Linux vs Minix
gre...@concentric.net
Yea, Minix.
I read the book by Tabenbaum that gave a good overview
of the Minix OS. Now I understand that Linux was written
because Linus T. didn't much care for Minix and wanted
somthing better. Is that true??
OK, be that as it may, Does Linux bear any simularity to
Minix (other than they are both UNIX knock-offs)? Or
are we trying to compare bicycles and Jet planes?
Any good books that describe the inner workings of Linux
like (or better than) the Tanenbaum book does for Minix?
Thanks
====================================
Receipt of unsolicited e-mail will result in
legal action pursuant to applicable federal laws.
Translation: NO SPAM! or else. (2 Spammers down, and counting)
Specific single replies to this posting are OK.
====================================
Paul Hughett
gre...@concentric.net wrote:
: Yea, Minix.
: I read the book by Tabenbaum that gave a good overview
: of the Minix OS. Now I understand that Linux was written
: because Linus T. didn't much care for Minix and wanted
: somthing better. Is that true??
: OK, be that as it may, Does Linux bear any simularity to
: Minix (other than they are both UNIX knock-offs)? Or
: are we trying to compare bicycles and Jet planes?
As I understand it, Minix was written primarily to teach
operating systems, rather than as a useful system. It
will--I think--run on the 286 while Linux will not (yet,
anyway). Linux has evolved into a complete Un*x system
capabile of running a wide variety of Unix software
fast enough to be useful as a workstation. There is a
news group concerning Minix; post a note there and they
should be able to give a pretty good idea of what Minix
is good for.
: Any good books that describe the inner workings of Linux
: like (or better than) the Tanenbaum book does for Minix?
Yes, there is a book called Linux Kernel Internals, by (I
think) Beck and several other authors that describes the
details of the kernel. Then you can go look at the source
code itself, plus some documentation that comes with it.
I haven't read Tannebaum's book, so I can't compare them,
but--given the differing motivations--I would expect it to
be less terse than the Linux internals book.
Paul Hughett
Oleg Dulin
> OK, be that as it may, Does Linux bear any simularity to
> Minix (other than they are both UNIX knock-offs)? Or
> are we trying to compare bicycles and Jet planes?
Well, Linux is far more than Minix. I have Minix on my old laptop. Linux
in comparing to Minix is like Windows NT in comparing to Windows 3.1 .
Think of Linux as a full-blown UNIX system, that SCO and Sun compete
with. Sun is offering a trade-in replacement, while SCO sent out a
letter offering Linux users to save $650 by purchasing SCO UnixWare for
only $1400 after they give up their $50 Red Hat Linux with unlimited
user licenses :)
Visit http://www.linux.org/, http://www.linuxmall.com/ and you will find
that Linux is much more than just Minix, or even UNIX for that matter.
Linux is a desktop operating system. Linux is UNIX for masses, not for
classes.
Oleg
Toon Moene
gre...@concentric.net wrote:
> Yea, Minix.
That's Tanenbaum for you :-)
> I read the book by Tabenbaum that gave a good overview
> of the Minix OS. Now I understand that Linux was written
> because Linus T. didn't much care for Minix and wanted
> somthing better. Is that true??
Well, I don't know what Linus' experience was, but I used (with emphasis on
"use"d) Minix in the summer of '87 on a 512 Kbyte IBM PC XT clone.
Unfortunately, that was it. Bootstrap system, type in ls and all that. No C
compiler, so the sources themselves were not that useful.
Like having them on microfiche - been there, done that.
--
Toon Moene (mailto:to...@moene.indiv.nluug.nl)
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Phone: +31 346 214290; Fax: +31 346 214286
g77 Support: mailto:for...@gnu.org; NWP: http://www.knmi.nl/hirlam
r.e.b...@usa.net
In article <672a2c$r...@examiner.concentric.net>,
gre...@concentric.net wrote:
> Yea, Minix.
> I read the book by Tabenbaum that gave a good overview
> of the Minix OS. Now I understand that Linux was written
> because Linus T. didn't much care for Minix and wanted
> somthing better. Is that true??
Minix was originally written for the Intel 8086, which did not have it's
own MMU, and relied on segmented memory. A commercial version, Coherant,
was available for a short while but never really took off.
> OK, be that as it may, Does Linux bear any simularity to
> Minix (other than they are both UNIX knock-offs)? Or
> are we trying to compare bicycles and Jet planes?
Linus studied the Minux code and tutorials, and proceded to write his own
version almost from the "bottom up". He started with the 32bit linear
addressing - which meant it wouldn't run on the old 286 machines, but
could handle ports of generic UNIX code.
Linus also added Memory Management, demand paged virtual memory, and
several other features. Over time, contributors added even better L1 and
L2 memory management, better disk caching, more reliable e2fs file
systems, and modules (drivers that could be loaded without recompiling
the kernel.
> Any good books that describe the inner workings of Linux
> like (or better than) the Tanenbaum book does for Minix?
There are several good books, including the Linux Documentation Project,
DR Linux, and the Linux Unleashed books. The Slackware releases often
include documentation of the kernel. The kernal source is also
interesting but a bit ugly and hard to read (the BSDI code is cleaner and
better documented).
> Thanks
>
> ====================================
> Receipt of unsolicited e-mail will result in
> legal action pursuant to applicable federal laws.
> Translation: NO SPAM! or else. (2 Spammers down, and counting)
> Specific single replies to this posting are OK.
> ====================================
I charge $25 for processing, storage, and handling of unsolicited e-mail.
If I'm interested, I won't charge you.
Rex Ballard
http://www.access.digex.net/~rballard
-------------------==== Posted via Deja News ====-----------------------
http://www.dejanews.com/ Search, Read, Post to Usenet
Nathan Hand
Toon Moene <to...@moene.indiv.nluug.nl> writes:
> Well, I don't know what Linus' experience was, but I used (with emphasis on
> "use"d) Minix in the summer of '87 on a 512 Kbyte IBM PC XT clone.
>
> Unfortunately, that was it. Bootstrap system, type in ls and all that. No
> C compiler, so the sources themselves were not that useful.
Errr, you must be missing a diskette. I fully remember using a
similar system with Minix, and being able to modify the kernel
source, recompile the entire kernel *on* *floppy* (no harddisk
in those days) and reboot to see my changes.
This was, of course, the original purpose for Minix.
--
The sticker on the side of the box said "Supported Platforms: Windows 95,
Windows NT 4.0, or better", so clearly Linux was a supported platform.
Russ Price
: Minix was originally written for the Intel 8086, which did not have it's
: own MMU, and relied on segmented memory. A commercial version, Coherant,
: was available for a short while but never really took off.
Actually, Coherent was around long before Minix - it was a product
of Mark Williams Company (is it still around?). I remember using
their "Let's C" compiler for MS-DOS way back when. It was OK,
as long as you only needed to compile small-memory-model (64k instruction/
64k data) programs...
I ran Minix for a little while, but I quickly found out that it just
wasn't ready for prime time...
: Linus studied the Minux code and tutorials, and proceded to write his own
: version almost from the "bottom up". He started with the 32bit linear
: addressing - which meant it wouldn't run on the old 286 machines, but
: could handle ports of generic UNIX code.
: Linus also added Memory Management, demand paged virtual memory, and
: several other features. Over time, contributors added even better L1 and
: L2 memory management, better disk caching, more reliable e2fs file
: systems, and modules (drivers that could be loaded without recompiling
: the kernel.
Isn't it amazing what a bunch of hobbyists can do when they put their
minds to it? I remember spending hours downloading the floppy images
for SLS way back in the 0.9x days -- via a long distance connection <ouch>,
since I had no 'Net access... Back then, my main use for Linux
was to run nethack :-), since it was far too bloated to run in
MS-DOS (who says that only commercial software is bloated?).
Now, I spend so much time surfing the Web and reading netnews, that I don't
have much time for games! <grin>
--
Russ Price - rjpri...@ais.net - "Doesn't 'Microsoft' mean 'small and limp?'"
