like I said, I've been months behind the latest kernel with
>no problems. Usually when I get about 20-25 versions behind I upgrade.

Because, when you want to add some software and get that from an ftp-site
it is not clear if it is meant for "the last stable kernel" or not.
Of course you could put that information in a README file, but I do
not think, that that is the right place for it. Software that works with
an old, but stable kernel should be put in a different directory and
should be mentioned in some "stable-software-list". Other newer software,
that is compiled under and works with a newer kernel, should be put
in another directory tree. There of course you could mention these
dependencies in a README-file. This README file is then ment for
somebody who wants to use this often better software, but is not
necessary the be read by the "average DOS-user" (ADU).

>> My 2 advises:
>>1) There should be a stable kernel for approximately one year. It is not
>>necessary that it supports all the latest hardware. The average user
>>is only interested in newer/fancier/faster etc. hardware when there
>>is a stable and bugfree kernel for it. So if you want a stable Linux
>>system then take a selection out of the hardwarelist, that belongs
>>to the last stable kernel.

>wasn't 100% stable, so it was patched. That finally stabilised around 1.0.9.
>This was back in March of last year. True, not a year yet, but very close.
>Stability doesn't mean it must stagnate.

I agree totally. However what I mean is, that there should be two different
Linux trees, one that only contains a total package of kernelsources
and binaries that are stable ( I know off course, that nothing is 100%
stable ), with stable approved drivers for a selection of hardware and
a selection of software. If Linux really catches on, what I really hope,
also for the ADU, hardware manufacturers are maybe pushed to deliver
stable binaries and sources also for the Linux community. I do not think,
one can expect that a manufacturer makes software for an everchanging
kernel.

The second tree is where "the action is going" and could provide better
and faster software. It could contain betaversions for drivers for new
hardware, new software, that depends on new features and so on.
The only difference with the present situation is, that it should
be easy to be recognized as such. The average DOS user knows that he could
face problems, when he uses this tree.

>So, if 1.1.75 supports a new function that enables me to do something no
>previous kernel did, then all my software should be beta, even though if it
>works 100% because someone is still using 1.0.9. I don't buy that. It may

"Beta" is only a word and not a qualifier for me. Software that is dependent
on and is made on a system with a 1.1.75 should only be placed in the
other tree. It should contain README files and so on,that is readable
for the "dynamic" Linux users and therefore could contain information
about what kernel-version one should use and so on. I do not think,
that the average DOS user is interested in differences between
1.0.9 and 1.1.75.

>look nice on paper, but then again, so does socialism. It is very simple
^^^^^^^^^
And every "ism", feminism, capitalism, anarchism, katholicism, they all look
nice on paper, but that is a totally different discussion :)


>>To convince a DOS user to try Linux, one should offer him a ( minimal )
>>and STABLE system ( including applications ), or else Linux stays a
>>system for hobbiests. ( AND A VERY GOOD SYSTEM, I DON'T WANT TO BE
>>MISUNDERSTOOD :-)

>And that is what the Slackware distribution does. The most recent picked a
>stable development kernel instead of 1.0.9, but all the applications are
>compiled against that kernel and those libraries.

Yes, but Slackware is only a selection of all the software that is
available for Linux. And I thougt that on the latest CD's with Slackware
kernel 1.1.54 is the standard kernel and the numbering convention says
it is not a "code-freezed", stable kernel.

What I hope, is that Linux (or some other Unix-version ? ) becomes
really popular and wipes out MSDOS.
What is MSDOS more then CPM1.4 with some stolen Unix-enhancements ?
How is it possible that somebody has become so rich by means of such
software ( is that capitalism ? ) ?


> -Dan

>--
>Dan Newcombe newc...@aa.csc.peachnet.edu
>-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
>"And the man in the mirror has sad eyes." -Marillion

Arie Maaskant

unread,
Jan 25, 1995, 4:48:05 AM1/25/95
to
pi...@gnu.ai.mit.edu (Pixelated!) writes:


>Try running Windows 3.11 on Dos 2.1. See how far you get. :)

You do not have to convince me of that.

>>The last time I bought a CD with Linux on it, was Oktober 94 ( so only 3
>> months
>>ago ). It had the kernelversion 1.0.9 on it.

>Yes, and it had source up to 1.1.59 or so, if you bought the InfoMagic one.
>It couldn't have been more than 1 or 2 patchlevels when it came out. Don't
>blame the CD vendor for not being able to keep up.

But you certainly can not expect the "average DOS-user" to keep up.

> Besides, I don't
>really trust DoubleSpace, even if it didn't trash my HD for the six months I

Yes I made the decision to by another drive for DOS, I need it to
run "reisplanner" ( :) :) )


>> As a matter of fact one needs to use the latest kernel, to be sure to be
>>able to run the last applications. For an average DOS user, it is NOT
>>realistic to upgrade the kernel every day.

>If a 486 is available, it'll only take a half-hour to upgrade once a week
>or so. Or maybe every 2 weeks. I know the last job I had (full time, even)
>left me enough free time to be able to.

I have a 486DX50 with 16Mbyte memory and it takes me app. 20 minutes
to compile a kernel. But that is something for you and me but not for
the "average DOS-user"

>> My 2 advises:
>>1) There should be a stable kernel for approximately one year. It is not

>That's 1.0.9. It' been around since at least July. I think it was even


>around last January. If so, that's been one year. I'm sorry, I don't
>see your point here.

Ok, longer than I thought. There should be "code-freeze" by now.

>>2) Now the more important thing I think. All the applications, that are made
>>for Linux, should run and should be able to compile on this last stable
>>Linux version and not only on some experimental version, that changes
>>every day. If there are applications made for newer versions, they
>>should be called "testversions" or "betaversions" or whatever. The
>>average DOS users, that want to be converted to Linux could avoid these
>>versions or jump on the bandwagon of the everchanging kernel.:{

>That defeats the purpose of a code freeze although I suppose you could
>try patching EIDE drivers (for example) into 1.0.9. Sticking with

Yes that should be done too, to put it in a stable system.

>old kernels means you limit yourself in what hardware support you have
>available. I think a better idea would be a list of "stable" 1.1.x kernels.
>F.e., 1.1.59 seemed o.k. to me; it never even hiccupped for me. ditto 1.1.72.

I thought that when a kernel is supposed to be "stable" the second number
should be even (or zero ).

>1.x.y series, wheree x is odd, *are* beta versions. when 1.2.0 comes out,
>it'll be silly to call 1.1.x versions "official" as they'll be obsolete.

>>To convince a DOS user to try Linux, one should offer him a ( minimal )
>>and STABLE system ( including applications ), or else Linux stays a
>>system for hobbiests.

>Again, if they can live without newer hardware support and things like
>VESA powesaver options, then they can use 1.0.9. But I'd rather use 1.1.72,
>which was (in my opinion) rock solid, and had a *lot* of support for stuff.

I use myself 1.1.54 because I needed the doublespeed drivers for my
CDU33. I could also use 1.0.9, it works OK, but only in single speed
mode.


>I don't want to muck around with manually patching stuff like support
>for my Panasonic CD-ROM, or the ide performance patches, into 1.0.9, though,
>which is why I use newer versions. And you should read the article about
>Virginia Power in Linux Journal #7. They use linux PCs to gather millions
>of points of data per day. The article's author claims to have not lost
>a single data point.

You do not have to convince me of the virtues of Linux. I only hope, that
the "average DOS-user" will be convinced. By that I mean f.i. the
butcher around the corner, that uses his PC for bookkeeping and
the administration of the soccer-club. He is not interested in upgrading
his kernel. He has no access to internet. If he buys some new hardware
he wants to be sure that it works on his system. The manufacucturer
gives hime eassy to be installed drivers for DOS. The manufaturers only
makes drivers for Linux if there is a great enough interested for it.
I can not imagine that they develop these drivers when the kernel
is changing so rapidly.
I want to be Linux ( or some other Unix version ? ) to be really popular
and wipe out MSDOS. But one should also at least try to convince the
butcher around the corner, instead of an "elite" group.

>> ( AND A VERY GOOD SYSTEM, I DON'T WANT TO BE
>>MISUNDERSTOOD :-)

>I should hope not. :) But this comment seems strained given the rest of
>your article, kind of like you're saying "well don't flame me, I was
>just joking." Maybe you're not, I don't want to say you are, but it just
>feels funny.

No, it is not a joke.


>--
>Richard Cooley Extraordinaire "Yeah. Arrgh."
> pi...@gnu.ai.mit.edu These are my opinions, not MIT's.
> rcoo...@dgl.ssc.mass.edu Linux Linux Linux Linux Linux

Big Thunder Rumblefish

unread,
Jan 25, 1995, 9:39:29 AM1/25/95
to
In article <3g54gp$o...@bit.fwi.uva.nl>, Arie Maaskant <a...@fwi.uva.nl> wrote:

>newc...@aa.csc.peachnet.edu (Dan Newcombe) writes:
>
>No, I do mean, there was a double-spaced filesystem reader made by somebody
>in Germany, it was possible to mount a Double-spaced filesystem in read-
>only mode. One only needed kernel version 1.1.18 ( I think ) with
>enhancements ( loadable modules ). Overall it was OK, but certainly not
>installable for the "average DOS-user".
>
>> Sounds like a personal problem to me :)
>
>No , it isn't. I do not think, that the "average DOS-user" is capable
>of using "loadable modules".
>
Ummm.... Isn't this exactly what they do when AUTOEXEC.BAT and CONFIG.SYS
load their drivers into memory?


