Is Linux written in C????
Frank Black
Ben Pfaff
> Just wondering if Linux is written in C.
Yes, but it uses system-specific features and GNU C extensions
that produce undefined behavior in strict ANSI C. It's really
written in GNU C, not ANSI C.
--
"What is appropriate for the master is not appropriate for the novice.
You must understand the Tao before transcending structure."
--The Tao of Programming
Tim Prince
topic.
"Frank Black" <purple...@hotmail.com> wrote in message
news:3a821a87$1...@spamkiller.newsfeeds.com...
Dan Pop
>Just wondering if Linux is written in C.
A meaningless question, as long as you don't specify which part of Linux
you're asking about.
Dan
--
Dan Pop
CERN, IT Division
Email: Dan...@cern.ch
Mail: CERN - IT, Bat. 31 1-014, CH-1211 Geneve 23, Switzerland
Micah Cowan
> In <3a821a87$1...@spamkiller.newsfeeds.com> "Frank Black" <purple...@hotmail.com> writes:
>
> >Just wondering if Linux is written in C.
>
> A meaningless question, as long as you don't specify which part of Linux
> you're asking about.
>
> Dan
No, it's not. Linux is a kernel. "Which part" would mean what -
which source file?
Micah
Dan Pop
>Dan...@cern.ch (Dan Pop) writes:
>
>> In <3a821a87$1...@spamkiller.newsfeeds.com> "Frank Black" <purple...@hotmail.com> writes:
>>
>> >Just wondering if Linux is written in C.
>>
>> A meaningless question, as long as you don't specify which part of Linux
>> you're asking about.
>>
>> Dan
>
>No, it's not. Linux is a kernel.
Since when? Could you point me to a Linux distribution consisting of a
kernel?
>"Which part" would mean what - which source file?
Nope. It would mean which component of Linux.
mike burrell
> In <yu87l30...@dyn-26-123.transmeta.com> Micah Cowan <mi...@cowanbox.com> writes:
>>Dan...@cern.ch (Dan Pop) writes:
>>> In <3a821a87$1...@spamkiller.newsfeeds.com> "Frank Black" <purple...@hotmail.com> writes:
>>>
>>> >Just wondering if Linux is written in C.
>>>
>>> A meaningless question, as long as you don't specify which part of Linux
>>> you're asking about.
>>
> Since when? Could you point me to a Linux distribution consisting of a
> kernel?
that there are distributions based on Linux doesn't change the fact that
it's a kernel. bad analogy:
>>>> Just wondering if fuse X is ceramic.
>>> A meaningless question, as long as you don't specify which part of fuse
>>> X you're asking about.
>> No, it's not. Fuse X is a fuse.
> Since when? Could you point me to an electrical appliance consisting of a
> fuse?
anyway, i thought the question was pretty cut and dry. Linux is almost
entirely written in GNU C.
--
/"\ m i k e b u r r e l l
\ / ASCII RIBBON CAMPAIGN mik...@home.com
X AGAINST HTML MAIL,
/ \ AND NEWS TOO, dammit finger mik...@mikpos.dyndns.org for GPG key
Dan Pop
>Dan Pop <Dan...@cern.ch> wrote:
>> In <yu87l30...@dyn-26-123.transmeta.com> Micah Cowan <mi...@cowanbox.com> writes:
>>>Dan...@cern.ch (Dan Pop) writes:
>>>> In <3a821a87$1...@spamkiller.newsfeeds.com> "Frank Black" <purple...@hotmail.com> writes:
>>>>
>>>> >Just wondering if Linux is written in C.
>>>>
>>>> A meaningless question, as long as you don't specify which part of Linux
>>>> you're asking about.
>>>
>>>No, it's not. Linux is a kernel.
>
>> Since when? Could you point me to a Linux distribution consisting of a
>> kernel?
>
>that there are distributions based on Linux doesn't change the fact that
>it's a kernel.
It is the kernel of what???
Micah Cowan
> In <yu87l30...@dyn-26-123.transmeta.com> Micah Cowan <mi...@cowanbox.com> writes:
>
> >Dan...@cern.ch (Dan Pop) writes:
> >
> >> In <3a821a87$1...@spamkiller.newsfeeds.com> "Frank Black" <purple...@hotmail.com> writes:
> >>
> >> >Just wondering if Linux is written in C.
> >>
> >> A meaningless question, as long as you don't specify which part of Linux
> >> you're asking about.
> >>
> >> Dan
> >
> >No, it's not. Linux is a kernel.
>
> Since when? Could you point me to a Linux distribution consisting of a
> kernel?
Linux distribution != Linux.
Linux distribution == operating system (specifically, the GNU one
running on a Linux kernel).