(ANTI-SPAM ALERT: Remove the animal from my email address to reply to me.)
"It was chosen in our tradition of selecting names that make marketeers wince."
-- Lucent's Plan 9 FAQ, about Plan 9's name
Buddy, could you spare me a CPU? - http://www.distributed.net/rc5/
Toon Moene
Nathan Hand <natha...@anu.edu.au> wrote:
> Toon Moene <to...@moene.indiv.nluug.nl> writes:
> > Well, I don't know what Linus' experience was, but I used (with emphasis
on
> > "use"d) Minix in the summer of '87 on a 512 Kbyte IBM PC XT clone.
> > Unfortunately, that was it. Bootstrap system, type in ls and all that.
No
> > C compiler, so the sources themselves were not that useful.
> Errr, you must be missing a diskette. I fully remember using a
> similar system with Minix, and being able to modify the kernel
> source, recompile the entire kernel *on* *floppy* (no harddisk
> in those days) and reboot to see my changes.
It could be we were done in by the 512 Kbyte - I vaguely remember that we
couldn't do several things that were supposed to work "Out of the Box".
Unfortunately, the IBM PC XT wasn't our own system, so we didn't have
extended time to figure out what was wrong and whether there was a method to
get everything running in 512 K.
Besides, my real job was to improve the NOS/BE operating system we got
(source code included) at work. Now there was an OS: 250 interactive users
in 192 Kwords (of 60 bits), ie. roughly 1.5 Mbyte. But then, *that* OS was
written by Seymour Cray :-)
Timothy Murphy
Russ Price <rjcow...@ais.net> writes:
>I ran Minix for a little while, but I quickly found out that it just
>wasn't ready for prime time...
What exactly do you mean?
I ran Linux for some time (since the first version came out in fact)
and found it a fine system on 8086's and 80286's.
Far better than the then current version of DOS,
and just as stable.
In fact the unofficial 386 version was also very good --
more reliable than the first versions of Linux, as I recall.
Tanenbaum's error was to reject the 386.
(This was also Linus' opportunity.)
If Tanenbaum had gone over to the 386,
we would all be running Minix today.
He had a crazy argument that there so many 8086's around
that there would always be a need for an 8086 OS.
[I've always suspected that in truth
he just got sick of the hassle of developping an OS.]
--
Timothy Murphy
e-mail: t...@maths.tcd.ie
tel: +353-1-2842366
s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland
James Youngman
>>>>> "Timothy" == Timothy Murphy <t...@maths.tcd.ie> writes:
Timothy> Tanenbaum's error was to reject the 386.
Timothy> (This was also Linus' opportunity.)
Timothy> If Tanenbaum had gone over to the 386,
Timothy> we would all be running Minix today.
IMHO not. Most people concede that the GPL was a major factor in the
success of Linux. Minix was not under nearly so free a licence.
Timothy Murphy
James Youngman <JYou...@vggas.com> writes:
> Timothy> Tanenbaum's error was to reject the 386.
> Timothy> (This was also Linus' opportunity.)
> Timothy> If Tanenbaum had gone over to the 386,
> Timothy> we would all be running Minix today.
>IMHO not. Most people concede that the GPL was a major factor in the
>success of Linux. Minix was not under nearly so free a licence.
I don't think the licence makes the slightest difference.
The vast majority of Linux users are basically hobbyists.
The small number of semi-professionals and amateur lawyers
who go on and on about licences
are a tiny minority (best avoided, in my view).
If Tanenbaum had brought out a 386 version of Minix,
I'm sure Linus would never have had the motivation
to develop a rival OS.
r.e.b...@usa.net
In article <67bihh$gon$1...@turing.maths.tcd.ie>,
t...@maths.tcd.ie (Timothy Murphy) wrote:
> James Youngman <JYou...@vggas.com> writes:
> > Timothy> Tanenbaum's error was to reject the 386.
> > Timothy> (This was also Linus' opportunity.)
> > Timothy> If Tanenbaum had gone over to the 386,
> > Timothy> we would all be running Minix today.
> >IMHO not. Most people concede that the GPL was a major factor in the
> >success of Linux. Minix was not under nearly so free a licence.
Very true. If AT&T hadn't "given" UNIX to the colleges back in 1977-79,
UNIX would have been nothing and something like a cross between Smalltalk
and Forth would have evolved into the dominant operating system. R&D
people are notorious for taking the path of least resistance. It's very
easy to get a piece of unrestricted GPL software into a "Research Box".
It's a nightmere in corporate beaurocracy to try and get a piece of
"unblessed" software it the office.
Back in 1982-3, getting a PC into the office was a trick, even if you
bought it yourself. Getting it plugged into the Mainframe was like
asking the Pope to spend the night with Madonna. If you proposed it,
they'd give you this blank stare, ask for your supervisor's name, and ask
if your supervisor knew you were Psychotic. This was for an IBM PC-1
(8088) running MS-DOS.
> I don't think the licence makes the slightest difference.
> The vast majority of Linux users are basically hobbyists.
> The small number of semi-professionals and amateur lawyers
> who go on and on about licences
> are a tiny minority (best avoided, in my view).
Very wrong! There have been over a dozen versions of UNIX for the Intel
machine including Xenix, Venix, PC-UX, AIX, SunOS/386, Solaris,
Interactive Unix, Univel, SCO, SysVR4, BSD, FreeBSD, BSDi, Minix, and of
course 7 distributions of Linux. The easiest to aquire was Linux. The
easiest to get past the corporate legal system was Linux. The easiest to
get upgrades for was Linux. The easiest to give to friends was Linux.
The easiest to "Try Out" was Linux. The easiest to distribute was Linux.
The easiest to get funding for was Linux.
I've paid more in Cab Fare than I have paid for most of my copies of
Linux. Because it was so cheap, lot's of people were willing to "tinker
with it". As a result, thousands of applications emerged, all under the
same GPL. As the distribution grew, and grew, and grew ..., it began to
excel. It's a challenge to get the full suite of applications for other
versions of UNIX than it is to get what comes "standard" with Linux.
Now, commercial vendors are adding or switching to gclib to be compatible
with Linux binaries.
The GNU compilers and tools are now consdered "standard" by most
experienced UNIX people. Unix people often request Linux on their
workstation because they want the tools they won't get on a UNIX server.
> If Tanenbaum had brought out a 386 version of Minix,
> I'm sure Linus would never have had the motivation
> to develop a rival OS.
It's hard to say. Mark Williams came up with a 386 version of Coherant,
for only $99. Linux still dominated, offering Xfree86, X11R6, SAMBA, IPX,
Java, and Corba not to mention Apache and 5 different web browsers. This
was largely due to the fact that Linux was protected by GPL. For many,
their contribution was "pay back", for others, technical support, and for
others PR work was their "pay back". You can't buy customer "good will".
You certianly can't buy that kind of loyalty. How many NT 4.0 developers
are publishing applications, with source code, and expecting no payback.
> Timothy Murphy
> s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland
Rex Ballard
Linus Torvalds
In article <67bihh$gon$1...@turing.maths.tcd.ie>,
Timothy Murphy <t...@maths.tcd.ie> wrote:
>James Youngman <JYou...@vggas.com> writes:
>
>>IMHO not. Most people concede that the GPL was a major factor in the
>>success of Linux. Minix was not under nearly so free a licence.
>
>The vast majority of Linux users are basically hobbyists.
>The small number of semi-professionals and amateur lawyers
>who go on and on about licences
>are a tiny minority (best avoided, in my view).
>
>I'm sure Linus would never have had the motivation
>to develop a rival OS.
Actually, there _was_ a 386-version of Minix, mostly written by Bruce
Evans. The Minix/386 project was severely hampered by the restrictions
of the Minix copyright - it wasn't very easy to get a system up and
running with all the tools. Ast didn't want to include the i386 code in
the standard distribution because while they made the system actually
useful they also made it more complex.