--
------------------------------------------------------------
| Kris Zoerhoff du...@freenet.grfn.org |
| http://www.grfn.org/~dust/ |
------------------------------------------------------------

Dan Newcombe

unread,
Jan 25, 1995, 1:24:23 PM1/25/95
to
In article <3g54gp$o...@bit.fwi.uva.nl> a...@fwi.uva.nl (Arie Maaskant) writes:
>No, I do mean, there was a double-spaced filesystem reader made by somebody
>in Germany, it was possible to mount a Double-spaced filesystem in read-
>only mode. One only needed kernel version 1.1.18 ( I think ) with
>enhancements ( loadable modules ). Overall it was OK, but certainly not
>installable for the "average DOS-user".

Sorry...I was unaware of the doublespace driver...I don't trust it
(Doublespace that is), and probably skipped over the driver announcement.
Learn something new everyday...

>No , it isn't. I do not think, that the "average DOS-user" is capable
>of using "loadable modules".

Sadly though, modules seem to be pretty easy compared to some of the other
stuff on a Linux system. If someone is moving from DOS, they are going to
have a harder time with some things like ppp/networking, and X. They are
going to have to want to learn some new stuff. It's very hard to have a
pre-slugged Linux system for all cases.

>>> As a matter of fact one needs to use the latest kernel, to be sure to be
>>>able to run the last applications. For an average DOS user, it is NOT
>>>realistic to upgrade the kernel every day.
>>Then why do so? Like I said, I've been months behind the latest kernel with
>>no problems. Usually when I get about 20-25 versions behind I upgrade.
>Because, when you want to add some software and get that from an ftp-site
>it is not clear if it is meant for "the last stable kernel" or not.

For most of the software, it is not the reliant on kernel versions. Only some
very hefty software, such as DOSEMU or Wine relies one some kernel versions.
This is mainly because they are trying to do very nasty things to the system
and need certain patches installed. A pain - yes...but needed.
The biggest problem I find is if I FTP some pre-compiled binary (to save
time), sometimes I don't have all the same library versions (libc, x libs,
etc..) that the person who compiled it had. This is where things start to
get ugly.

>Of course you could put that information in a README file, but I do
>not think, that that is the right place for it. Software that works with
>an old, but stable kernel should be put in a different directory and

The Linux Software Map usually lists these dependencies. I pesonally think it
would be easier for a user to check the lsm for a certain program before
ftping it would be a lot easier than having seperate directories. If I'm
looking for the latest version of Mosaic, I don't want to have to look in
.../linux/stable/mosaic
.../linux/unstable/X11R5/libc4.0.2/mosaic
.../linux/unstable/X11R6/libc4.5.2/mosaic
etc...

when I can easily go to
.../linux/system/Network/info-system/Mosaic
and find all the versions right there, along with readme's/lsm's telling me
what works with what.

>should be mentioned in some "stable-software-list". Other newer software,
>that is compiled under and works with a newer kernel, should be put
>in another directory tree. There of course you could mention these
>dependencies in a README-file. This README file is then ment for
>somebody who wants to use this often better software, but is not
>necessary the be read by the "average DOS-user" (ADU).

Perhaps a better idea that could suit both of use would be a stable directory
with links to programs/source/etc... that are known to work on that system.
Kernel differences usually don't break that many programs. That would make
for a very ugly system...if I had to recompile lots of stuff everytime I
upgraded my kernel, I'd be pretty pissed. The programs that a kernel upgrade
usually does break are those that the ADU would not usually be trying to
upgrade on their own (ps, e2fsck, etc...)

If the ADU is just trying to get a new version of sc because of new
functionallity, it's not going to matter much if the person that compiled it
had 1.0.9 or 1.1.50. The only thing that'll matter is what version of libc
someone had, and whether that is compatiable. If it's just source code, the
ADU should have even less of a problem.

>>wasn't 100% stable, so it was patched. That finally stabilised around
1.0.9. >>This was back in March of last year. True, not a year yet, but very
close. >>Stability doesn't mean it must stagnate.

>I agree totally. However what I mean is, that there should be two different
>Linux trees, one that only contains a total package of kernelsources
>and binaries that are stable ( I know off course, that nothing is 100%
>stable ), with stable approved drivers for a selection of hardware and
>a selection of software. If Linux really catches on, what I really hope,
>also for the ADU, hardware manufacturers are maybe pushed to deliver
>stable binaries and sources also for the Linux community. I do not think,
>one can expect that a manufacturer makes software for an everchanging
>kernel.

Just look at people with Motif. From what I read it doesn't work that well
with X11R6, so they have to upgrade. Fortunatly the X11 updates are not that
frequent in big ways like R5 -> R6. Libc upgrades happen, but less often.
One of the problems I see for a manufacturer is how to decide what you release
your binaries for. Do you go for 1.0.9 because that is supposed to be stable,
but maybe doesn't use the most up to date libraries, or do you go for the most
up to date stuff, but maybe exclude some older stuff. Either way, some people
won't be able to run the product. You could statically link it, but then you
may be running into memory problems. I'm glad I'm not a commercial developer,
and can just upload source!

>>So, if 1.1.75 supports a new function that enables me to do something no
>>previous kernel did, then all my software should be beta, even though if it
>>works 100% because someone is still using 1.0.9. I don't buy that. It may

>"Beta" is only a word and not a qualifier for me. Software that is dependent
>on and is made on a system with a 1.1.75 should only be placed in the
>other tree. It should contain README files and so on,that is readable
>for the "dynamic" Linux users and therefore could contain information
>about what kernel-version one should use and so on. I do not think,
>that the average DOS user is interested in differences between
>1.0.9 and 1.1.75.

To tell you the truth, I don't think the ADU is interested in Linux, but
that's a different ball game. It sounds to me that what you think would be
good is a
/pub/Linux/1.0.x
/pub/Linux/1.1.x
structure or something like that (trying to avoid words like beta and
unstable). I would also suggest a /pub/Linux/both tree or something where
source that works just as well on either tree would be placed.

For instance:
/pub/Linux/1.0.x
would have
a) program binaries (login, getty, bash, etc...)
b) libraries (libc, libm, libsvga, libX11) etc...
c) kernel
that would all be at a set/defined/agreed upon revision, so that all programs
under this tree would be compiled or guarenteed to work with this setup.
Sounds like you'd need some sort of committee or group, like the File Standard
people to work on this one. They would figure out when to upgrade, what
should be there, etc...
/pub/Linux/1.1.x
would have binaries, libraries, kernel, etc... that are linked or designed for
whatever version that happen to be for. Anarchy at best :)
/pub/Linux/both
would have source code for programs that don't rely on specific dependencies.
I say source only, becuase binaries would mean they were linked against
certain library versions, which could cause problems.

>>look nice on paper, but then again, so does socialism. It is
very simple > ^^^^^^^^^
>And every "ism", feminism, capitalism, anarchism, katholicism, they all look
>nice on paper, but that is a totally different discussion :)

True...but lets start it and cross post it to 20 alt. groups...uh, no.

>>>To convince a DOS user to try Linux, one should offer him a ( minimal )
>>>and STABLE system ( including applications ), or else Linux stays a
>>>system for hobbiests. ( AND A VERY GOOD SYSTEM, I DON'T WANT TO BE
>>>MISUNDERSTOOD :-)
>>And that is what the Slackware distribution does. The most recent picked a
>>stable development kernel instead of 1.0.9, but all the applications are
>>compiled against that kernel and those libraries.
>Yes, but Slackware is only a selection of all the software that is
>available for Linux. And I thougt that on the latest CD's with Slackware
>kernel 1.1.54 is the standard kernel and the numbering convention says
>it is not a "code-freezed", stable kernel.

It's not a frozen kernel. It is one that has proven itself to be stable.
Yes, it is just a subset of software, but I don't think you'll ever get all of
the software for a stable version. Lets say we decided 1.0.9, Xfree2.1, and
libc 4.4.1 were to be that stable version. I have 1.1.76, Xfree 3.1, and
a libc which is greater than 4.4.1. If I wanted to release something new or
updated, I am not going to hunt down a 1.0.9 compliant system to compile with,
I'm just gonna release the source and binary for what I have.

True, the standard committee I mentioned before could handle watching for new
sources, compiling on a standard system, but what if it's a binary only
release?

>What I hope, is that Linux (or some other Unix-version ? ) becomes
>really popular and wipes out MSDOS.
>What is MSDOS more then CPM1.4 with some stolen Unix-enhancements ?
>How is it possible that somebody has become so rich by means of such
>software ( is that capitalism ? ) ?

Actually, I think it's luck. :)

Matthew Cravit

unread,
Jan 25, 1995, 10:04:53 PM1/25/95
to
In article <3g56ol$p...@bit.fwi.uva.nl>, strange alien beings caused Arie Maaskant (a...@fwi.uva.nl) to write:

> Yes that should be done too, to put it in a stable system.

But if you patch the EIDE drivers into the 1.0.9 kernel, it may not be
stable anymore. It seems to me that you want both sides of the coin
here. The 1.0.9 kernel is stable as is. If you add EIDE drivers, and
double speed CD-ROM drivers, and so forth, you get...a 1.1.x
kernel.

> gives hime eassy to be installed drivers for DOS. The manufaturers only
> makes drivers for Linux if there is a great enough interested for it.
> I can not imagine that they develop these drivers when the kernel
> is changing so rapidly.

Hmm...I'm not sure I see the connection here. One of the primary
reasons the kernel keeps changing is that drivers keep being added to
it. Developers won't write drivers for it because they are afraid of
other drivers being added to it? I don't understand. The kernel should
not change enough from version to version to require complete rewrites
of the drivers, I don't think.

/Matthew
(Linux laptop...$1500 to go and counting :=) )

--
Matthew Cravit N9VWG | "After all is said and done,
DePaul University | more is said than done."
Chicago, IL 60614 | -- Author Unknown
crav...@interaccess.com http://www.interacces.com/users/cravitma

Warwick Allison

unread,
Jan 25, 1995, 2:11:09 AM1/25/95
to
I've often seen this claim that people feel they need to be forever
upgrading. YOU DON'T. Linux kernel 1.0.9 is more stable than Windoze
has ever been, so people slaver for new less buggy versions of Windoze.
But Linux users don't have to be on that treadmill.