Linux == kernel
Unfortunately, many people (including myself occasionally) mistakenly
refer to Linux as an operating system, which would be more correctly
called GNU/Linux (or Linux/GNU I suppose). Every part of the
operating system besides the kernel is GNU on most systems, or at
least mostly so. In fairness, there is some borrowing from BSD, such
as the init process and a few other things, so the most correct term
would be GNU/BSD/Linux. IMO, if you have to shorten the name, GNU is
more correct, because it embodies the majority of the system.
Micah Cowan
A GNU/Linux user/devotee and supporter of free software
Dan Pop
>Dan...@cern.ch (Dan Pop) writes:
>
>> In <yu87l30...@dyn-26-123.transmeta.com> Micah Cowan <mi...@cowanbox.com> writes:
>>
>> >Dan...@cern.ch (Dan Pop) writes:
>> >
>> >> In <3a821a87$1...@spamkiller.newsfeeds.com> "Frank Black" <purple...@hotmail.com> writes:
>> >>
>> >> >Just wondering if Linux is written in C.
>> >>
>> >> A meaningless question, as long as you don't specify which part of Linux
>> >> you're asking about.
>> >>
>> >> Dan
>> >
>> >No, it's not. Linux is a kernel.
>>
>> Since when? Could you point me to a Linux distribution consisting of a
>> kernel?
>
>Linux distribution != Linux.
>
>Linux distribution == operating system (specifically, the GNU one
> running on a Linux kernel).
GNU is *not* an operating system, even if that was the original intention
of RMS. The operating system name is Linux. You don't have to take my
word for that:
ues12:~/tmp 16> uname --help | grep sysname
-s, --sysname print the operating system name
ues12:~/tmp 17> uname -s
Linux
Like most operating systems, it has a kernel. But the kernel is part
of the OS, not the OS itself.
Many important components of the OS have no connection with GNU (i.e.
they are not projects of the FSF).
BTW, I'm extensively using GNU tools on my Digital Unix workstation.
Does that make it a GNU workstation? :-)
>Linux == kernel
Linux itself has proved you wrong :-)
>Unfortunately, many people (including myself occasionally) mistakenly
>refer to Linux as an operating system, which would be more correctly
>called GNU/Linux (or Linux/GNU I suppose).
I have yet to see anybody calling it like that. That is, except for the
GNU homepage :-) I wonder why they don't stick to their own OS (Hurd),
that apparently nobody wants to use ;-)
BTW, RMS tried once to call it Lignux and got the whole Linux community
mad at him :-)
>Every part of the
>operating system besides the kernel is GNU on most systems, or at
>least mostly so.
Which of the following is GNU on the usual distributions:
1. The boot loader.
2. The init program.
3. The X server.
4. SVGAlib.
5. Perl (arguably the most popular scripting language in the Unix world).
6. tcsh and zsh, two of the most popular shells.
7. The various vi clones.
8. Netscape and Lynx.
9. Pine (probably the most popular mail reader on Linux systems).
Without the first two components you can't even boot the system properly.
Without the others you'd get a system so boring that few people would want
to use it these days.
GLOBAL LINK 2001
>of RMS. The operating system name is Linux. You don't have to take my
>word for that:
>
> ues12:~/tmp 16> uname --help | grep sysname
> -s, --sysname print the operating system name
> ues12:~/tmp 17> uname -s
> Linux
>
>Like most operating systems, it has a kernel. But the kernel is part
>of the OS, not the OS itself.
Ok you are correct in one respect: GNU is not an operating system
But Linux is not either, Linux is a kernel maintained by Linus Travolds
Infact as I am sure you know Linux does not even have it's own C librairy is
uses the GNU C librairy, it also does not have its own base packages like
passwd, wc, etc etc.
So you see Linux is just a kernel, you don't get an operating system until you
add a C librairy etc and other base packages
So what you are running is a distro, with Linux kernel X.X.X
An operating system like freebsd for example has all it's base software
maintained by the same people and is included with a native C librairy, boot
loader, base packages and many other things so you can say that you are running
freebsd etc
you can say you are running Linux because it is the popular slang, but it is
not true, while the kernel is the heart of the OS and just about the whole
thing when you get technical, it is not quite there (i.e. init, bootstrap). and
you could not have a functional OS with just the linux kernel by any stretch of
the imagination.
and if you do not believe me that linux is just a kernel, go to any linux
website and look at the little intro paragraph on everyone you will see
something like "Linux is a *kernel* maintained by Linus Travolds"
and about this thing with 'uname' when you see where it says linux, you will
see it gives the kernel version not the version of your distro so this proves
nothing except that you are using a linux kernel
Z
> In <UuGg6.79379$K8.30...@news1.rdc1.ab.home.com> mike burrell <mik...@home.com> writes:
>>>>No, it's not. Linux is a kernel.