The i386 patches didn't actually use paging etc at all, they still used
the original 286 segment approach but they expanded it to 32 bits. But
that was still what allowed me to run gcc on top of minix. All hail
Bruce Evans. (Bruce was also a great sounding board for my very first
testprograms that were later to become Linux).
People have later made "real" 386 versions of minix with more advanced
memory subsystems available, but while they were better than not having
it, they really never were very good. The basic minix design was so
fundamentally flawed by a really bad microkernel implementation that it
never really got off the ground.
Linus
Russ Price
Timothy Murphy <t...@maths.tcd.ie> wrote:
:>I ran Minix for a little while, but I quickly found out that it just
:>wasn't ready for prime time...
: What exactly do you mean?
[snip]
: Tanenbaum's error was to reject the 386.
: (This was also Linus' opportunity.)
You answered my question. :-) I was also rather disappointed in Minix's lack
of support for things like gcc, modems, network cards, etc.
Also, Minix's restrictive licensing made it problematic at best when it came
to distributing enhancements... putting Linux under GPL, and using an
open development model, was a big win, IMHO.
(As an aside, a BSD-style license wouldn't have been a bad way to go, either.)
Funny how the hot '386 of ten years ago has become so obsolete today... yet,
with Linux, it can still be put to good use.
Timothy Murphy
r.e.b...@usa.net writes:
>> If Tanenbaum had brought out a 386 version of Minix,
>> I'm sure Linus would never have had the motivation
>> to develop a rival OS.
>It's hard to say. Mark Williams came up with a 386 version of Coherant,
>for only $99. Linux still dominated, offering Xfree86, X11R6, SAMBA, IPX,
>Java, and Corba not to mention Apache and 5 different web browsers. This
>was largely due to the fact that Linux was protected by GPL.
This is manifest nonsense, since FreeBSD has all these programs
without the "benefit" of GPL.
If there is one thing worse than lawyers, it is amateur lawyers.
All this talk about licences is utter BS.
The success of Linux has absolutely nothing to do with GPL.
I doubt if 1% of Linux users care in the slightest
about the licence it runs under.
In fact I doubt if 1% have even read the GPL licence.
--
Timothy Murphy
e-mail: t...@maths.tcd.ie
tel: +353-1-2842366
Timothy Murphy
torv...@transmeta.com (Linus Torvalds) writes:
>>If Tanenbaum had brought out a 386 version of Minix,
>>I'm sure Linus would never have had the motivation
>>to develop a rival OS.
>Actually, there _was_ a 386-version of Minix, mostly written by Bruce
>Evans.
I know this -- I actually contributed a tiny bit to it
(a floating point library).
My point was that Tanenbaum, far from giving it his blessing,
more or less obstructed its development.
>The Minix/386 project was severely hampered by the restrictions
>of the Minix copyright - it wasn't very easy to get a system up and
>running with all the tools. Ast didn't want to include the i386 code in
>the standard distribution because while they made the system actually
>useful they also made it more complex.
It seemed to me at the time that Tanenbaum was becoming
less and less "open".
Having originally let almost anyone do anything,
he gradually found more and more reasons
to stop people developing Minix.
I came to the conclusion that he had actually lost interest in it,
or maybe found new interest in other things,
and was just trying to find a way to ease himself out.
>The basic minix design was so
>fundamentally flawed by a really bad microkernel implementation that it
>never really got off the ground.
Let's hope ast doesn't read this newsgroup!
James Youngman
>>>>> "Timothy" == Timothy Murphy <t...@maths.tcd.ie> writes:
Timothy> James Youngman <JYou...@vggas.com> writes:
Timothy> Tanenbaum's error was to reject the 386.
Timothy> (This was also Linus' opportunity.)
Timothy> If Tanenbaum had gone over to the 386,
Timothy> we would all be running Minix today.
>> IMHO not. Most people concede that the GPL was a major factor in the
>> success of Linux. Minix was not under nearly so free a licence.
Timothy> I don't think the licence makes the slightest difference.
Timothy> The vast majority of Linux users are basically hobbyists.
Timothy> The small number of semi-professionals and amateur lawyers
Timothy> who go on and on about licences
Timothy> are a tiny minority (best avoided, in my view).
No, the success of Linux is due to its reliability and feature set.
Linux would not have been in this position were it not for its army of
volunteers. This army of vounteers is there because of the GPL. The
"vast majority of Linux users are basically hobbyists" may indeed not
care a bit about the GPL, but it's because of the GPL that Linux is
good enough for them to use.
Timothy> If Tanenbaum had brought out a 386 version of Minix,
Timothy> I'm sure Linus would never have had the motivation
Timothy> to develop a rival OS.
*That* may be true.
James Youngman
>>>>> "Timothy" == Timothy Murphy <t...@maths.tcd.ie> writes:
Timothy> r.e.b...@usa.net writes:
>>> If Tanenbaum had brought out a 386 version of Minix,
>>> I'm sure Linus would never have had the motivation
>>> to develop a rival OS.
>> It's hard to say. Mark Williams came up with a 386 version of Coherant,
>> for only $99. Linux still dominated, offering Xfree86, X11R6, SAMBA, IPX,
>> Java, and Corba not to mention Apache and 5 different web browsers. This
>> was largely due to the fact that Linux was protected by GPL.
Timothy> This is manifest nonsense, since FreeBSD has all these programs
Timothy> without the "benefit" of GPL.
Nope, Samba is GPLed. Hence FreeBSD has it *with* the benefit of the
GPL.
Tim Smith
Toon Moene <to...@moene.indiv.nluug.nl> wrote:
>Unfortunately, that was it. Bootstrap system, type in ls and all that. No C
>compiler, so the sources themselves were not that useful.
Minix came with a C compiler (considering that the whole point of Minix was
to have an OS that came with source that students could modify and rebuild,
it would be rather silly to not have a compiler). It did not come with the
source to the C compiler, however.
--Tim Smith
Tim Smith
In article <882481764...@dejanews.com>, <r.e.b...@usa.net> wrote:
>Very wrong! There have been over a dozen versions of UNIX for the Intel
>machine including Xenix, Venix, PC-UX, AIX, SunOS/386, Solaris,
>Interactive Unix, Univel, SCO, SysVR4, BSD, FreeBSD, BSDi, Minix, and of
>course 7 distributions of Linux. The easiest to aquire was Linux. The
>easiest to get past the corporate legal system was Linux. The easiest to
>get upgrades for was Linux. The easiest to give to friends was Linux.
I don't see how Linux is easier to give to friends than FreeBSD or OpenBSD
or NetBSD, unless Linux has better CD-ROM duplication software.
--Tim Smith
Tim Smith
>for only $99. Linux still dominated, offering Xfree86, X11R6, SAMBA, IPX,
>Java, and Corba not to mention Apache and 5 different web browsers. This
Yet X, Java, and Apache are not GPL'ed. The thing that gave Linux the big
boost at the start was not GPL. It was the legal dispute between AT&T and
Berkeley that delayed the freeing of 4.4BSD, and put the free BSD projects
a year or so behind where they would have been otherwise.
--Tim Smith
Toon Moene
t...@halcyon.com (Tim Smith) wrote:
Yep, I think you're right (it has been a long time).
As I pointed out in a previous post, we couldn't run many of the commands (on
a 512 Kbyte system) and certainly not the C compiler.
Because, at that time, I had a day time job where we had the sources to both
the OS and the compilers, I quickly lost interest.
Not in the least because _that_ OS (NOS/BE) was written by a Real Programmer
(Seymour Cray).
Rajat Datta
On 20 Dec 1997 12:00:41 -0800, Tim Smith <t...@halcyon.com> wrote:
>Yet X, Java, and Apache are not GPL'ed. The thing that gave Linux the big
>boost at the start was not GPL. It was the legal dispute between AT&T and
>Berkeley that delayed the freeing of 4.4BSD, and put the free BSD projects
>a year or so behind where they would have been otherwise.