I've only used 2 kernel versions:

1.1.19 (first slackware one with NCR SCSI support)
1.1.54 (I wanted to compile in sound because I got a SB16)

All programs run. Find a kernel that works for you (ie. supports your
hardware) and stick with it until you buy some more hardware, or want
to run some special new program (99.99% of programs run on even very
old kernels). Then when an `official' stable version arrives, switch
to that just for the sake of any performance increases you'll get.

We are NOT all kernel hackers and beta testers! Some people I'm sure
love hacking the kernel and finding bugs and fixes and would be hacking
Windoze code if it was free. Others have other Linux or Unix or X
projects to contribute and use Linux as a rock-solid foundation on
which to work.

--
Warwick
--
_-_|\ war...@cs.uq.oz.au \ Microsoft is not the answer, Microsoft
/ * <- Comp Science Department,\ is the question. NO is the answer.
\_.-._/ University of Queensland,)
v Brisbane, Australia. / Intel Inside? Don't Divide!

Bernd Bernie Meyer

unread,
Jan 25, 1995, 7:54:19 AM1/25/95
to
a...@fwi.uva.nl (Arie Maaskant) writes:

>The last time I bought a CD with Linux on it, was Oktober 94 ( so only 3 months
>ago ). It had the kernelversion 1.0.9 on it. However I had also an MSDOS
>partition on it ( doublespaced ) and wanted to be able to read this partition
>too. Well, one needs a kernel higher then app. 1.1.18. If you want to use
>DOS under Linux, you'd better use a higher version then app. 1.1.60.
>These are only two examples.

I think that the doublespace-FS (does it exist at all? Must have missed it)
certainly is ALPHA or BETA software. MS-DOS FS works flawlessly in 1.0.9.
The dosemu you tried was probably called something like "pre53_36...." or so
and comes from the /pub/linux/ALPHA/dosemu/Development directory. That
directory has a large README which tells you that you are dealing with work
in progress. Dosemu 0.52, OTOH, runs flawlessly on 1.0.9 :-)

>These betaversions can only be called "official" versions, when there
>arrives a new "stable" kernel.

Which is probably exactly what will happen.
And on a related note: I recently returned from a 1 year trip to Australia,
played back my backup of a pre-1.0 system, and started installing new
software. The only thing that did _not_ run was the new dosemu, and I knew
that beforehand.

Bernie

Frank Richard Ammeter

unread,
Jan 26, 1995, 5:38:11 AM1/26/95
to
Of course Linux is not an alternative for DOS. The average DOS user does
not need Linux, since he does word processing only. Very few people really
need a good OS: this explains Mircosoft's success.

--
|--------------------------------------------------------------------|
| |
| '[Hmm... this is actually pretty dumb...]' |
| Frank Ammeter (famm...@iiic.ethz.ch), Student of computer science |
| |
|--------------------------------------------------------------------|

Arie Maaskant

unread,
Jan 26, 1995, 8:48:32 AM1/26/95
to
crav...@flowbee.interaccess.com (Matthew Cravit) writes:

>In article <3g56ol$p...@bit.fwi.uva.nl>, strange alien beings caused Arie Maaskant (a...@fwi.uva.nl) to write:

>> Yes that should be done too, to put it in a stable system.

>But if you patch the EIDE drivers into the 1.0.9 kernel, it may not be
>stable anymore. It seems to me that you want both sides of the coin
>here. The 1.0.9 kernel is stable as is. If you add EIDE drivers, and
>double speed CD-ROM drivers, and so forth, you get...a 1.1.x
>kernel.

>> gives hime eassy to be installed drivers for DOS. The manufaturers only
>> makes drivers for Linux if there is a great enough interested for it.
>> I can not imagine that they develop these drivers when the kernel
>> is changing so rapidly.

>Hmm...I'm not sure I see the connection here. One of the primary
>reasons the kernel keeps changing is that drivers keep being added to

That's right, when a manufacturer adds a Linux-driver of his hardware
to the Linux community, it de facto should change the version number.
But the manufacurer could advertise, that it is compilable and/or
usable under the last stable Linux-version. That is the same as saying
that your driver works with DOS6.2 ( or whatever number ). The version
number of DOS doesn't change either, if you add a driver to it.

Dan Newcombe

unread,
Jan 26, 1995, 5:53:14 PM1/26/95
to
In article <3g897g$i...@hermes.fwi.uva.nl> a...@fwi.uva.nl (Arie Maaskant) writes:
>>Hmm...I'm not sure I see the connection here. One of the primary
>>reasons the kernel keeps changing is that drivers keep being added to

>That's right, when a manufacturer adds a Linux-driver of his hardware
>to the Linux community, it de facto should change the version number.
>But the manufacurer could advertise, that it is compilable and/or
>usable under the last stable Linux-version. That is the same as saying
>that your driver works with DOS6.2 ( or whatever number ). The version
>number of DOS doesn't change either, if you add a driver to it.

That is where the modules need to be concentrated on. Right now there are
three ways to implement a driver:
1) Have it be a part of the official kernel, which means a change to it is a
patch level increase of the kernel.
2) Have it be a standalone file, where the person that needs it must get
(ftp?) the file and patch it on their own. (And pray it works with the
current version of the kernel.
3) Have a loadable module, which doesn't affect the kernel source at all.

Of these three, I would say that the loadable modules (much like driver
loading in config.sys) would be the way for a manufacturer to go, and for most
of us to go also!

Joe Sloan

unread,
Jan 26, 1995, 2:55:16 PM1/26/95
to
Let's face the facts here:

Linux is a 32-bit, multiuser, multitasking, networking OS.
What in the world is the average DOS user going to do with such a thing?
To the DOS/windoze user, the OS is a program loader. period. It loads
a word processor, or a game, or something of the sort, and the DOS user
is happy.

On the other hand, for those who wish to break free of the limitations
of a single-user, single-tasking OS, who are of the hacker nature, who
want to be able to do their C/C++ or X programming at home, or want to
experiment with computer networking, linux is the inevitable solution.

Just bear in mind that a certain level of sophistication is needed
before a person will even feel the need for more than DOS/windoze.
The tragedy is that so often the user in need of more power will look
to miscro$oft to supply the $olution, at inflated prices and inferior
functionality to what is available in linux...

jjs

--
Email to: | Running Linux! (Slackware)
j...@dostoevsky.ucr.edu | because a 486 is a terrible
j...@ucrengr.ucr.edu | thing to waste...


Peter Mutsaers

unread,
Jan 26, 1995, 4:52:55 PM1/26/95
to
>> On 26 Jan 1995 03:04:53 GMT, crav...@flowbee.interaccess.com
>> (Matthew Cravit) said:

MC> Hmm...I'm not sure I see the connection here. One of the primary
MC> reasons the kernel keeps changing is that drivers keep being added to
MC> it. Developers won't write drivers for it because they are afraid of
MC> other drivers being added to it? I don't understand. The kernel should
MC> not change enough from version to version to require complete rewrites
MC> of the drivers, I don't think.

The latest versioned modules should solve this problem: the kernel
stays the same and you load the drivers as external modules. And if
the kernel is changed the modules can remain the same (which was not
the case up till now).

More and more drivers and other parts of the kernel will be moved into
modules.
--
Peter Mutsaers | AT Computing bv, P.O. Box 1428,
p...@atcmp.nl | 6501 BK Nijmegen, The Netherlands
tel. work: +31 (0)80 527248 |
tel. home: +31 (0)3405 71093 | "Trust me, I know what I'm doing."

Joseph W. Vigneau

unread,
Jan 27, 1995, 2:57:58 AM1/27/95
to
In article <newcombe.12...@aa.csc.peachnet.edu>,
Dan Newcombe <newc...@aa.csc.peachnet.edu> wrote:

>Of these three, I would say that the loadable modules (much like driver
>loading in config.sys) would be the way for a manufacturer to go, and for most
>of us to go also!

Where can I find information about how to make loadable kernel modules?
--
jo...@wpi.edu WPI Computer Science '97 Linux!
<a href="http://www.wpi.edu/~joev"> Click Here! </a>

Jean Liddle

unread,
Jan 26, 1995, 10:07:22 PM1/26/95
to
>Perhaps a better idea that could suit both of use would be a stable directory
>with links to programs/source/etc... that are known to work on that system.

I believe the on-line distributions (slackware, etc.) already do much to
fulfill this need. I wouldn't want to burden the ftp-site administrators
with the task of trying to maintain such a tree, even if they were willing.

>To tell you the truth, I don't think the ADU is interested in Linux, but
>that's a different ball game. It sounds to me that what you think would be
>good is a
> /pub/Linux/1.0.x
> /pub/Linux/1.1.x
>structure or something like that (trying to avoid words like beta and
>unstable). I would also suggest a /pub/Linux/both tree or something where
>source that works just as well on either tree would be placed.
>
>For instance:
> /pub/Linux/1.0.x
>would have
> a) program binaries (login, getty, bash, etc...)
> b) libraries (libc, libm, libsvga, libX11) etc...
> c) kernel
>that would all be at a set/defined/agreed upon revision, so that all programs
>under this tree would be compiled or guarenteed to work with this setup.
>Sounds like you'd need some sort of committee or group, like the File Standard
>people to work on this one. They would figure out when to upgrade, what
>should be there, etc...
> /pub/Linux/1.1.x
>would have binaries, libraries, kernel, etc... that are linked or designed for
>whatever version that happen to be for. Anarchy at best :)
> /pub/Linux/both

I like the above suggestion, with a couple of modifications.