>>
>>> Since when? Could you point me to a Linux distribution consisting of a
>>> kernel?
>>
>>that there are distributions based on Linux doesn't change the fact that
>>it's a kernel.
>
> It is the kernel of what???
Debian, RedHat, SuSE, Slackware (if that one still exists),
LST or any of the countless distributions. As a matter of fact
when Torvalds started working on this project (which was
more or less for training on C programming and i find a kernel
to be a very bad training project :-)) after a while he sent
what he had so far to some newsgroups (I don't remember which)
this is what is today know as the linux kernel. The term linux
today has become an elastic word. Anything based on the linux
kernel is merrily named linux.
--
Z (Zoran....@daimlerchrysler.com)
"LISP is worth learning for the profound enlightenment experience
you will have when you finally get it; that experience will make you
a better programmer for the rest of your days." -- Eric S. Raymond
Gib Bogle
flurry of hair-splitting.
Gergo Barany
OK, so those are not GNU packages, and the system should not be
called GNU. It doesn't follow that it should be called Linux
instead. Linux is a software package maintained by Linus Torvalds
that can be used as an operating system kernel. It doesn't include
any of the things you listed above.
Gergo
--
Depends on how you define "always". :-)
-- Larry Wall in <1997102116...@wall.org>
Richard Bos
> But Linux is not either, Linux is a kernel maintained by Linus Travolds
^^^^^^^^
Is it? That's new to me.
Yes, spelling flames suck. So does the rest this thread. May it die.
Richard
Dan Pop
>Once upon a while "Dan Pop" <Dan...@cern.ch> wrote:
>
>> In <UuGg6.79379$K8.30...@news1.rdc1.ab.home.com> mike burrell <mik...@home.com> writes:
>
>>>>>No, it's not. Linux is a kernel.
>>>
>>>> Since when? Could you point me to a Linux distribution consisting of a
>>>> kernel?
>>>
>>>that there are distributions based on Linux doesn't change the fact that
>>>it's a kernel.
>>
>> It is the kernel of what???
>
>Debian, RedHat, SuSE, Slackware (if that one still exists),
uname -s on RedHat and Slackware (and I'm pretty sure on other
distributions, too) returns "Linux" as the operating system name.
Dan Pop
>OK, so those are not GNU packages, and the system should not be
>called GNU. It doesn't follow that it should be called Linux
>instead. Linux is a software package maintained by Linus Torvalds
>that can be used as an operating system kernel.
And what is the name of the operating system using this kernel?
According to the uname command it's Linux. According to the overwhelming
majority of its users, it's Linux, too. If you disagree, please cite
your source(s).
BTW, the comp.os.linux sub-hierarchy looks like this:
comp.os.linux.advocacy Benefits of Linux compared to other operating systems.
comp.os.linux.announce Announcements important to the Linux community. (Moderated)
comp.os.linux.answers FAQs, How-To's, READMEs, etc. about Linux. (Moderated)
comp.os.linux.development.apps Writing Linux applications, porting to Linux.
comp.os.linux.development.system Linux kernels, device drivers, modules.
comp.os.linux.hardware Hardware compatibility with the Linux operating system.
comp.os.linux.m68k Linux operating system on 680x0 Amiga, Atari, VME.
comp.os.linux.misc Linux-specific topics not covered by other groups.
comp.os.linux.networking Networking and communications under Linux.
comp.os.linux.setup Linux installation and system administration.
comp.os.linux.x Linux X Window System servers, clients, libs and fonts.
These guys look quite convinced that Linux is an OS, too :-)
Z
>
> uname -s on RedHat and Slackware (and I'm pretty sure on other
> distributions, too) returns "Linux" as the operating system name.
The important note of mine was that the term linux
has become an elastic word. But you snipped it.
Richard Heathfield
>
> Frank Black wrote:
> >
> > Just wondering if Linux is written in C.
> >
> This is funny. A simple question with a simple answer (yes) elicits a
> flurry of hair-splitting.
Hair-splitting is our *job*.
--
Richard Heathfield
"Usenet is a strange place." - Dennis M Ritchie, 29 July 1999.
C FAQ: http://www.eskimo.com/~scs/C-faq/top.html
K&R answers, C books, etc: http://users.powernet.co.uk/eton
Lesser
> It is the kernel of what???
Snipped from http://www.gnu.org/
...Variants of the GNU operating system, which use the kernel Linux, are
now widely used; though these systems are often referred to as ``Linux'',
they are more accurately called GNU/Linux systems.