>
Also the attitude of the original BSD people. The current people in
the project are far more tolerant than the original folks of the
unwashed like us that wanted to run Microsoft OSs and BSD on the same
machine. I know a number of people, including me, who got turned off
because of extremely intolerant responses to our questions about DOS
multiboot, non-SCSI adapters, etc. Linux, at the time, was far behind,
but much more inclusive.
rajat
gre...@concentric.net
>fundamentally flawed by a really bad microkernel implementation that it
>never really got off the ground.
>
>
I never cease to be amazed by who I run into around here.
Although Ast hasn't yet made an appearance.
Anyway, having just read Ast's book I am intrieged by your
comment about Minix's "fundamentaly flawed micro kernel".
I myself could see several possible improvments. In your
opinion, what were Minix's flaws and how were they
addressed in Linux??
THanks.
Linus Torvalds
In article <67kpke$s...@examiner.concentric.net>,
<gre...@concentric.net> wrote:
>In <67cimd$ol5$1...@palladium.transmeta.com>, torv...@transmeta.com (Linus Torvalds) writes:
>>
>>People have later made "real" 386 versions of minix with more advanced
>>memory subsystems available, but while they were better than not having
>>it, they really never were very good. The basic minix design was so
>>fundamentally flawed by a really bad microkernel implementation that it
>>never really got off the ground.
>
>comment about Minix's "fundamentaly flawed micro kernel".
>I myself could see several possible improvments. In your
>opinion, what were Minix's flaws and how were they
>addressed in Linux??
The minix microkernel implementation is essentially single-threaded (and
fundamentally so, partly due to the message passing primitives - they
couldn't queue a message), which means that once you have multiple
kernel services depending on each others you end up having to be _very_
careful about deadlocks. What was supposed to be a simple system
becomes a nightmare of call graphs..
For example, this makes handling paging rather nasty. Imagine doing a
large "write()" call, and the systems needs more memory for buffers. So
it wants to write out a few pages to disk. Sadly, the filesystem server
is already busy doing the write, and due to its single-threaded nature
you're now dead-locked.
So to avoid all circular dependencies, at least the original minix had
very strict static rules (processes couldn't grow at run-time, for
example, and you obviously couldn't take page faults etc). These
limitations were still true in the i386-minix that I used, but at least
the 32-bit issues made the system much more pleasant in other ways.
I think some of the minix patches tried to work around some of these
problems, but basically the system was never very well designed. Last
time I brought this up, ast still considered multithreaded servers just
a "small performance issue", rather than a major fundamental thing.
As to how these kinds of problems are addressed in Linux, the answer is
simple: the above kind of problem is not an issue in a traditional
monolithic kernel. It's a fundamental flaw in microkernels - although
it can be (reasonable easily assuming a good design) handled by making
the servers re-entrant which is something that Minix never did as far as
I know.
In a monolithic system the kernel is quite naturally multi-threaded
without any added special logic (in fact, much special logic is needed
to _protect_ regions that aren't naturally re-entrant - which is a lot
better than the opposite approach because it forces people to be very
aware of when they aren't re-entrant).
The reason ast didn't want to make minix servers re-entrant was that he
felt that it wasn't all that important (total BS, imnsho) and that it
makes the system more complex (quite true). The fact that the lack of
re-entrancy made the system almost useless for real work seems to not
have been an issue.
To me, microkernel people are these shifty used car salesmen: they
promise great systems, but the end result always seems to have most of
the corners cut. They seem to offer one of "speed", "simplicity" and
"stability", but never a combination of all three. QNX seems to be the
best of the lot by far, and even QNX seems to be mainly a special-
purpose system.
Linus
Phil Fraering
torv...@transmeta.com (Linus Torvalds) writes:
>In a monolithic system the kernel is quite naturally multi-threaded
>without any added special logic (in fact, much special logic is needed
>to _protect_ regions that aren't naturally re-entrant - which is a lot
>better than the opposite approach because it forces people to be very
>aware of when they aren't re-entrant).
Just curious. Maybe I'm out of my depth here... but I've heard that
the new glibc libraries are meant to facilitate threaded and properly
re-entrant code... does any of this do anything to help the kernel?
Does the kernel incorporate much library code?
>The reason ast didn't want to make minix servers re-entrant was that he
>felt that it wasn't all that important (total BS, imnsho) and that it
>makes the system more complex (quite true). The fact that the lack of
>re-entrancy made the system almost useless for real work seems to not
>have been an issue.
>To me, microkernel people are these shifty used car salesmen: they
>promise great systems, but the end result always seems to have most of
>the corners cut. They seem to offer one of "speed", "simplicity" and
>"stability", but never a combination of all three. QNX seems to be the
>best of the lot by far, and even QNX seems to be mainly a special-
>purpose system.
I take it from what you're saying that QNX is re-entrant? I might
give it a try, when I get the floppy swapped out.
Phil
> Linus
--
Phil Fraering "You will cooperate with Microsoft, for the
p...@globalreach.net good of Microsoft and for your own survival."
/Will work for *tape*/ - Navindra Umanee
Ben Lindstrom
Tim Smith <t...@halcyon.com> wrote:
> In article <882481764...@dejanews.com>, <r.e.b...@usa.net> wrote:
>>machine including Xenix, Venix, PC-UX, AIX, SunOS/386, Solaris,
>>Interactive Unix, Univel, SCO, SysVR4, BSD, FreeBSD, BSDi, Minix, and of
>>course 7 distributions of Linux. The easiest to aquire was Linux. The
>>easiest to get past the corporate legal system was Linux. The easiest to
>>get upgrades for was Linux. The easiest to give to friends was Linux.
> I don't see how Linux is easier to give to friends than FreeBSD or OpenBSD
> or NetBSD, unless Linux has better CD-ROM duplication software.
Only thing easier is the fact that you don't find FreeBSD/OpenBSD at your
local Best Buy store where Linux/Redhat is at Best Buy, and 3 other
local computer stores. =)
Anyways...back to working on IPNAT under OpenBSD.=) Still have one little
bug.=) UDP/Diablo...WHY UDP?!?! Blech.
Navindra Umanee
At one point in cyberspacetime,
Ben Lindstrom <mou...@newton.pconline.com> wrote:
> Anyways...back to working on IPNAT under OpenBSD.=) Still have one little
> bug.=) UDP/Diablo...WHY UDP?!?! Blech.
UDP == Unreliable Data Protocol... So why not? Wouldn't it be better
and faster than TCP in this case?
What does this have to do with OpenBSD/Linux? Can you play Diablo on
either?
Navin
Kai Henningsen
ph...@lungold.globalreach.net (Phil Fraering) wrote on 22.12.97 in <ng7n76.5d.ln@localhost>:
> torv...@transmeta.com (Linus Torvalds) writes:
>
> >In a monolithic system the kernel is quite naturally multi-threaded
> >without any added special logic (in fact, much special logic is needed
> >to _protect_ regions that aren't naturally re-entrant - which is a lot
> >better than the opposite approach because it forces people to be very
> >aware of when they aren't re-entrant).
>
> Just curious. Maybe I'm out of my depth here... but I've heard that
> the new glibc libraries are meant to facilitate threaded and properly
> re-entrant code... does any of this do anything to help the kernel?
>
> Does the kernel incorporate much library code?
None at all. The libc is a pure user level beast; the kernel uses nothing
from it.
Kai
--
Internet: k...@khms.westfalen.de
Bang: major_backbone!khms.westfalen.de!kai
http://www.westfalen.de/private/khms/
Kai Henningsen
torv...@transmeta.com (Linus Torvalds) wrote on 22.12.97 in <67l60t$c0p$1...@palladium.transmeta.com>:
> In article <67kpke$s...@examiner.concentric.net>,
> <gre...@concentric.net> wrote:
> >In <67cimd$ol5$1...@palladium.transmeta.com>, torv...@transmeta.com (Linus
> >>Torvalds) writes:
> >>People have later made "real" 386 versions of minix with more advanced
> >>memory subsystems available, but while they were better than not having
> >>it, they really never were very good. The basic minix design was so
> >>fundamentally flawed by a really bad microkernel implementation that it
> >>never really got off the ground.