#1 - instead of /pub/Linux/both I suggest /pub/Linux/contrib, which would
only contain source ... no binaries.
#2 - binaries under /pub/Linux/binaries (as you noted, most binaries don't
depend on kernel versions at all)
#3 - both of the above organized in the same fashion (e.g. identical
directory structures, probably the same as under /pub/Linux today,
modula the kernel and kernel specific binaries noted above
#4 - only kernel, driver, libraries, and kernel specific sources and
binaries under /pub/Linux/1.1.x and /pub/Linux/1.0.x, possibly with
those subtrees broken out by contrib and binaries if that was felt to be
useful. This would include such things as DOSEMU and WINE.

I think for binaries, the LSM is sufficient for determining which
libraries are required.

The advantage of this approach is that
A) users know whether or not applications will compile for their kernel
by which tree they pull it out of.
B) the organization doesn't change all that much from the current one, other
than adding 1 level to the directory structure
C) It explicitly reflects kernel dependencies where they do arise
D) It makes the search for binaries/sources easier (you don't download one
expecting the other)

Disadvantages
A) We need a way to make it easy for site admins to know where to put
/Incoming stuff. Perhaps a new field in the LSM indicating which
tree the *.tar.gz file or whatever should go in. I dunno.

Comments, thoughts?

Jean.

CHaley

unread,
Jan 27, 1995, 7:51:11 PM1/27/95
to
On Thu, 26 Jan 1995, Dan Newcombe wrote:

> In article <3g897g$i...@hermes.fwi.uva.nl> a...@fwi.uva.nl (Arie Maaskant) writes:
> >>Hmm...I'm not sure I see the connection here. One of the primary
> >>reasons the kernel keeps changing is that drivers keep being added to
>
> >That's right, when a manufacturer adds a Linux-driver of his hardware
> >to the Linux community, it de facto should change the version number.
> >But the manufacurer could advertise, that it is compilable and/or
> >usable under the last stable Linux-version. That is the same as saying
> >that your driver works with DOS6.2 ( or whatever number ). The version
> >number of DOS doesn't change either, if you add a driver to it.
>
> That is where the modules need to be concentrated on. Right now there are
> three ways to implement a driver:
> 1) Have it be a part of the official kernel, which means a change to it is a
> patch level increase of the kernel.
> 2) Have it be a standalone file, where the person that needs it must get
> (ftp?) the file and patch it on their own. (And pray it works with the
> current version of the kernel.
> 3) Have a loadable module, which doesn't affect the kernel source at all.
>
> Of these three, I would say that the loadable modules (much like driver
> loading in config.sys) would be the way for a manufacturer to go, and for most
> of us to go also!
>
> -Dan

The crappy thing about a config.sys driver-type system is that, for
example, when you load the mouse driver out of config.sys, it takes up
otherwise useful memory. the advantage of a /dev directory is being able
to add on devices that only use memory when they are being utilized.
This is one of the features of UNIX which has made it the standard for
REAL computers for years.

__
/ /\
/ / \ CHaley
/ / /\ \
/ / /\ \ \ home:(505)293-8910 * cha...@m-net.arbornet.org
/ /_/__\ \ \ ch00...@arriba.nm.org * ch00...@pi.lanl.gov
/________\ \ \
\___________\/


Kazimir Kylheku

unread,
Jan 28, 1995, 2:49:37 PM1/28/95
to
In article <3g7u2j$5...@neptune.ethz.ch>,

Frank Richard Ammeter <famm...@iiic.ethz.ch> wrote:
>Of course Linux is not an alternative for DOS. The average DOS user does
>not need Linux, since he does word processing only. Very few people really
>need a good OS: this explains Mircosoft's success.

No, very few people _perceive_ the need for a good OS because they
have no idea of what one is. The average computer owner thinks that
computers inherently lack reliability and robustness. For this
perception, yo can thank the likes of Microsoft and Apple.

Bernd Bernie Meyer

unread,
Jan 28, 1995, 8:33:37 PM1/28/95
to
c2a...@ugrad.cs.ubc.ca (Kazimir Kylheku) writes:

>No, very few people _perceive_ the need for a good OS because they
>have no idea of what one is. The average computer owner thinks that
>computers inherently lack reliability and robustness. For this
>perception, yo can thank the likes of Microsoft and Apple.

A wonderful point. My girlfriend used to work at a hotel, and was
pretty satisfied with the computer they had there --- until I took all my
spare parts together and built her a Windows machine. That made her use
dirty language about the user interface of their hotel app, and praise Word
for Windows. Next, I told her not to use WfW for her diploma thesis, but
emTeX. Now she uses dirty language for all those flashy looking low quality
apps.

I wonder if she would have liked linux --- we split up before she had the
hardware for it, though :-(

Bernie

Bernd Bernie Meyer

unread,
Jan 28, 1995, 8:42:23 PM1/28/95
to
je...@flowbee.interaccess.com (Jean Liddle) writes:

>I like the above suggestion, with a couple of modifications.

>#1 - instead of /pub/Linux/both I suggest /pub/Linux/contrib, which would
> only contain source ... no binaries.

[...]
>Comments, thoughts?

I would rather like to see _original_ sources accompanied by, if
required, patches for the to compile with linux. This would make life a lot
easier for those non-linux users who try to use archie.

Of course, it would provide everybody with useful insight into the porting
traps out there, too.

Bernie

Hallvard Paulsen

unread,
Jan 31, 1995, 5:50:12 AM1/31/95
to
In article <3ge74h...@keats.ugrad.cs.ubc.ca>,
|>
|> No, very few people _perceive_ the need for a good OS because they
|> have no idea of what one is. The average computer owner thinks that
|> computers inherently lack reliability and robustness. For this
|> perception, yo can thank the likes of Microsoft and Apple.
|>

Indeed, do we need computers, and do we need cars. Maybe
we could use a T-ford or a horse? And so on...

The fact is that linux is here, it for most purposes it
does a better job than DOS and therefore most
people that use DOS today will benefit from a change
to linux.

Isn't it as simple as this?

Hallvard P.

Nathan Hand

unread,
Jan 31, 1995, 9:23:26 AM1/31/95
to
Hallvard Paulsen (hall...@imm.unit.no) wrote:

: The fact is that linux is here, it for most purposes it


: does a better job than DOS and therefore most
: people that use DOS today will benefit from a change
: to linux.

: Isn't it as simple as this?

No... you havent defined what "a better job" is.

If you mean reliable, faster, neater, powerful, cheaper then linux does
the better job.

If you mean popular, commercial, supported, widespread then dos/windows
does the better job.

I wouldnt foist linux onto a 14 year old who wants games. Nor onto a 40
year old who wants Word for typing up office letters. By all means tell
them that linux exists: but its not "as simple" as making everying just
"change to linux".

--
-------------------------------------+----------------------------
Nathan Hand - nat...@bin.anu.edu.au | h930...@student.anu.edu.au
"help, help, im being repressed" | perfection is achieved only
"ahhh, what you gonna do about it" | on the point of collapse

Neale Pickett

unread,
Feb 1, 1995, 3:34:27 PM2/1/95
to
This thread has been around for quite a while, and that extraneous
apostrophe has been there with it in each new message, so I'm following up
in the (vain) hope that we might eventually phase out hackers who can't
spell.

Oh, yeah. Linux is *not* an alternative for DOS, in the sense that an IBM
PC is not an alternative to an Atari 2600. I sure as hell wouldn't
recommend Linux for high-speed data acquisition, but there's no way I'd
tell someone to use DOS for a multi-user workstation. DOS is good for what
it was designed--direct interface to the hardware. Because PCs were so
slow for so long, users didn't want to put up with all the overhead of a
real OS. Now, finally, the huge momentum started by the PC is dying down
and people are starting to think seriously about using an Operating System
on their computer (OS/2, Windows NT, Linux, FreeBSD...) but there are still
applications for which DOS (or a variant of DOS) will always be used.
That's one of the neat things about the PC architecture, and that's why it
sold so well. You can use it for almost anything (imagine trying to set up
a Mac to monitor your businesses phone system, or environmental systems.
It simply wasn't made to do that).

Okay, so we're all going to use correct punctuation, right?
--
___________________________________
zep...@nmt.edu This is the .signature file!

"How do you get down here and make this be on top?"

Pekka J Taipale

unread,
Feb 2, 1995, 1:17:23 AM2/2/95
to
In article <3glh4u$v...@manuel.anu.edu.au> nat...@bin.anu.edu.au (Nathan Hand) writes:
>I wouldnt foist linux onto a 14 year old who wants games.

I would. Heaven forbid, as "Real Programmers Don't Write Pascal" says,
"Unix is a glorified video game".

And a fun (and useful) game at that.

--
Pekka....@ntc.nokia.com

Hallvard Paulsen

unread,
Feb 2, 1995, 2:22:44 AM2/2/95
to
In article <3glh4u$v...@manuel.anu.edu.au>, nat...@bin.anu.edu.au (Nathan Hand) writes:
|>
|> No... you havent defined what "a better job" is.
|>
|> If you mean reliable, faster, neater, powerful, cheaper then linux does
|> the better job.

You said it. The part about reliable is the most important.
I cannot imagine anybody wanting the less reliable solution
if they ever had a real choice. Also the fact that linux
is a multiuser/multitasking 32-bit, is important when
it comes to "usability". (On the other side, I *can*
see some very special applications, where a dos based
solution would be better, since there are ways in which
compleate controll of the HW can be achieved. Since DOS
only runs a single program, I would gess that in theory
this single program could be made to run faster than it
is possible under linux).

|>
|> If you mean popular, commercial, supported, widespread then dos/windows
|> does the better job.

I don't think DOS is popular. Most users are not given any
choice, and do not know what else there is on the marked.