Richard Heathfield
>
> Once upon a while "Dan Pop" <Dan...@cern.ch> wrote:
> >
> > uname -s on RedHat and Slackware (and I'm pretty sure on other
> > distributions, too) returns "Linux" as the operating system name.
>
> The important note of mine was that the term linux
> has become an elastic word. But you snipped it.
<shrug> That's part of Usenet. People reply to the bits they want, not
the bits you want. Selective quotation happens.
Dan's done that to me, too, so I know how it feels. I don't think it's
deliberate, because I don't believe Dan is intellectually dishonest.
Dan Pop
>Once upon a while "Dan Pop" <Dan...@cern.ch> wrote:
>>
>> uname -s on RedHat and Slackware (and I'm pretty sure on other
>> distributions, too) returns "Linux" as the operating system name.
>
>The important note of mine was that the term linux
>has become an elastic word. But you snipped it.
It's a lot less elastic than you may think: *all* the essential
components of the OS are the same, the main difference is in the way
the distribution is packaged and installed. Joe User is quite unlikely
to see any difference between one distribution and another, if the same
packages are installed.
When I switched from Slackware to RedHat, the only detectable difference
was in the way some system administration tasks (e.g. network
configuration) were performed.
That's why people say "I'm a Linux user" without feeling that mentioning
the distribution they use is important.
Micah Cowan
> >that there are distributions based on Linux doesn't change the fact that
> >it's a kernel.
>
> It is the kernel of what???
Any operating system which chooses to use it, currently the only one
which uses it is GNU.
Micah
Micah Cowan
> >that there are distributions based on Linux doesn't change the fact that
> >it's a kernel.
>
> It is the kernel of what???
Sorry to reply a second time to the same post, but if you really want
more information, go to
http://www.gnu.org/gnu/linux-and-gnu.html
And argue with them if you desire (but please don't; they have enough
things to deal with).
Micah
Micah Cowan
> And what is the name of the operating system using this kernel?
> According to the uname command it's Linux. According to the overwhelming
> majority of its users, it's Linux, too. If you disagree, please cite
> your source(s).
<news list snipped>
>
> These guys look quite convinced that Linux is an OS, too :-)
>
> Dan
because many people believe something does not indicate fact. Please
support it with evidence. uname is acceptable evidence, but is not
the only source. GNU cites very compelling reasons why it should be
called GNU/Linux (see previously cited URL). And it is worth noting
there is no such thing as a Debian Linux distribution. According to
Debian, it's GNU/Linux (They are also developing a GNU/HURD
distribution). Also, when you download a kernel from ftp.kernel.org,
it is linux-2.4.1.tar.bz2 or somesuch, not linux-kernel-2.4.1.tar.bz2.
It is probably time to lay this thread to rest, and draw our own
conclusions based on the ample evidence presented on all sides.
In any case, it is clear that there is a very clear and unambiguous
meaning attached to the OP's question, which is easily answerable, and
in fact has already been answered.
Micah
Gergo Barany
> In <slrn987jmg.pe...@hold.otthon.at> gergo....@gmx.net (Gergo Barany) writes:
>
> >OK, so those are not GNU packages, and the system should not be
> >called GNU. It doesn't follow that it should be called Linux
> >instead. Linux is a software package maintained by Linus Torvalds
> >that can be used as an operating system kernel.
>
> And what is the name of the operating system using this kernel?
Linux, if you really insist. Do we agree that the term is
overloaded? If yes, do you still disagree with the statement that
started this subthread, namely "Linux is a kernel"?
Gergo
--
Heisenberg may have slept here...
Dan Pop
>This is funny. A simple question with a simple answer (yes) elicits a
>flurry of hair-splitting. ^^^
The simple answer above is almost certainly wrong :-)
Dan Pop
>>GNU is *not* an operating system, even if that was the original intention
>>of RMS. The operating system name is Linux. You don't have to take my
>>word for that:
>>
>> ues12:~/tmp 16> uname --help | grep sysname
>> -s, --sysname print the operating system name
>> ues12:~/tmp 17> uname -s
>> Linux
>>
>>Like most operating systems, it has a kernel. But the kernel is part
>>of the OS, not the OS itself.
>
>Ok you are correct in one respect: GNU is not an operating system
>
>But Linux is not either, Linux is a kernel maintained by Linus Travolds
>
>Infact as I am sure you know Linux does not even have it's own C librairy is
>uses the GNU C librairy,
Before switching to glibc (a few years ago), Linux did had its own
library.
And, having been involved in the glibc port to one Linux platform, I can
tell you that the glibc used by Linux is effectively its own C library:
major components of the library are written specially for Linux.
Look at what platforms are supported by glibc version 2 and be prepared
for a surprise.
>it also does not have its own base packages like passwd, wc, etc etc.