> >
> >Anyway, having just read Ast's book I am intrieged by your
> >comment about Minix's "fundamentaly flawed micro kernel".
> >I myself could see several possible improvments. In your
> >opinion, what were Minix's flaws and how were they
> >addressed in Linux??
>
> The minix microkernel implementation is essentially single-threaded (and
> fundamentally so, partly due to the message passing primitives - they
> couldn't queue a message), which means that once you have multiple
> kernel services depending on each others you end up having to be _very_
> careful about deadlocks. What was supposed to be a simple system
> becomes a nightmare of call graphs..
Oh yes. I still remember there was this awfully complicated and hard-to-
understand (and obviously even harder to get right - races and deadlocks
whenever something changed) code where you had to call one piece of the
kernel from another, and _couldn't_ use messages because they where
unqueuable.
Tanenbaum freely admitted that problem. However, he absolutely refused to
have message queues implemented, because it would "complicate" the system
"too much". Or at least that's how he explained it to me. (Funny, _I_'d
said that it would make the system far easier to understand.)
That was about the moment when I decided to give up on Minix. Obviously,
its very real flaws weren't ever going to be fixed.
I experimented a little with the very beginnings of writing my own, but
never got really far, mostly due to not investing enough time. I'm very
glad Linus did better at this :-)
At the time, I hoped to eventually use BSD. However, the BSD bandwagon,
far from taking off, seemed to be only interested in slowing down, not
only because of the ATT lawsuit. And then there was Linux.
Part of it is a matter of timing. Part of it, however, is a matter of
licenses - the Minix license, combined with ast's attitude, effectively
killed Minix as a serious alternative to Linux (Ast wouldn't make it a
real system, and nobody else was allowed to).
> The reason ast didn't want to make minix servers re-entrant was that he
> felt that it wasn't all that important (total BS, imnsho) and that it
> makes the system more complex (quite true). The fact that the lack of
> re-entrancy made the system almost useless for real work seems to not
> have been an issue.
Not only that, it _also_ lead to unnecessary complexity whenever you had
to do something that would ordinarily have been done via re-entrancy.
There were several corners in Minix that were _very_ ugly because of this;
I'm still of the opinion that this was quite a match for the complexity
introduced the other way.
Ben Lindstrom
Navindra Umanee <navi...@cs.mcgill.ca> wrote:
> At one point in cyberspacetime,
> Ben Lindstrom <mou...@newton.pconline.com> wrote:
>> Anyways...back to working on IPNAT under OpenBSD.=) Still have one little
>> bug.=) UDP/Diablo...WHY UDP?!?! Blech.
> UDP == Unreliable Data Protocol... So why not? Wouldn't it be better
> and faster than TCP in this case?
Ask Blizzard. UDP has very limited uses IMHO. And when used right it
can be a blessing. However, I'm stuck using UDP as a transport.
> What does this have to do with OpenBSD/Linux? Can you play Diablo on
> either?
Sorry, should have made it clear. =)
{Win95,Linux} --Internal Net-->OpenBSD 2.2--Modem Line-->ISP.
For a few reason.
1... To play with IPNAT and IP Filtering.
2... To protect Windows 95 and Linux from the outside world.
3... To "play" on the "Dark Side" (Or as what many Linuxer claim the BSD are)
We are off the topic now.=) What this has to do with Minux/Linux is beyound
me.=)
r.e.b...@usa.net
In article <67cnkg$e...@graves.maths.tcd.ie>,
t...@maths.tcd.ie (Timothy Murphy) wrote:
> r.e.b...@usa.net writes:
> >> If Tanenbaum had brought out a 386 version of Minix,
> >> I'm sure Linus would never have had the motivation
> >> to develop a rival OS.
>
> >for only $99. Linux still dominated, offering Xfree86, X11R6, SAMBA, IPX,
> >Java, and Corba not to mention Apache and 5 different web browsers. This
> >was largely due to the fact that Linux was protected by GPL.
> This is manifest nonsense, since FreeBSD has all these programs
> without the "benefit" of GPL.
As late as 1994, you could not get FreeBSD in a working compiled version.
There was some proprietary code that was still restricted by the AT&T
license. BSDi is $400 and more for an unlimited user license. For a Web
server, you must have an unlimited user license, or know the price. Even
at $1/user, the bill can get huge quickly.
> If there is one thing worse than lawyers, it is amateur lawyers.
> All this talk about licences is utter BS.
> The success of Linux has absolutely nothing to do with GPL.
Linux enjoys much of the support it gets because people who develop for
Linux are protected by GPL. They don't have to worry about writing an
application, or adding an enhancement, or fixing a bug, only to have "Big
Brother Bill" cripple their version with "system upgrades" and replace it
with the "Big Brother Bill" version. Thousands of people contributed to
BSD from 4.1 to 4.4, and many found their products showing up as Windows
applications, with no recourse.
> I doubt if 1% of Linux users care in the slightest
> about the licence it runs under.
> In fact I doubt if 1% have even read the GPL licence.
It might be a little larger. There are about 20 million Linux users, and
about 100,000 Linux contributors. The 20 million users think GPL is cool
because they can buy a copy and pass it around at work or school -
legally. They know enough about the license to know they can give it to
their buddies. The other 100,000 Linux contributors know that if someone
uses their source code to create a Windows implementation, they are
protected by the copyright and the dirivative work (Windows
implementation) is subject to the same copyright.
Finally, Linux got free coverage in usenet news groups, and exposure to
the national media, primarily because those who promoted it knew that
their promotional efforts were going to a worthy cause and wouldn't be
exploited by "Big Brother Bill". To this day, Linux and Apache dominate
the Internet server market. Linux workstation usage is increasing at a
rate of nearly 15%/month. The usage is still part-time (dual-boot), but
the percentage of time spent in Linux is increasing as applications
become available.
> Timothy Murphy
> e-mail: t...@maths.tcd.ie
> tel: +353-1-2842366
> s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland
Rex Ballard
r.e.b...@usa.net
In article <67vglb$nos$1...@bell.pconline.com>,
Ben Lindstrom <mou...@newton.pconline.com> wrote:
>
> Tim Smith <t...@halcyon.com> wrote:
> > In article <882481764...@dejanews.com>, <r.e.b...@usa.net> wrote:
> >>Very wrong! There have been over a dozen versions of UNIX for the Intel
> >>machine including Xenix, Venix, PC-UX, AIX, SunOS/386, Solaris,
> >>Interactive Unix, Univel, SCO, SysVR4, BSD, FreeBSD, BSDi, Minix, and of
> >>course 7 distributions of Linux. The easiest to aquire was Linux. The
> >>easiest to get past the corporate legal system was Linux. The easiest to
> >>get upgrades for was Linux. The easiest to give to friends was Linux.
>
> > I don't see how Linux is easier to give to friends than FreeBSD or OpenBSD
> > or NetBSD, unless Linux has better CD-ROM duplication software.
Linux is easier to duplicate (legally) because of the GPL. There are
several qualified restrictions on BSD.
> Only thing easier is the fact that you don't find FreeBSD/OpenBSD at your
> local Best Buy store where Linux/Redhat is at Best Buy, and 3 other
> local computer stores. =)
Also, you don't have over 21 book titles, each with at least one
complimentary copy of the FreeBSD Operating System.
You do have over 21 book titles, each with at least one (sometimes 3)
complimentary copy of the Linux Operating System. Linux didn't even make
the Computer Retailer "Top Seller List" because 80% of the Linux copies
sold were sold as BOOKS. Linux is simultaneously keeping a very low
profile, and selling copies by the millions.