Hallvard P.

jpsb

unread,
Feb 2, 1995, 2:20:32 PM2/2/95
to
This seems like a good place to ask all the DOS guru's
out there for HELP!

>Linux is a 32-bit, multiuser, multitasking, networking OS.
>What in the world is the average DOS user going to do with such a thing?
>To the DOS/windoze user, the OS is a program loader. period. It loads
>a word processor, or a game, or something of the sort, and the DOS user
>is happy.

My Company is tring to port a Unix base software system to windows 3.1
Hasn't been easy but we are getting there. The other day we hit a real
problem, the application will launch an another binary and wait for
the process to complete before continuing execution. "system(whatever);
Now NOBODY thought this was a problem, pretty standard stuff. But no,
on the Windows side it lanuches the binary and then keeps on trucking.
Acting like a "system(whatever&);". Windows (VC++) doesn't wait for the
binary to complete. This is messing us all up. Any idea's on how to get
Windows to wait for the system call
to complete?
jim shirreffs
thanks in advance

Francis J. Bruening

unread,
Feb 2, 1995, 5:22:43 PM2/2/95
to
>>>>> "Neale" == Neale Pickett <zep...@nmt.edu> writes:
In article <1995Feb1.2...@nmt.edu> zep...@nmt.edu (Neale Pickett) writes:

[snip...]

Neale> and that's why it sold so well. You can use it for almost
Neale> anything (imagine trying to set up a Mac to monitor your
Neale> businesses phone system, or environmental systems. It
Neale> simply wasn't made to do that).


Not quite true. I'm not a Mac fan, (I run Linux myself) but I recently
saw a home grown robotics system at a lumber mill in Roseburg Oregon.
(Sunstuds Lumber - for the curious)

One of their operators had set up a series of laser sensors along
the milling line and tied the sensors along with the automated
cutting machine (who knows what they called it) in an old Mac SE
up in a control room.

This system would take the dimensional information from the laser
measuring system, and compute the most efficient piece of lumber
to make out of that particular piece of rough cut. It then
would set the cutter up and slice and dice away. Total time
per board, about 3-4 seconds...

I was surprised to see something like this here, and I spoke
to the guy who actually wrote the system. Nice guy, who has
no background in CS, but considered himself a "hobbiest".

The system has been up for over a year, and has been running
flawlessly the last 10 months or so.

The code was C, real ugly C, but it worked for them and worked
well. (and besides, the guy who wrote it didn't have 4+ years
of schooling on how to right "proper" code... I wish I could get
my programmers to write code which was "correct" not just pretty ;-))

Just another data point.

--
______________________________________________________________________________
Francis J. Bruening Francis.J...@TEK.COM
Tektronix Inc. T&M Central Software
PO Box 500 M/S:39-704 Voice: (503) 627-1023
Beaverton OR. 97076-4600 Fax: (503) 627-5587
______________________________________________________________________________

Joe Sloan

unread,
Feb 2, 1995, 6:14:02 PM2/2/95
to
In article <PTAIPALE.95Feb2081723@pc139_92.trs.ntc.nokia.com>,

Yes, quite - it all depends in how one looks at it... I understand the
point that the original poster was making, but I must point out that
networked linux doom is one of the coolest games around...

--
j...@wintermute.ucr.edu / You can't figure out how to
A linux machine! because a 486 / ACCELERATE your Windoze NT box?
is a terrible thing to waste! / -9.8 m/s^2 works quite well!!!

Joe Sloan

unread,
Feb 2, 1995, 6:17:35 PM2/2/95
to
In article <3grba0$f...@Starbase.NeoSoft.COM>,

jpsb <jp...@Starbase.NeoSoft.COM> wrote:
>This seems like a good place to ask all the DOS guru's
>out there for HELP!
...

>on the Windows side it lanuches the binary and then keeps on trucking.
>Acting like a "system(whatever&);". Windows (VC++) doesn't wait for the
>binary to complete. This is messing us all up. Any idea's on how to get
>Windows to wait for the system call
>to complete?

Sorry, guy, you're asking the wrong newsgroup -
No offense, but windoze can go take a hike for all I care

jjs

Juan Molinari

unread,
Feb 2, 1995, 9:19:12 PM2/2/95
to
jp...@Starbase.NeoSoft.COM (jpsb) writes:
>My Company is tring to port a Unix base software system to windows 3.1
>Hasn't been easy but we are getting there. The other day we hit a real
>problem, the application will launch an another binary and wait for
>the process to complete before continuing execution. "system(whatever);
>Now NOBODY thought this was a problem, pretty standard stuff. But no,
>on the Windows side it lanuches the binary and then keeps on trucking.
>Acting like a "system(whatever&);". Windows (VC++) doesn't wait for the
>binary to complete. This is messing us all up. Any idea's on how to get
>Windows to wait for the system call
>to complete?

Well, i'm reluctant to give you this fix, as you've already
demonstrated that VC++ for Windows (which I've never used
or heard of) doesn't operate under the same rules as C/UNIX.

Instead of using the system() call, use fork(), get the pid
for the new process (ha!) and then issue a wait().

int pid;
pid = fork();
if (pid == 0) { /* child */
system("bongo");
exit(0);
}
wait(pid); /* parent */
...


I'm about 95% sure this would work as described on a UNIX machine;
it would be a fun experiment under windows......

--
____ Juan G. Molinari (716) 424-8613
\HI/ ju...@clark.net Rochester, NY
\/ "Let's sacrifice him to our gods! ... Well, why not?
we used to do it all the time in the 30's..." -- Abraham Simpson

jpsb

unread,
Feb 4, 1995, 1:01:48 PM2/4/95
to
>>on the Windows side it lanuches the binary and then keeps on trucking.
>>Acting like a "system(whatever&);". Windows (VC++) doesn't wait for the
>>binary to complete. This is messing us all up. Any idea's on how to get
>>Windows to wait for the system call
>>to complete?
>
>Sorry, guy, you're asking the wrong newsgroup -
>No offense, but windoze can go take a hike for all I care

None taken, I'm a Unix advocate who has been dragged kicking and screaming
into the Windoze world. I posted here because i like linux and linux
users (i hate the backspace thou) just though maybe someone might have
an idea. I think i need to build a DLL, God help me I don't even know
how to compile on Windows, any body got an example of how to build a DLL?
This was so easy under Unix, why does the world turn to shit when you boot
dos? Does this mead i'll have to learn how to use dosemu? Well i've been
wanting to do that anyway. to all the windows fans out there, this is a
examble, a real world examble of why windows sucks. Maybe windows 95 or
NT will act like a multi-tasking OS shouuld. Maybe i'm just ignoant of
how windows works, but the Unix side of the house got pulled in because
the Dos side of the house couldn't.

jim shirreffs

Reagan Blundell

unread,
Feb 4, 1995, 8:04:02 PM2/4/95
to
Arie Maaskant (a...@fwi.uva.nl) wrote:
: newc...@aa.csc.peachnet.edu (Dan Newcombe) writes:

: >In article <3g38pm$h...@bit.fwi.uva.nl> a...@fwi.uva.nl (Arie Maaskant) writes:


: No, I do mean, there was a double-spaced filesystem reader made by somebody
: in Germany, it was possible to mount a Double-spaced filesystem in read-
: only mode. One only needed kernel version 1.1.18 ( I think ) with
: enhancements ( loadable modules ). Overall it was OK, but certainly not
: installable for the "average DOS-user".

: > Sounds like a personal problem to me :)

: No , it isn't. I do not think, that the "average DOS-user" is capable
: of using "loadable modules".


Why not? Dont they do it all the time with their config.sys file? Its
almost the same thing as the loadable modules, with the difference that
you can only load them at boot-time, and you cant unload them.


--
Reagan Blundell | "I Hope you know that
no...@desire.apana.org.au | this will go down on your
APANA Brisbane Regional | Permanent Record"
Management Committee | -- Violent Femmes

Stacy Leggitt

unread,
Feb 7, 1995, 11:02:58 PM2/7/95
to
>The fact is that linux is here, it for most purposes it
>does a better job than DOS and therefore most
>people that use DOS today will benefit from a change
>to linux.
>

Okay, "speak to me like I'm 2 years old." :) Exactly what does it do
better than DOS, and how. For somebody that's only nominally into
programming.


Bruce Scott TOK

unread,
Feb 9, 1995, 3:54:01 PM2/9/95
to

In a word, networking.

--
Gruss,
Dr Bruce Scott The deadliest bullshit is
Max-Planck-Institut fuer Plasmaphysik odorless and transparent
b...@ipp-garching.mpg.de -- W Gibson

muza...@smixedsignal.com

unread,
Feb 9, 1995, 8:48:57 PM2/9/95
to

In article <3hdvd9...@slcbds.aug.ipp-garching.mpg.de>, > In article
<3h9fpi$9...@news.u.washington.edu>, sleg...@u.washington.edu (Stacy Leggitt)
writes:
> |> >The fact is that linux is here, it for most purposes it
> |> >does a better job than DOS and therefore most
> |> >people that use DOS today will benefit from a change
> |> >to linux.
> |> >
> |>
> |> Okay, "speak to me like I'm 2 years old." :) Exactly what does it do
> |> better than DOS, and how. For somebody that's only nominally into
> |> programming.
>
> In a word, networking.

hmm, does it connect to a Netware server ? What about to a WfW 3.11 ? Does
it to NetBIOS ? XNS ? Lantastic networks ? LanMan ?

Yes, Linux talks to another Linux box but what about those 100 M machines
running other protocols ?

Muzaffer

standard disclaimer


Nigel

unread,
Feb 9, 1995, 11:45:13 PM2/9/95
to
In article <NEWTNews.10619.7...@omer1.smixedsignal.com>,

<muza...@smixedsignal.com> wrote:
>
>it to NetBIOS ? XNS ? Lantastic networks ? LanMan ?
>
>Yes, Linux talks to another Linux box but what about those 100 M machines
>running other protocols ?
I'm not sure about NetWare connectivity. The kernel _does_ have support
for the IPX protocol, so I assume that some kind of compatibility is
there.