If you believe that *each* Unix flavour implements these utilities from
scratch, you're either naive or ignorant.
Should we stop calling Solaris and IRIX by their names, because their code
base is SVR4?
>So you see Linux is just a kernel, you don't get an operating system until you
>add a C librairy etc and other base packages
The Linux kernel plus these other compnents is called the Linux operating
system. Which eactly what I've proved by showing the output of uname -s.
Mark McIntyre
If you said "linux HAS a kernel" then everyone would shut up.
--
CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
Dan Pop
>Dan Pop <Dan...@cern.ch> wrote:
>> In <slrn987jmg.pe...@hold.otthon.at> gergo....@gmx.net (Gergo Barany) writes:
>>
>> >OK, so those are not GNU packages, and the system should not be
>> >called GNU. It doesn't follow that it should be called Linux
>> >instead. Linux is a software package maintained by Linus Torvalds
>> >that can be used as an operating system kernel.
>>
>> And what is the name of the operating system using this kernel?
>
>Linux, if you really insist.
So, we seem to get somewhere.
>Do we agree that the term is overloaded?
No. In common usage, Linux means a free, Unix-like operating system.
The "official" description of the Linux newsgroups clearly shows this.
>If yes, do you still disagree with the statement that
>started this subthread, namely "Linux is a kernel"?
Yup, of course. This statement is no truer today than it was yesterday.
I'm surprised that people who were supposed to have a clue (or at least
half of it) can't distinguish between an OS and its kernel.
Dan Pop
What *exactly* makes them an authoritative source on the issue?
I know RMS's arguments, and I don't buy them. He has exactly zero
credibility on this issue, because he's a (very) interested party.
When the Hurd project failed to achieve its intended goal, the only hope
left to RMS was to "hijack" Linux.
The Linux community at large simply ignored him.
Gergo Barany
> In <slrn988gvq.sf...@hold.otthon.at> gergo....@gmx.net (Gergo Barany) writes:
>
> >Dan Pop <Dan...@cern.ch> wrote:
> >> In <slrn987jmg.pe...@hold.otthon.at> gergo....@gmx.net (Gergo Barany) writes:
> >>
> >> >OK, so those are not GNU packages, and the system should not be
> >> >called GNU. It doesn't follow that it should be called Linux
> >> >instead. Linux is a software package maintained by Linus Torvalds
> >> >that can be used as an operating system kernel.
> >>
> >> And what is the name of the operating system using this kernel?
> >
> >Linux, if you really insist.
>
> So, we seem to get somewhere.
But that's not the major point. The kernel's name is what is being
discussed.
> >Do we agree that the term is overloaded?
>
> No. In common usage, Linux means a free, Unix-like operating system.
So if someone announces that "Linux 2.4" has been released, they
speak of an operating system? No, in this case the term refers to a
single peace of software that happens to be the kernel of the free
operating system of the same name. Both uses are common.
> The "official" description of the Linux newsgroups clearly shows this.
They deal with the OS called Linux. That doesn't mean there can't be
a kernel with the same name.
> >If yes, do you still disagree with the statement that
> >started this subthread, namely "Linux is a kernel"?
>
> Yup, of course. This statement is no truer today than it was yesterday.
So you say that there is *no* software package called "Linux" that
can be used as an operating system's kernel?
> I'm surprised that people who were supposed to have a clue (or at least
> half of it) can't distinguish between an OS and its kernel.
The issue here is not the distinction. The issue is that both the OS
and the kernel have the same name. The question "Is Linux written in
C?" does not give enough context to know which one was meant.
Gergo
--
Fat people of the world unite, we've got nothing to lose!
Micah Cowan
Dan...@cern.ch (Dan Pop) writes:
> >Do we agree that the term is overloaded?
>
> No. In common usage, Linux means a free, Unix-like operating system.
> The "official" description of the Linux newsgroups clearly shows
> this.
OH, REALLY!?!?!?
Well, how about the "official" FAQ-list for the Linux newsgroups (more
specifically, news:comp.os.linux.misc?
http://www.mainmatter.com/linux-faq/toc.html
FAQ 1.1: What Is Linux?
Linux is the KERNEL OF OPERATING SYSTEMS that look like and perform as
well or better than the famous operating system from AT&T Bell
Labs. Linus Torvalds and a loosely knit team of volunteer hackers from
across the Internet wrote (and still are writing) Linux from
scratch. It has all of the features of a modern, fully fledged
operating system: true multitasking, threads, virtual memory, shared
libraries, demand loading, shared, copy-on-write executables, proper
memory management, loadable device driver modules, video frame
buffering, and TCP/IP networking.
Most people, however, refer to the operating system kernel, system
software, and application software, collectively, as ``Linux,'' and
that convention is used in this FAQ as well.