In addition, a Linux Update 'freeze release' is published every month and
Red Hat, Slackware, Debian, and S.U.S.E. each publish update every 3-4
months. The legacy software is frequently given to friends or relatives
(who try it because it's free and quickly buy the upgrades).
It's ironic actually. I've probably spent nearly $200 on boxes and books
containing copies of Linux, not including the Applix package that came
with a complimentary copy of Linux. I still feel that I got a very good
deal. (I've given away several copies that were returned when, after a
"test drive", the borrowers decided to buy their own copy).
Flagship (WGS)was selling quantities of Red Hat Linux for $2/copy. I
thought about giving those out along with Christmas Cards.
Timothy Murphy
r.e.b...@usa.net writes:
>> The success of Linux has absolutely nothing to do with GPL.
>Linux enjoys much of the support it gets because people who develop for
>Linux are protected by GPL.
I don't believe it would have made the slightest difference
if Linus Torvalds had decided to follow the FreeBSD licence,
or for that matter if FreeBSD had decided to follow the GPL licence.
The development of TeX has been very similar in many ways
to that of Linux,
but the licence is completely different.
>They don't have to worry about writing an
>application, or adding an enhancement, or fixing a bug, only to have "Big
>Brother Bill" cripple their version with "system upgrades" and replace it
>with the "Big Brother Bill" version. Thousands of people contributed to
>BSD from 4.1 to 4.4, and many found their products showing up as Windows
>applications, with no recourse.
As you say, many people contributed to BSD
without the benefit of GPL.
I have met many of these people,
and cannot recall any of them ever complaining along the lines you suggest.
>To this day, Linux and Apache dominate the Internet server market.
This is certainly true of Apache, but is it true of Linux?
I'd like to think it was, but I very much doubt it.
--
Marc Slemko
In <883787826...@dejanews.com> r.e.b...@usa.net writes:
>In article <67cnkg$e...@graves.maths.tcd.ie>,
> t...@maths.tcd.ie (Timothy Murphy) wrote:
[...]
>> If there is one thing worse than lawyers, it is amateur lawyers.
>> All this talk about licences is utter BS.
>> The success of Linux has absolutely nothing to do with GPL.
>Linux enjoys much of the support it gets because people who develop for
>Linux are protected by GPL. They don't have to worry about writing an
>application, or adding an enhancement, or fixing a bug, only to have "Big
>Brother Bill" cripple their version with "system upgrades" and replace it
>with the "Big Brother Bill" version. Thousands of people contributed to
>BSD from 4.1 to 4.4, and many found their products showing up as Windows
>applications, with no recourse.
[...]
>Finally, Linux got free coverage in usenet news groups, and exposure to
>the national media, primarily because those who promoted it knew that
>their promotional efforts were going to a worthy cause and wouldn't be
>exploited by "Big Brother Bill". To this day, Linux and Apache dominate
>the Internet server market. Linux workstation usage is increasing at a
>rate of nearly 15%/month. The usage is still part-time (dual-boot), but
>the percentage of time spent in Linux is increasing as applications
>become available.
You say that Linux and Apache "dominate the Internet server market".
While I wouldn't necessarily agree with that, I fail to see how this
supports your contention that the GPL is the cause of greatness.
Apache is not GPLed. Your assertion that the GPL is the reason
for developer support of Linux isn't supported by anything you say.
r.e.b...@usa.net
In article <68ms1b$5mm$1...@scanner.worldgate.com>,
Marc Slemko <ma...@znep.com> wrote:
>
> In <883787826...@dejanews.com> r.e.b...@usa.net writes:
>
> >In article <67cnkg$e...@graves.maths.tcd.ie>,
> > t...@maths.tcd.ie (Timothy Murphy) wrote:
>
> [...]
>
> >> If there is one thing worse than lawyers, it is amateur lawyers.
> >> All this talk about licences is utter BS.
> >> The success of Linux has absolutely nothing to do with GPL.
It's a fundamental issue of "Path of Least Resistance". Linux is
successful BECAUSE of GPL. If Linus had tried to keep Linux to himself,
his key interests would have been addressed (memory management,
scheduling, caching). But there wouldn't have been a Linux Documentation
project. There wouldn't have been EMACS, the GNU clib, the hundreds of
commands, bash. The korn shell, csh, and sh were originally copyrighted
and restricted. Lex and Yacc were protected by Berkely.
AT&T would not have put the millions of staff-years worth of support that
UNIX, BSD, Minix, ... and Linux have enjoyed. It certainly wouldn't have
encouraged the porting of UNIX applications to these "non-unix"
platforms. In fact, AT&T system 3, which had no BSD utilities was so
"crippled" that corporations were choosing Berkely 4.2 instead. With
System V, AT&T traded certain kernel copyrights for the contributed BSD
utilities.
They used to joke about UNIX, calling it "that system created by college
kids". Of course, when those "college kids" graduated, they formed
companies like SUN and SGI.
> >Linux enjoys much of the support it gets because people who develop for
> >Linux are protected by GPL. They don't have to worry about writing an
> >application, or adding an enhancement, or fixing a bug, only to have "Big
> >Brother Bill" cripple their version with "system upgrades" and replace it
> >with the "Big Brother Bill" version. Thousands of people contributed to
> >BSD from 4.1 to 4.4, and many found their products showing up as Windows
> >applications, with no recourse.
Have you noticed that there isn't this huge repository of source code
available for Microsoft windows? In fact, the repository of high quality
third party shareware binaries for NT 4.0 is rather sparse. It seems
like the only folks doing real innovative development for Microsoft these
days is Microsoft (I know this isn't true, but it just seems like it).
> [...]
> >Finally, Linux got free coverage in usenet news groups, and exposure to
> >the national media, primarily because those who promoted it knew that
> >their promotional efforts were going to a worthy cause and wouldn't be
> >exploited by "Big Brother Bill". To this day, Linux and Apache dominate
> >the Internet server market. Linux workstation usage is increasing at a
> >rate of nearly 15%/month. The usage is still part-time (dual-boot), but
> >the percentage of time spent in Linux is increasing as applications
> >become available.
> You say that Linux and Apache "dominate the Internet server market".
According to the Netcraft survey (http://www.netcraft.co.uk/Survey),
Apache still commands 45% of the market. Linux and Solaris vie for
top spot depending on who is counting what (Linux has more cages but
Solaris has more pages and sites, Solaris also has 64 processor cages).
Many baby Linux sites grow up to be Solaris, SGI, AIX, or HP-UX machines.
Others just put Linux on Sparc, Mips, PPC, or Alpha machines.
> While I wouldn't necessarily agree with that, I fail to see how this
> supports your contention that the GPL is the cause of greatness.
> Apache is not GPLed. Your assertion that the GPL is the reason
> for developer support of Linux isn't supported by anything you say.
There are many variations on the FSF GPL, but fundamentally, the core
provisions - copyrighted software, availability of freely distributed
source code, and some assurance that the copyright holder isn't going to
use your contributions to create a Microsoft "Superproduct" provide a
structure that encourages contribution, cooperation, and user generated
promotion.
Consider that UNIX has been around for 20 years, and 99% of what you can
find on a commercial UNIX box can also be obtained via GPL on a Linux
distribution source code CD-ROM.
This doesn't mean that GPL guarantees "success", or that the a GPL product
will always win out over a commercial product. Netscape is a good example
of a company that competes effectively with GPL by actually supporting
GPL as well as providing it's own "Ease of Use" functions.
Rex
r.e.b...@usa.net
In article <68k675$i...@graves.maths.tcd.ie>,
t...@maths.tcd.ie (Timothy Murphy) wrote:
> >Linux are protected by GPL.
>
> if Linus Torvalds had decided to follow the FreeBSD licence,
> or for that matter if FreeBSD had decided to follow the GPL licence.