As for NetBIOS/LANman/WfWg, etc....

Samba will allow Linux to connect to all of these protocols as just
another machine running LANman (at least from the other machines' point of
view).

>
>Muzaffer
>
See above.


--
------------------------------------------------------------
| Kris Zoerhoff du...@freenet.grfn.org |
| http://www.grfn.org/~dust/ |
------------------------------------------------------------

Steve Fuller

unread,
Feb 10, 1995, 1:01:32 AM2/10/95
to
On Thu, 09 Feb 95 18:48:57 PDT, muza...@smixedsignal.com wrote:

>> In a word, networking.

>hmm, does it connect to a Netware server ? What about to a WfW 3.11 ? Does
>it to NetBIOS ? XNS ? Lantastic networks ? LanMan ?

>Yes, Linux talks to another Linux box but what about those 100 M machines
>running other protocols ?

WfW and Netware have TCP/IP available. That's a nice easy way to get them
to talk to the linux box. I expect Netware support from something like
UnixWare, not from a free product like Linux. Granted it would be nice to
have, but I'm willing to live and work within the current operating
parameters and deal with it as best I can.

You don't use XTs to run Pagemaker. You don't use a Linux box to talk to
something that can't speak TCP/IP.

--
Steve Fuller
sfu...@worf.infonet.net

Nathan Hand

unread,
Feb 10, 1995, 3:12:27 AM2/10/95
to
muza...@smixedsignal.com wrote:
: > In a word, networking.

: hmm, does it connect to a Netware server ?

Dunno.. probably

: What about to a WfW 3.11 ?

Yes

: Does
: it to NetBIOS ?

Yes

: XNS ? Lantastic networks ? LanMan ?

What?

Also NFS and Appletalk. Have yet to see non-unix boxes do this
without massive amounts of $$$.

: Yes, Linux talks to another Linux box but what about those 100 M machines
: running other protocols ?

Yes... are you bashing linux out of knowledge or out of fear? Ill
leave the decision as an exercise to the reader.

--
"Ive never been so insulted in my life" +-----------------------
"Well, its early yet" +-----------------+ nat...@bin.anu.edu.au
----------------------+ I read the news reguarly -- sad, isnt it

Lawrence Foard

unread,
Feb 10, 1995, 11:33:04 AM2/10/95
to
>hmm, does it connect to a Netware server ? What about to a WfW 3.11 ? Does
>it to NetBIOS ? XNS ? Lantastic networks ? LanMan ?
>
>Yes, Linux talks to another Linux box but what about those 100 M machines
>running other protocols ?

Reformat them and install Linux :)
If you really need DOS/Windows install Windows+Work group for windows on it
and samba on one of the Linux machines...
--
------ Call the skeptic hotline 1-900-555-5555 talk to your own personal .
\ / skeptic 24 hours/day. Just say no to victimless crimes. . .
\ / High quality Linux application development available. . . .
\/ Violence is a lousy substitute for sex and drugs. . . . .

Bernd Bernie Meyer

unread,
Feb 11, 1995, 8:06:29 PM2/11/95
to
sleg...@u.washington.edu (Stacy Leggitt) writes:

>Okay, "speak to me like I'm 2 years old." :) Exactly what does it do
>better than DOS, and how. For somebody that's only nominally into
>programming.


It allows you to go on working (or reading news) while you wait for your
printout/calculation/simulation/text processing/whatever to finish

It allows you to password-protect your files from all those 2 year olds who
listened when talked to like grown ups and now like to play xinvaders

It allows you to dial in from your notebook when you realize that you just
forgot to copy that one file and you are at the wrong coast

It allows you to run the answering machine software for your modem without
tying up your whole computer

It allows your husband/partner/sister/kids/whoever to use the power of your
Pentium while not occupying its keyboard and monitor

Need some more?

Bernie

Highlander C H B M

unread,
Feb 11, 1995, 9:09:19 PM2/11/95
to
Stacy Leggitt (sleg...@u.washington.edu) wrote:
: >The fact is that linux is here, it for most purposes it

Well ... about everything . For starters it's a real OS . Then it
has real multiprocessing ( i'm writting this in my Linux box connect to my
univ unix machine while formatting some floppies and compiling a new gizmo )
Then it's free ... apps are free , it has wide free suport . Then it uses
your machine while doz turns your state-of-the-art pentium box into a 8086
with 640Kb ram . It is enough ? Ok , a little more , Linux users only have to
hit the power switch two times a day , and we get up and when we go to sleep .
A linux box can run indefinetly without a reboot , wich is more than I can say
for doz/windoze . ( ps: check my sig )


--

+---------------------------------+-------------------------------------------+
| Carlos Hugo Baptista Morgado | Are you the proud owner of a ZXSpectrum ? |
| e-mail: l39...@alfa.ist.utl.pt | So why are you a proud DOZ luser ? |
| l39...@ci.ist.utl.pt | Get a REAL OS , *NIX is the way to go ! |
| PGP key on finger +-------------------------------------------+
| Instituto Superior Tecnico de Lisboa -- Universidade Tecnica de Lisboa |
| ( Technical Superior Institute of Lisbon - Portugal ) |
+-----------------------------------------------------------------------------+

Jon M. Taylor

unread,
Feb 12, 1995, 4:38:49 PM2/12/95
to
In article <3hjqkf$u...@alfa.ist.utl.pt>,

Highlander C H B M <l39...@ci.ist.utl.pt> wrote:
>Stacy Leggitt (sleg...@u.washington.edu) wrote:
>: >The fact is that linux is here, it for most purposes it
>: >does a better job than DOS and therefore most
>: >people that use DOS today will benefit from a change
>: >to linux.
>: >
>
>: Okay, "speak to me like I'm 2 years old." :) Exactly what does it do
>: better than DOS, and how. For somebody that's only nominally into
>: programming.
> Well ... about everything . For starters it's a real OS.

Some would argue that a "real" OS is anything that will let you
run a program on your computer. Not me, but....

>Then it
>has real multiprocessing

It has multiTASKING. Multiprocessing is the ability to use
multiple CPUs concurrently, which Linux does not (yet) have. Some folx
are working on a multithreaded, multiprocessing kernel (Viper), but I
haven't heard much about it for some time now.

>( i'm writting this in my Linux box connect to my
>univ unix machine while formatting some floppies and compiling a new gizmo )

This is multitasking.

>Then it's free

Yes.

> ... apps are free,

Most, but not all. There's no prohibition against selling
commercial software that runs on Linux, and in fact some examples of this
already exist (Worperfect, that clipper clone thingy, Maple when it's
released, and others).

> it has wide free suport.

AND there are a growing number of companies that specialize in
for-pay Linux support. This will be a boon to those who have to have
support they can count on and hold responsible - the support on the net is
very good, but it is not reliable enough to be used in certain business
environments.

>Then it uses
>your machine while doz turns your state-of-the-art pentium box into a 8086
>with 640Kb ram . It is enough ? Ok , a little more , Linux users only have to
>hit the power switch two times a day , and we get up and when we go to sleep .
>A linux box can run indefinetly without a reboot , wich is more than I can say
>for doz/windoze . ( ps: check my sig )

This is not an absolute, however I cannot remember the last day at
work wherein I didn't have Windoze crash at least once. The difference
with Linux is that you can push your system MUCH harder and not have to
worry that it'll die on you.


--
-----------------------------------------
Jon M. Taylor <tay...@gaia.ecs.csus.edu>
-----------------------------------------
-----BEGIN PGP PUBLIC KEY BLOCK-----

Highlander C H B M

unread,
Feb 12, 1995, 7:29:17 PM2/12/95
to
[snip]
: >Then it
: >has real multiprocessing

: It has multiTASKING. Multiprocessing is the ability to use
: multiple CPUs concurrently, which Linux does not (yet) have. Some folx
: are working on a multithreaded, multiprocessing kernel (Viper), but I
: haven't heard much about it for some time now.

You're right ! ( my mistake ) btw : as anybody heard of multiprocessing
for PC ? ( bet not :-( )
[snip]
: > ... apps are free,

: Most, but not all. There's no prohibition against selling
: commercial software that runs on Linux, and in fact some examples of this
: already exist (Worperfect, that clipper clone thingy, Maple when it's
: released, and others).

Long live GNU :-)
: > it has wide free suport.

: >Then it uses

: >your machine while doz turns your state-of-the-art pentium box into a 8086
: >with 640Kb ram . It is enough ? Ok , a little more , Linux users only have to
: >hit the power switch two times a day , and we get up and when we go to sleep .
: >A linux box can run indefinetly without a reboot , wich is more than I can say
: >for doz/windoze . ( ps: check my sig )

: This is not an absolute, however I cannot remember the last day at
: work wherein I didn't have Windoze crash at least once. The difference
: with Linux is that you can push your system MUCH harder and not have to
: worry that it'll die on you.

Only once ? I guess you're one of the few :-)

: --

: -----------------------------------------
: Jon M. Taylor <tay...@gaia.ecs.csus.edu>
: -----------------------------------------
: -----BEGIN PGP PUBLIC KEY BLOCK-----

--

Bruce D. Scott

unread,
Feb 13, 1995, 11:55:01 AM2/13/95
to
muza...@smixedsignal.com wrote:

: In article <3hdvd9...@slcbds.aug.ipp-garching.mpg.de>, > In article

Let them learn to speak TCP/IP. It is the international language.