[emphasis mine]
Which explains both that, among other things, you are wrong about
Linux /not/ being a kernel; and that the use of the term "Linux" to
refer to an OS is conventional, not technical.
> I'm surprised that people who were supposed to have a clue (or at least
> half of it) can't distinguish between an OS and its kernel.
ditto...
Micah
Dan Pop
>Dan...@cern.ch (Dan Pop) writes:
>
>> And what is the name of the operating system using this kernel?
>> According to the uname command it's Linux. According to the overwhelming
>> majority of its users, it's Linux, too. If you disagree, please cite
>> your source(s).
>
><news list snipped>
>>
>> These guys look quite convinced that Linux is an OS, too :-)
>>
>> Dan
>
>because many people believe something does not indicate fact. Please
>support it with evidence. uname is acceptable evidence, but is not
>the only source.
It is an *objective* and *reliable* source.
>GNU cites very compelling reasons why it should be
>called GNU/Linux (see previously cited URL).
You discard the beliefs of *many* people, then, immediately, invoke the
opinion of *one* person, RMS, who has a vested interest in the issue,
and therefore very little (if any) credibility. Is this your preferred
method of establishing what is a fact? :-)
> And it is worth noting
>there is no such thing as a Debian Linux distribution. According to
>Debian, it's GNU/Linux (They are also developing a GNU/HURD
>distribution).
What does uname -s display on the Debian Linux distribution?
>Also, when you download a kernel from ftp.kernel.org,
>it is linux-2.4.1.tar.bz2 or somesuch, not linux-kernel-2.4.1.tar.bz2.
What kind of facts can be derived from a file name?
When the site name is ftp.kernel.org, what could you expect to find
there?
>In any case, it is clear that there is a very clear and unambiguous
>meaning attached to the OP's question,
And what would that meaning be?
Dan Pop
>Dan Pop <Dan...@cern.ch> wrote:
>> In <slrn988gvq.sf...@hold.otthon.at> gergo....@gmx.net (Gergo Barany) writes:
>>
>> >Dan Pop <Dan...@cern.ch> wrote:
>> >> In <slrn987jmg.pe...@hold.otthon.at> gergo....@gmx.net (Gergo Barany) writes:
>> >>
>> >> >OK, so those are not GNU packages, and the system should not be
>> >> >called GNU. It doesn't follow that it should be called Linux
>> >> >instead. Linux is a software package maintained by Linus Torvalds
>> >> >that can be used as an operating system kernel.
>> >>
>> >> And what is the name of the operating system using this kernel?
>> >
>> >Linux, if you really insist.
>>
>> So, we seem to get somewhere.
>
>But that's not the major point. The kernel's name is what is being
>discussed.
Nope. The meaning of *unqualified* "Linux" in an *unspecified context*
is being discussed. See the subject line and my first reply, that
triggered this whole discussion.
>> >Do we agree that the term is overloaded?
>>
>> No. In common usage, Linux means a free, Unix-like operating system.
>
>So if someone announces that "Linux 2.4" has been released, they
>speak of an operating system? No, in this case the term refers to a
>single peace of software that happens to be the kernel of the free
>operating system of the same name. Both uses are common.
This is no longer *unqualified* "Linux". The extra information provides
enough context to figure out that "Linux" is used as short for "the Linux
kernel, version".
>> The "official" description of the Linux newsgroups clearly shows this.
>
>They deal with the OS called Linux. That doesn't mean there can't be
>a kernel with the same name.
It's normally referred to as "the Linux kernel", which is the common usage
for the kernels of other operating systems, too. When I say "Solaris",
I mean the whole OS, when I want to refer to its kernel I say "the
Solaris kernel". That is, unless the context of the discussion is
*explicitly* kernels.
>> >If yes, do you still disagree with the statement that
>> >started this subthread, namely "Linux is a kernel"?
>>
>> Yup, of course. This statement is no truer today than it was yesterday.
>
>So you say that there is *no* software package called "Linux" that
>can be used as an operating system's kernel?
I'm merely saying that substituting a part for the whole and vice versa
belongs to literature, not to technical discussions.
>> I'm surprised that people who were supposed to have a clue (or at least
>> half of it) can't distinguish between an OS and its kernel.
>
>The issue here is not the distinction. The issue is that both the OS
>and the kernel have the same name.
In different contexts, yes. In an unspecified context, Linux is the OS
and its kernel is referred to as the Linux kernel.
>The question "Is Linux written in
>C?" does not give enough context to know which one was meant.
According to common usage, as shown above, it refers to the whole OS.
If this newsgroup were dedicated to kernel programming, then it would
have referred to the Linux kernel.