Even as late as 1994, there were restrictions on the FreeBSD code that
created barriers to it's distribution. In addition, FreeBSD and BSDi
are often in competition or percieved to be competitors. BSDi is great
stuff, but at $400 it's not as easy to give as a gift as a Linux CD-Rom.
I don't mind spending $15, $20, $40, or even $60 to give a boss, coworker,
friend, family member, or even a young student - a copy of Linux. When
we start talking $400, I have a harder time getting someone to choose
BSDi over NT 4.0.
Linux is cheap enough to counter the advantage Microsoft has of having
manufacturers pre-install MS-Windows. Perhaps some manufacturers will
start offering co-resident Linux/MS or UNIX/MS systems. If Caldera wins
it's suit, it could nullify some of Microsoft's "exclusive" licensing
agreements.
As Linux continues to gain popularity, and as other UNIX vendors begin
to exploit that market, we will probably begin to see computer makers
offering PCs with Preinstalled "Dual-Boot" systems.
> The development of TeX has been very similar in many ways
> to that of Linux,
> but the licence is completely different.
The core features of the GPL (copyright protection, assurance of source
code availability, restrictions on "proprietary" updates, and right to
freely redistribute) are still present. Remember that EMACS also
supports TeX and TeXInfo files - which adds leverage to TeX. In fact,
the EMACS TeXInfo files were the granddaddy of "hypertext", back before
we had mice and 1280x1024 graphics displays. Back when the same program
had to work with a vt100, Concept 108, or heathkit H19 terminal. (I
worked at one company where we had over 12 different kinds of terminal
connected to the same VAX - running BSD 4.2).
> >They don't have to worry about writing an
> >application, or adding an enhancement, or fixing a bug, only to have "Big
> >Brother Bill" cripple their version with "system upgrades" and replace it
> >with the "Big Brother Bill" version. Thousands of people contributed to
> >BSD from 4.1 to 4.4, and many found their products showing up as Windows
> >applications, with no recourse.
>
> As you say, many people contributed to BSD without the benefit of GPL.
> I have met many of these people,
> and cannot recall any of them ever complaining along the lines you suggest.
So which people are thrilled that "Brother Bill" has obsconded with their
source code and integrated it into the MS Suite. What are they doing
today? Are they getting royalty checks?
> >To this day, Linux and Apache dominate the Internet server market.
> This is certainly true of Apache, but is it true of Linux?
> I'd like to think it was, but I very much doubt it.
Linux has a higher "Box Count". But Solaris has more sites and pages.
> Timothy Murphy
> e-mail: t...@maths.tcd.ie
> tel: +353-1-2842366
> s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland
-------------------==== Posted via Deja News ====-----------------------
Marc Slemko
In <884098344...@dejanews.com> r.e.b...@usa.net writes:
>In article <68ms1b$5mm$1...@scanner.worldgate.com>,
> Marc Slemko <ma...@znep.com> wrote:
>> While I wouldn't necessarily agree with that, I fail to see how this
>> supports your contention that the GPL is the cause of greatness.
>> Apache is not GPLed. Your assertion that the GPL is the reason
>> for developer support of Linux isn't supported by anything you say.
>There are many variations on the FSF GPL, but fundamentally, the core
>provisions - copyrighted software, availability of freely distributed
>source code, and some assurance that the copyright holder isn't going to
>use your contributions to create a Microsoft "Superproduct" provide a
>structure that encourages contribution, cooperation, and user generated
>promotion.
You are completely missing the point. You are claiming that the GPL is
responsible for the success of Linux; the GPL specifically, not any
old license allowing free distribution and use.
You use Apache as an example to support this in general. Apache is
not GPLed and does not have a GPL-like license. Therefore, your
argument makes no sense.
James Youngman
[Mailed and posted]
>>>>> "r" == r e ballard <r.e.b...@usa.net> writes:
r> It's a fundamental issue of "Path of Least Resistance". Linux is
r> successful BECAUSE of GPL. If Linus had tried to keep Linux to
r> himself, his key interests would have been addressed (memory
r> management, scheduling, caching). But there wouldn't have been a
r> Linux Documentation project. There wouldn't have been EMACS, the
Emacs would have existed anyway. In fact, of course, it existed
first.
r> GNU clib, the hundreds of commands, bash. The korn shell, csh,
glibc 1.x also predated Linux. Bash would still have worked on
Linux because Linus ported it himself.
r> and sh were originally copyrighted and restricted. Lex and Yacc
r> were protected by Berkely.
No they weren't. Lex was written my Mike Lesk, at AT&T.
r> They used to joke about UNIX, calling it "that system created by
r> college kids". Of course, when those "college kids" graduated,
r> they formed companies like SUN and SGI.
And others went on to work for companies and specified Unix for new
systems.
o r c e l l . p o r t l a n d . o r . u s
In article <883793682...@dejanews.com>, <r.e.b...@usa.net> wrote:
>In article <67vglb$nos$1...@bell.pconline.com>,
> Ben Lindstrom <mou...@newton.pconline.com> wrote:
>>
>> Tim Smith <t...@halcyon.com> wrote:
>> > In article <882481764...@dejanews.com>, <r.e.b...@usa.net> wrote:
>> >>Very wrong! There have been over a dozen versions of UNIX for the Intel
>> >>machine including Xenix, Venix, PC-UX, AIX, SunOS/386, Solaris,
>> >>Interactive Unix, Univel, SCO, SysVR4, BSD, FreeBSD, BSDi, Minix, and of
>> >>course 7 distributions of Linux. The easiest to aquire was Linux. The
>> >>easiest to get past the corporate legal system was Linux. The easiest to
>> >>get upgrades for was Linux. The easiest to give to friends was Linux.
>>
>> > I don't see how Linux is easier to give to friends than FreeBSD or OpenBSD
>> > or NetBSD, unless Linux has better CD-ROM duplication software.
>
>Linux is easier to duplicate (legally) because of the GPL. There are
>several qualified restrictions on BSD.
And those are?
I've not found any significant restriction with any of the xBSD code
that isn't easily bested by the GPL; and this in both commercial and
non-commercial applications.
____
david parsons \bi/ the only thing that saves me from the GPL is that
\/ CD blanks are dirt cheap.
r.e.b...@usa.net
In article <69ea2r$5...@pell.pell.portland.or.us>,
o r c @ p e l l . p o r t l a n d . o r . u s (david parsons) wrote:
>
> In article <883793682...@dejanews.com>, <r.e.b...@usa.net> wrote:
> >In article <67vglb$nos$1...@bell.pconline.com>,
> > Ben Lindstrom <mou...@newton.pconline.com> wrote:
> >>
> >> Tim Smith <t...@halcyon.com> wrote:
> >> > In article <882481764...@dejanews.com>, <r.e.b...@usa.net> wrote:
> >> > I don't see how Linux is easier to give to friends than FreeBSD or OpenBSD
> >> > or NetBSD, unless Linux has better CD-ROM duplication software.
> >Linux is easier to duplicate (legally) because of the GPL. There are
> >several qualified restrictions on BSD.
> And those are?
The biggest one is that BSD allows you to create proprietary extensions
to existing binaries and to take "ownership" of the source code. As a
result, much of the better "ease of use" features are only offered in
BSDi, which is not even inexpensive (as much as $1400 for an "unimited
user" license.)
Linux has relied almost entirely on contributed software for it's
existence. As a result, Linux has been able to solve many of the "ease
of use" problems that plagued other versions of UNIX. It is a major leap
of faith for a user to take on the drama of shrinking the Windows
partition, creating the Linux Partitions, and installing the software
correctly. The trauma should be a short as possible, to get them using a
system that is as friendly as MS-Windows yet as powerful, fast, and
reliable as UNIX. It will take the Price/Performance of Linux to pierce
the "Armor" of the Microsoft "Per Processor" license. Until PC vendors
or retailers are willing to perform the installation of Linux or BSD
themselves, many of your more timid customers, especially Managers and
Corporate executives will be reluctant to try Linux and will be
supporting Microsoft out of Ignorance. Once managers can get a fully
functional version of Linux on their own Desktop, and have a pleasant
experience with it, they will be more inclined to consider some variant
of Linux or UNIX, including BSDi, Solaris, or SCO, on the basis of the
availability of ongoing support.