Nathan Hand

unread,
Feb 13, 1995, 10:29:19 PM2/13/95
to
Highlander C H B M (l39...@ci.ist.utl.pt) wrote:
: : >Then it
: : >has real multiprocessing

: : It has multiTASKING. Multiprocessing is the ability to use
: : multiple CPUs concurrently, which Linux does not (yet) have. Some folx
: : are working on a multithreaded, multiprocessing kernel (Viper), but I
: : haven't heard much about it for some time now.

: You're right ! ( my mistake ) btw : as anybody heard of multiprocessing
: for PC ? ( bet not :-( )

I believe NT supports multiple processors. You can buy dual Pentiums
right now. NT runs your old Win16 apps so if you run >1 application
you will be multiprocessing.

Steve Fuller

unread,
Feb 13, 1995, 11:08:18 PM2/13/95
to
On 14 Feb 1995 03:29:19 GMT, Nathan Hand (nat...@bin.anu.edu.au) wrote:

>I believe NT supports multiple processors. You can buy dual Pentiums
>right now. NT runs your old Win16 apps so if you run >1 application
>you will be multiprocessing.

Running 0 applications on that machine it could be multiprocessing. The
number of applications does not make a difference.

--
Steve Fuller Forsan et haec olim meminisse iuvabit.
sfu...@worf.infonet.net --Virgil
INS Online Help Personality

CHaley

unread,
Feb 14, 1995, 10:10:50 AM2/14/95
to
On 14 Feb 1995, Nathan Hand wrote:
> Highlander C H B M (l39...@ci.ist.utl.pt) wrote:
> : : >Then it
> : : >has real multiprocessing
> : : It has multiTASKING. Multiprocessing is the ability to use
> : : multiple CPUs concurrently, which Linux does not (yet) have. Some folx
> : : are working on a multithreaded, multiprocessing kernel (Viper), but I
> : : haven't heard much about it for some time now.
> : You're right ! ( my mistake ) btw : as anybody heard of multiprocessing
> : for PC ? ( bet not :-( )
> I believe NT supports multiple processors. You can buy dual Pentiums
> right now. NT runs your old Win16 apps so if you run >1 application
> you will be multiprocessing.
false.
Myth #1
Multiprocessing IS the ability of the *OS* to use multiple processors
concurrently, however, the Dual-P is not considered a
multiprocessing-able machine because it's architecture makes the OS think
that there is only 1 chip when the motherboard utilizes both.
Myth #2
Windows NT DOES support dual-Ps, as does Linux, however, neither of these
would run on REAL multiprocessing-able machines (see myth #1).

:)

-CHaley

----==-- _ // CHaley (HAC)
---==---(_)__ __ ____ __ // cha...@unm.edu, ch00...@pi.lanl.gov
--==---/ / _ \/ // /\ \/ / // Linux, the OS of the future, available
-=====/_/_//_/\_,_/ /_/\_\ // today from anonymous FTP (sunsite.unc.edu)
A proud member of TeamLinux// ask me about it!

Mark A. Horton KA4YBR

unread,
Feb 14, 1995, 9:29:08 PM2/14/95
to
Stacy Leggitt (sleg...@u.washington.edu) wrote:
: >The fact is that linux is here, it for most purposes it

Two words.... true multitasking.
Two more..... true multiuser.

--
"InfoMagic Linux Developer's Resource - we support it!"
------------------------------------------------------------
Mark A. Horton ka4ybr m...@ka4ybr.atlanta.com
Systems and Network Performance Tuning m...@ka4ybr.atl.ga.us
+1.404.371.0291 : 33 45 31 N / 084 16 59 W m...@ka4ybr.com

Paul English

unread,
Feb 14, 1995, 11:51:42 PM2/14/95
to

> In article <3glh4u$v...@manuel.anu.edu.au> nat...@bin.anu.edu.au (Nathan Hand) writes:
> >I wouldnt foist linux onto a 14 year old who wants games.
>
> I would. Heaven forbid, as "Real Programmers Don't Write Pascal" says,
> "Unix is a glorified video game".
>
> And a fun (and useful) game at that.

And of course Unix looked like a great arcade game in Jurassic Park
("This is Unix, I know this!"). Now if only we can get SGI to port
their graphical interface over to Linux... ;-)

Since the Linux version of Doom was released who needs DOS? No more
fiddling around with config.sys etc

Cheers,

Paul.

--
,-_|\ | Paul English (p.en...@nepean.uws.edu.au) =-=-=-=-=-=-=-=-=-=
/ \ | Associate Lecturer | PO Box 10 | Ph: +61 47 36 0607
\_.--_/ <- | Department of Computing | Kingswood 2747 | Fax: +61 2 678 5570
v | UWS - Nepean | NSW, AUSTRALIA |
<a href=http://www.st.nepean.uws.edu.au/~paule>My WWW Home Page</a>

Mark Addinall

unread,
Feb 15, 1995, 1:53:30 AM2/15/95
to
> I believe NT supports multiple processors. You can buy dual Pentiums
> right now. NT runs your old Win16 apps so if you run >1 application
> you will be multiprocessing.
>
> --
> "Ive never been so insulted in my life" +-----------------------
> "Well, its early yet" +-----------------+ nat...@bin.anu.edu.au
> ----------------------+ I read the news reguarly -- sad, isnt it


SCO has had MPX support for a couple of years, COMPAQ multi-CPU machines
were running it back in '91.

+----------------------------+--------+---------------------------------------+
|Mark Addinall |Opinions|1. Never share a Foxhole with someone |
|Senior Software Engineer | unlike | braver than yourself. |
|AWA Defence Industries | ORIONS |2. Your Rifle was made by the lowest |
|PH 02 8877 111 | belong | bidder. |
|MOB 015 895 977 | solely |3. Teamwork is essential, it gives them|
|inet: madd...@awadi.com.au |to me| them someone else to shoot at ;-) |
+-----------------------------------------------------------------------------+


Donal K. Fellows

unread,
Feb 15, 1995, 6:50:41 AM2/15/95
to
In article <3hjqkf$u...@alfa.ist.utl.pt>,
Highlander C H B M <l39...@ci.ist.utl.pt> wrote (in his sig):

>| Carlos Hugo Baptista Morgado | Are you the proud owner of a ZXSpectrum ?

Yes! Lets have a big round of applause for the Z80; the little chip
that could!

Donal.
--
Donal K. Fellows, A.K.A. ``I'll get a life when I can find the FTP site...''
--
Dept. of Computer Science, | 6, Randall Place, Heaton,
University of Manchester | Bradford, BD9 4AE
U.K. Tel: ++44-161-275-6137 | U.K. Tel: ++44-1274-401017
fell...@cs.man.ac.uk (preferred) | do...@ugglan.demon.co.uk (if you must)
--
Please do not quote this .signature, it isn't worth it! :)

Russell Marks

unread,
Feb 17, 1995, 11:33:40 AM2/17/95
to
| : >The fact is that linux is here, it for most purposes it
| : >does a better job than DOS and therefore most
| : >people that use DOS today will benefit from a change
| : >to linux.
| : >
|
| : Okay, "speak to me like I'm 2 years old." :) Exactly what does it do
| : better than DOS, and how. For somebody that's only nominally into
| : programming.
|
| Two words.... true multitasking.
| Two more..... true multiuser.

In addition, how about...

Linux is a LOT more crashproof than DOS, or DOS/Windows for that
matter.

It's actually quite hard to crash a Linux system unless you have root
access, and not so very easy to do it then except indirectly.

Of course, all these points apply to Unix generally.

[from another post to this thread...]

| > >I wouldnt foist linux onto a 14 year old who wants games.

I learned Z80 m/c when I was a 14 year old who wanted games. I suppose
that's a little atypical, but I just thought I'd mention it. I really,
really wish I could have used a Unix box when I was 14. I would have
absolutely *loved* it.

| > I would. Heaven forbid, as "Real Programmers Don't Write Pascal" says,
| > "Unix is a glorified video game".
| >
| > And a fun (and useful) game at that.

Indeed. Nethack is peanuts compared to Unix. :-)

| Since the Linux version of Doom was released who needs DOS? [...]

People who want to play other decent PC games. Seriously. That's what
I use DOS for these days, as a 21 year old who wants games. :-) Now,
if there were a Linux port of Sensible Soccer, I might reconsider...

[and from yet another post to this thread :-)]

| >| Carlos Hugo Baptista Morgado | Are you the proud owner of a ZXSpectrum ?
|
| Yes! Lets have a big round of applause for the Z80; the little chip
| that could!

*cheer*

A speccy +3 running CP/M makes an interesting terminal for a Linux
box, actually... :-)

-Rus.

--
/ : russell marks :: mr...@gre.ac.uk :: speak softly and carry a +6 kitten : \
| GCS -d+ -p+ c++++ l++ u++ e+(*) m+@ s+/++ n--(---) h+(*) f+ !g w+ t+ r- y? |
\ ::: "His world is under anaesthetic - subdivided and synthetic" - Rush ::: /

Alan Cox

unread,
Feb 21, 1995, 1:06:45 PM2/21/95
to
In article <NEWTNews.10619.7...@omer1.smixedsignal.com> muza...@smixedsignal.com writes:
>> In a word, networking.
>hmm, does it connect to a Netware server ? What about to a WfW 3.11 ? Does
>it to NetBIOS ? XNS ? Lantastic networks ? LanMan ?
>
>Yes, Linux talks to another Linux box but what about those 100 M machines
>running other protocols ?

Well lets see NFS is an international standard. Linux talks that
WfW 3.11 SMB over IP is supported (including long file name for NT etc)
Appletalk is supported for both Phase 1 and Phase 2.
Lan manager is SMB - please read before you write
XNS is dead (IMHO)
Netware is a Novell protected secret so we are all waiting for Undocumented
Netware so we can write Novell client/server support. You can use Novell
in a DOS window of course.
Lantastic I don't have any docs on again.

Oh and all the available stuff is free, and with source code.