GLOBAL LINK 2001
>for the kernels of other operating systems, too. When I say "Solaris",
>I mean the whole OS, when I want to refer to its kernel I say "the
>Solaris kernel". That is, unless the context of the discussion is
>*explicitly* kernels.
You see this is what the whole discussion is about, Solaris, FreeBSD, IRIX etc,
have a kernel, and they also have there own C librairy, there own init process,
and there own base software packages like adduser, ps , cd and such, linux is
not like this, the only thing linus travolds and the people who develop linux
around the world maintain is the kernel, The people who develop FreeBSD handle
the kernel the C librairy the base software packages bootstrap init processes
etc etc, everything that you need to have an OS, with linux on the other hand
you do not have a functional OS even by the largest stretch of the imagination
until you add a boot strap a C librairy an init process all which are borrowed
or picked and choose between distributions (well the C librairy is always GNU
but you get the idea) so the fact is when you get an OS like freebsd, freebsd
IS an OS everything that the developers take care of and maintain and develop
could make a functional OS, linux on the other hand would not be functional
without GNU and other sources they obtain things from etc etc, so you see just
because people may say they "are running linux" what they really mean is "I am
running X distro with linux kernel X.X.X (and usually glibc upgraded to
x.x.x-x.x :-) ) because when you really get down to it the facts are, the C
librairy IS NOT linux, the init process IS NOT linux, the software IS NOT linux
the kernel IS LINUX which means linux on it's own could not be an OS (would be
kind of hard with out an init process or boot strap etc etc) so they need third
party software to become and OS and would not be able to be an OS with just
"linux", i rest my case
-art
Dan Pop
I read this text as: "some obscure and unspecified authority has decreed
that unqualified ``Linux'' refers to a kernel, but pretty much everyone
else uses it as the name of the operating system using that kernel and
this is how it will be used in this document, too".
If there is *any* mention to the *technical* meaning of the word in the
quoted text, I've missed it.
Thanks for providing more support for my position ;-)
DiGGr _spamtrap_
>I read this text as: "some obscure and unspecified authority has decreed
>that unqualified ``Linux'' refers to a kernel, but pretty much everyone
>else uses it as the name of the operating system using that kernel and
>this is how it will be used in this document, too".
Can you specify an authority whose decree you *are* prepared to abide
by?
In the absence of such, this whole argument centres around the fact that
you have chosen to use the term "Linux" in a particular way simply
because a majority of others do.
You are attempting, fallaciously IMO, to counter appeal to authority
with appeal to majority.
>If there is *any* mention to the *technical* meaning of the word in the
>quoted text, I've missed it.
What does Linus himself refers to Linux as? A kernel, or an OS? Surely
that should settle it to *everyone's* satisfaction?
Dan Pop
>Dan...@cern.ch (Dan Pop) wrote:
>
>>I read this text as: "some obscure and unspecified authority has decreed
>>that unqualified ``Linux'' refers to a kernel, but pretty much everyone
>>else uses it as the name of the operating system using that kernel and
>>this is how it will be used in this document, too".
>
>Can you specify an authority whose decree you *are* prepared to abide
>by?
Yup, the majority of Linux users. I can't think of any other authority
on this particular matter.
>In the absence of such, this whole argument centres around the fact that
>you have chosen to use the term "Linux" in a particular way simply
>because a majority of others do.
Isn't this the most sensible thing to do, rather than trying to impose
a minority view?
>You are attempting, fallaciously IMO, to counter appeal to authority
>with appeal to majority.
To what "authority"???
>>If there is *any* mention to the *technical* meaning of the word in the
>>quoted text, I've missed it.
>
>What does Linus himself refers to Linux as? A kernel, or an OS?
Ask him and post his answer.
>Surely that should settle it to *everyone's* satisfaction?
Nope. Linus' authority is limited to kernel issues only.
DiGGr _spamtrap_
>In <3a8bc6c7....@news.ntlworld.com> DiGGr _spamtrap_ <Di...@bigfoot.com> writes:
>
>>Dan...@cern.ch (Dan Pop) wrote:
>>
>>>I read this text as: "some obscure and unspecified authority has decreed
>>>that unqualified ``Linux'' refers to a kernel, but pretty much everyone
>>>else uses it as the name of the operating system using that kernel and
>>>this is how it will be used in this document, too".
>>
>>Can you specify an authority whose decree you *are* prepared to abide
>>by?
>
>Yup, the majority of Linux users. I can't think of any other authority
>on this particular matter.
The majority view is not necessarily correct. "Flat Earth" and
"geocentric astronomy" are but two examples that spring to mind.
Putting that to one side for the moment.
You claim that your authority is definitive and they claim that theirs
is definitive. Neither is prepared to compromise. Impasse.