FreeBSD is now in the back of at least on major book, hardcover, and
at a cost of nearly $70, not including any great Office Suites.
> I've not found any significant restriction with any of the xBSD code
> that isn't easily bested by the GPL; and this in both commercial and
> non-commercial applications.
Actually, there are fewer restrictions on the baseline code. You can
make your own proprietary enhancements and charge as much as you like,
and keep them as your own private domain. This is the very thinking that
enabled Microsoft to exploit the splits in the UNIX community and capture
a major segment of the market using "Vaporware" (NT didn't even exist
when Gates first announced it, he was reacting to the comparisons between
Windows 3.1 and SunOS 3.x.. At at least one show, the Sun, running
nothing but a bunch of trivial applications totally humiliated Windows,
runnig special demonstration software and graphics intended to look good
in a show. Even the presence of a Microsoft Rep didn't help.
I'm willing to try FreeBSD again, but I haven't seen an overwhelming case
for dropping Linux based on what I've seen so far.
> ____
> david parsons \bi/ the only thing that saves me from the GPL is that
> \/ CD blanks are dirt cheap.
Rex Ballard
http://www.access.digex.net/~rballard
Brian Kimball
r.e.b...@usa.net wrote:
> It will take the Price/Performance of Linux to pierce
> the "Armor" of the Microsoft "Per Processor" license.
A license which MS stopped using in '94.
Roger Christie
Well after its work was done.
--
+=======================================+
Microsoft: Where do you want to go today?
Linux: Been there, done that.
Roger Christie
rog...@castle.ultranet.com
arathorn@Castle D'Image MUSH
+======================================+
Brian Kimball
Roger Christie wrote:
>
> Brian Kimball wrote:
> >
> > r.e.b...@usa.net wrote:
> >
> > > It will take the Price/Performance of Linux to pierce
> > > the "Armor" of the Microsoft "Per Processor" license.
> >
> > A license which MS stopped using in '94.
>
> Well after its work was done.
True, but based on Rex's comments in this thread and others, he seems to
think they still use it.
brian
o r c e l l . p o r t l a n d . o r . u s
In article <884999570....@dejanews.com>, <r.e.b...@usa.net> wrote:
>In article <69ea2r$5...@pell.pell.portland.or.us>,
> o r c @ p e l l . p o r t l a n d . o r . u s (david parsons) wrote:
>>
>> In article <883793682...@dejanews.com>, <r.e.b...@usa.net> wrote:
>> >In article <67vglb$nos$1...@bell.pconline.com>,
>> > Ben Lindstrom <mou...@newton.pconline.com> wrote:
>> >>
>> >> Tim Smith <t...@halcyon.com> wrote:
>> >> > In article <882481764...@dejanews.com>, <r.e.b...@usa.net> wrote:
>
>> >> > I don't see how Linux is easier to give to friends than FreeBSD or OpenBSD
>> >> > or NetBSD, unless Linux has better CD-ROM duplication software.
>
>> >Linux is easier to duplicate (legally) because of the GPL. There are
>> >several qualified restrictions on BSD.
>> And those are?
>
>The biggest one is that BSD allows you to create proprietary extensions
>to existing binaries and to take "ownership" of the source code.
Take ownership of your modifications of the source code, you mean; the
original code remains as it was.
>> I've not found any significant restriction with any of the xBSD code
>> that isn't easily bested by the GPL; and this in both commercial and
>> non-commercial applications.
>
>Actually, there are fewer restrictions on the baseline code. You can
>make your own proprietary enhancements and charge as much as you like,
>and keep them as your own private domain.
Right. That's one of the argued points against the freely
redistributable nature of Berkeley-style copyrights in favor of the
more restrictive GPL copyright. If you want to restrict people from
doing as they wish with the code, but still have the source available
for public consumption, the GPL is a much better licence. But it IS
more restrictive, in that I cannot add proprietary enhancements to it
and keep the source for myself.
____
david parsons \bi/ If someone wants to use my code for proprietary stuff,
\/ they are certainly welcome to it.
r.e.b...@usa.net
In article <34C0F56A...@pobox.com>,
The actual contract terms have changed to satisfy DOJ and more
importantly, to encourage the use of both 95 and NT in a mix that
encourages use of NT by business and use of 95 by individuals. Microsoft
still has some strong-arm tactics that it uses to discourage the
preinstallation of UNIX in boxes shipped from the factory. In the
fiercely competitive MS-Windows platform market, a little tuning or
detuning from Microsoft can radically alter benchmark results. A good
relationship with Microsoft can provide press and co-op.
The contracts are very complex and under strict non-disclosure.
Vendors are feeling pressure from both sides. Sun is now producing very
low cost, high performance ultaSparc workstations which is putting a cap
an the high end of the PC MS-Windows Market. Meanwhile, "Bare-Bones"
machines configured to be "Linux Ready" are bringing up the expectations
of the "Low-End". For a few hundred dollars I can have an Internet-ready
Client/Server Linux system. For a few thousand dollars, I can have a
fully supported high-end workstation. This leaves the "Microsoft-Only"
market getting squeezed for more price/performance with few concessions
coming from Microsoft.
Meanwhile, to compete against Linux in the "low-end", Microsoft has
to make MS-Windows platforms more "hardware-flexible". It makes it easier
to use cheaper hardware for 95 and NT. Unfortunately, this means slitting
the throats of their "High End" OEMs.
Microsoft is also trying to protect it's "Low-End" with Windows CE, which
has to be custom tuned to each piece of hardware. This is intended to
thwart the "NC your PC with Linux" movement. While it might lead to some
cheap "Web-TV" type boxes, there are still 80 million PCs currently
running Windows 3.1 and another 80 million running 95 that can benefit
from the installation of Linux as a supplimentary operating system.
As people begin to use Linux, there is a very high likelihood that they
will start to prefer it over MS-Windows 3.1 or Bare-Bones MS-Windows 95.
The press has already discovered that Linux can provide tremendous "added
value" to a basic MS-Windows PC. It will be only a matter of time before
a substantial portion (50-70%) see a greater value in Linux than they see
in MS-Windows.
> brian
Chris
Hmmm, I'm not so sure. The needs of your low-end computer users - Web & Email
- fit pretty nicely with WebTV and/or Windows CE devices. I wonder if desktop
machines for low-enders, meaning P-100s with Office 95, aren't kind of a hack,
in that they're a technically 'wrong' solution that's hateful and will go
away.
Bernd Paysan
Chris wrote:
> Hmmm, I'm not so sure. The needs of your low-end computer users - Web & Email
> - fit pretty nicely with WebTV and/or Windows CE devices.
Or Corel's NC. Which runs Linux. But IMHO the typical low-end computer
user's needs are best fit by a Nintendo N64.
--
Bernd Paysan
"Late answers are wrong answers!"
http://www.jwdt.com/~paysan/
Chuck Adams
>> Hmmm, I'm not so sure. The needs of your low-end computer users - Web & Email
>> - fit pretty nicely with WebTV and/or Windows CE devices.
>
>Or Corel's NC. Which runs Linux. But IMHO the typical low-end computer
>user's needs are best fit by a Nintendo N64.
It does not run Linux, it runs some kind of BSD.
Alan Shutko
>>>>> "C" == Chuck Adams <chuck@but_i_dont_like_spam.org> writes:
>> Or Corel's NC. Which runs Linux.
C> It does not run Linux, it runs some kind of BSD.
Perhaps you're thinking of Oracle's? According to
http://www.corelcomputer.com/products/vnc/overview/swhighlights.htm#1,
it's running Linux.
--
Alan Shutko <a...@acm.org> - By consent of the corrupted
Veni, Vidi, Borde. I came, I saw, I got bored.