I'd say novell is the real issue and that regretably isn't made easy by
Novell themselves.


Alan
--
..-----------,,----------------------------,,----------------------------,,
// Alan Cox // iia...@www.linux.org.uk // GW4PTS@GB7SWN.#45.GBR.EU //
``----------'`--[Anti Kibozing Signature]-'`----------------------------''
One two three: Kibo, Lawyer, Refugee :: Green card, Compaq come read me...

Alan Cox

unread,
Feb 24, 1995, 7:26:52 AM2/24/95
to
In article <3hs8da...@myall.awadi.com.au> madd...@awadi.com.au writes:
>SCO has had MPX support for a couple of years, COMPAQ multi-CPU machines
>were running it back in '91.

And TOPS-10 was MP just a while before that, as was stuff like GCOS-3/TSS

Manfred Waßmann

unread,
Mar 5, 1995, 8:24:31 PM3/5/95
to
In article <D4I7w...@info.swan.ac.uk>,
Alan Cox <iia...@iifeak.swan.ac.uk> wrote:
... something I don't want to comment on.
I just had to laugh hard when I read this Subject line. It sounds to me like
"MS-DOS is NOT a real alternative for the master boot record of my hard disk".

Well, that's the only thing MS-DOS really is useful for. That you can run
loadlin.exe from a batch file i.e. starting linux, thus having a comfortable
way to provide the linux kernel with extra parameters.

Linux is the only reason I bought a PC. Because a computer is nothing without
software and software for the PC was either too expensive or too broken.

I would have sticked to my 8 year old Atari ST with 4MB RAM in a linear
adressable memory space, an OS that's worth the name includeing a GUI, and a
whole bunch of excellent and cheap shareware.

On my jobs I had to work with MS-DOS, even did some programming, so I knew
how much I would loose if I gave up my ST for a PC.

The only alternative was a Mac but that's $$$ (at least you get something for
your money ;-).

manolo
--
The name of the game is X not X-Windows \/ Manfred "manolo" Waßmann
because Windows is the spell of Microsoft /\ mail me @ dawn.oche.de

Dan Pop

unread,
Mar 6, 1995, 7:49:07 AM3/6/95
to
In <D4zvw...@dawn.oche.de> man...@dawn.oche.de writes:

>I just had to laugh hard when I read this Subject line. It sounds to me like
>"MS-DOS is NOT a real alternative for the master boot record of my hard disk".
>
>Well, that's the only thing MS-DOS really is useful for. That you can run
>loadlin.exe from a batch file i.e. starting linux, thus having a comfortable
>way to provide the linux kernel with extra parameters.

If you use LILO instead, you can have a Microsoft free boot disk.

Dan
--
Dan Pop
CERN, CN Division
Email: dan...@cernapo.cern.ch
Mail: CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland

Manfred Waßmann

unread,
Mar 6, 1995, 8:07:25 PM3/6/95
to
In article <PAULE.95F...@dodinas.st.nepean.uws.edu.au>,

Paul English <pa...@dodinas.st.nepean.uws.edu.au> wrote:
>
>Since the Linux version of Doom was released who needs DOS? No more
>fiddling around with config.sys etc
>
And on a german keyboard you can leave it with "y" - you have to use "z" with
the DOS version ;-)

Manfred Waßmann

unread,
Mar 6, 1995, 8:03:05 PM3/6/95
to
In article <danpop.794494147@rscernix>, Dan Pop <dan...@cernapo.cern.ch> wrote:
>In <D4zvw...@dawn.oche.de> man...@dawn.oche.de writes:
>
>>I just had to laugh hard when I read this Subject line. It sounds to me like
>>"MS-DOS is NOT a real alternative for the master boot record of my hard disk".
>>
>>Well, that's the only thing MS-DOS really is useful for. That you can run
>>loadlin.exe from a batch file i.e. starting linux, thus having a comfortable
>>way to provide the linux kernel with extra parameters.
>
>If you use LILO instead, you can have a Microsoft free boot disk.
>
At home I use LILO of course, but I didn't get my employee to remove all these
MS viruses from his hard disks yet ;-)

But DOS really is a comfortable utility to boot linux, though I think
the program could be much simpler without loosing. It is Microsoft of
course ;-)

Clara Mass

unread,
Mar 10, 1995, 12:25:55 PM3/10/95
to
Dan Pop (dan...@cernapo.cern.ch) wrote:
: In <D4zvw...@dawn.oche.de> man...@dawn.oche.de writes:

: >I just had to laugh hard when I read this Subject line. It sounds to me like
: >"MS-DOS is NOT a real alternative for the master boot record of my hard disk".
: >
: >Well, that's the only thing MS-DOS really is useful for. That you can run
: >loadlin.exe from a batch file i.e. starting linux, thus having a comfortable
: >way to provide the linux kernel with extra parameters.

: If you use LILO instead, you can have a Microsoft free boot disk.

I can honestly say I have been "clean" (ie - no msdos) for
over a year now.


: Dan

Jiann-Ming Su

unread,
Mar 10, 1995, 5:10:06 PM3/10/95
to
In article <D4I7w...@info.swan.ac.uk>,
Alan Cox <iia...@iifeak.swan.ac.uk> wrote:
>In article <3hs8da...@myall.awadi.com.au> madd...@awadi.com.au writes:
>>SCO has had MPX support for a couple of years, COMPAQ multi-CPU machines
>>were running it back in '91.
>

MS is the minority owner of SCO, yet their WWW site is running BSD.
Just a thought. . .

--
Jiann-Ming Su mailto:j...@microwave2.ph.msstate.edu
http://microwave1.ph.msstate.edu/~js1
For F1 and IndyCar Picutures: ftp://microwave1.ph.msstate.edu/pub/racing

Clara Mass

unread,
Mar 11, 1995, 9:49:02 AM3/11/95
to
Manfred Waßmann (man...@dawn.oche.de) wrote:

: In article <danpop.794494147@rscernix>, Dan Pop <dan...@cernapo.cern.ch> wrote:
: >In <D4zvw...@dawn.oche.de> man...@dawn.oche.de writes:
: >
: >>I just had to laugh hard when I read this Subject line. It sounds to me like
: >>"MS-DOS is NOT a real alternative for the master boot record of my hard disk".
: >>
: >>Well, that's the only thing MS-DOS really is useful for. That you can run
: >>loadlin.exe from a batch file i.e. starting linux, thus having a comfortable

Yeah.. On our HP boxes there is a CMOS "boot" command area (that
I have rarely had to use) called "IBL" (I think). Basicly, you
get in the IBL(?) prompt, and you can specify boot devices,
alternate devices/addresses..

... This is what my MSDOS should be. .. My "Unix boot interface"..

Dan Newcombe

unread,
Mar 14, 1995, 4:35:53 PM3/14/95
to
cm...@freenet2.scri.fsu.edu (Clara Mass) wrote:
>: >>Well, that's the only thing MS-DOS really is useful for. That you can run
>: >>loadlin.exe from a batch file i.e. starting linux, thus having a comfortable
> Yeah.. On our HP boxes there is a CMOS "boot" command area (that
> I have rarely had to use) called "IBL" (I think). Basicly, you
> get in the IBL(?) prompt, and you can specify boot devices,
> alternate devices/addresses..

Actually it's called the ISL. It is basically LILO :) It allows you to boot
different images from different devices. The cool thing about the HP boxes
(at least the big ones) is that you can attach a modem to the serial port, and
then dial in and communicate with the ISL without HPUX even being up an running.
If you then bring up HPUX from over this serial line, it becomes the console.

-Dan


--
Dan Newcombe newc...@aa.csc.peachnet.edu
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
"And the man in the mirror has sad eyes." -Marillion

Mark Addinall

unread,
Mar 14, 1995, 11:21:30 PM3/14/95
to
In article j...@freenet3.scri.fsu.edu, cm...@freenet2.scri.fsu.edu (Clara Mass) writes:
> Manfred Waßmann (man...@dawn.oche.de) wrote:
> : In article <danpop.794494147@rscernix>, Dan Pop <dan...@cernapo.cern.ch> wrote:
> : >In <D4zvw...@dawn.oche.de> man...@dawn.oche.de writes:
> : >
>
[snip, for you know who ]


> Yeah.. On our HP boxes there is a CMOS "boot" command area (that
> I have rarely had to use) called "IBL" (I think). Basicly, you
> get in the IBL(?) prompt, and you can specify boot devices,
> alternate devices/addresses..
>
> ... This is what my MSDOS should be. .. My "Unix boot interface"..
>


Anyone remember the Burroughs Pcs back in the early 80's? You hat to boot a
load then a base operating system then DOS over the lot. Used to take a lot
of floppy feeds to get the machine up :-)

pig...@walden.mo.net

unread,
Jun 16, 1995, 3:00:00 AM6/16/95
to
Mark A. Horton KA4YBR (m...@ka4ybr.com) wrote:

: Stacy Leggitt (sleg...@u.washington.edu) wrote:
: : >The fact is that linux is here, it for most purposes it
: : >does a better job than DOS and therefore most
: : >people that use DOS today will benefit from a change
: : >to linux.
: : >

Maybe this to add:
windows and dos have been running a 16/"32" bit platform.
Which is to say it crashes. Better-> it isnt crash tolerent.


Linux IS a 32-bit disk operating system/networking operating system
And yes it crashed on me one time. But I have been running it for
a little over a year. Xwindows has crashed several times due
to the mem hog that it is (I am running w/ only 8megs)
And also due to the fact that it is a bitch to config.
(well can be)

Windows(95) will crash. - and if they ever code it correctly (hah)
right for a real 32 bit platform it still will always stand a better
risk of crashing due to the graphical nature.

Linux/Unix/Dos | Xwindows/Windows
sorta...
Well you need to see it.. what can I say.


0 new messages