>>In the absence of such, this whole argument centres around the fact that
>>you have chosen to use the term "Linux" in a particular way simply
>>because a majority of others do.
>
>Isn't this the most sensible thing to do, rather than trying to impose
>a minority view?
Generally, no. In ideas, as in many other things, it is quality not
quantity that is important.
As Nitpicker Extrordinaire to the ISO C Standard, you of all people
should know that. *Many people* cast the return value from malloc().
That doesn't mean it's right.
>>You are attempting, fallaciously IMO, to counter appeal to authority
>>with appeal to majority.
>
>To what "authority"???
To the authority of the comp.os.linux.misc FAQ. You may not accept it
as an authority, but that is nevertheless the capacity in which it is
being presented.
They say "this authoritative document says X" - appeal to authority;
you say "that is not authoritative, and anyway most people say !X" -
appeal to majority.
*Both* of these arguments are fallacious.
By the way, your ques key appears to be bouncing.
>>>If there is *any* mention to the *technical* meaning of the word in the
>>>quoted text, I've missed it.
>>
>>What does Linus himself refers to Linux as? A kernel, or an OS?
>
>Ask him and post his answer.
I don't do email.
>>Surely that should settle it to *everyone's* satisfaction?
>
>Nope. Linus' authority is limited to kernel issues only.
He invented Linux and, I understand, wrote a good portion of it. I
imagine he should have a pretty clear idea of what it is.
So, if you claim that Linus wrote Linux and that Linus only wrote the
kernel, then by equivalence you are claiming that Linux is the kernel.
Conversely, if you are *not* making this claim, then you cannot
legitimately restrict Linus' authority to the kernel.
Richard Heathfield
>
<snip>
>
> I read this text as: "some obscure and unspecified authority has decreed
> that unqualified ``Linux'' refers to a kernel, but pretty much everyone
> else uses it as the name of the operating system using that kernel and
> this is how it will be used in this document, too".
Here's what the ISO C Standard has to say on the matter:
Joona I Palaste
The ISO C Standard sure doesn't talk too much, does it?
--
/-- Joona Palaste (pal...@cc.helsinki.fi) ---------------------------\
| Kingpriest of "The Flying Lemon Tree" G++ FR FW+ M- #108 D+ ADA N+++|
| http://www.helsinki.fi/~palaste W++ B OP+ |
\----------------------------------------- Finland rules! ------------/
"To know me IS to love me."
- JIPsoft
Gib Bogle
operating system?
--
------------------------------------------------------------
Gib Bogle Ph.D. bo...@ihug.co.nz
Water Engineering & Modelling Tel: (64-9) 525-6878
1/44 Arthur St Fax: (64-9) 525-6878
Ellerslie, N.Z.
------------------------------------------------------------
Richard Heathfield
>
> Richard Heathfield wrote:
> >
> > Gib Bogle wrote:
> > >
> > > Frank Black wrote:
> > > >
> > > > Just wondering if Linux is written in C.
> > > >
> > > This is funny. A simple question with a simple answer (yes) elicits a
> > > flurry of hair-splitting.
> >
> > Hair-splitting is our *job*.
> >
> Someone pays you to determine whether "Linux" refers to a kernel or an
> operating system?
That's not what I said. I said that hair-splitting is our job. That's
not the same as saying that someone pays us to determine whether "Linux"
refers to a kernel or an operating system.
FWIW I think it /does/ refer to a kernel or an operating system. That
doesn't mean it doesn't refer to other stuff too, of course.
GLOBAL LINK 2001
>that should settle it to *everyone's* satisfaction?
>
>
and this should settle it too: The C librairy is not linux it is GNU the INIT
process is not linux it is a combination of things but that is another story
but it is not linux, and everything else that is not the kernel is not linux.
so let me ask you this: can you have an OS without an INIT process? how could
you login start everything up etc? how about without a C librairy? so unless
you guys are arguing that GNU IS LINUX and that BSD/whatever/something INIT
process IS LINUX i don't see how you can call linux an OS
GLOBAL LINK 2001
are you dense? linux IS the kernel damnit!
name one thing besides the kernel that is Linux name ONE thing go ahead i dare
you :-)
Clark S. Cox, III
> Dan...@cern.ch (Dan Pop) wrote:
> So, if you claim that Linus wrote Linux and that Linus only wrote the
> kernel, then by equivalence you are claiming that Linux is the kernel.
Linux *is* the kernel, nothing more, nothing less, the rest of the
OS is GNU. Linus wrote the kernel, therefore, he wrote Linux. The proper
name of the OS is GNU/Linux. Look at
<http://www.gnu.org/gnu/why-gnu-linux.html>
> Conversely, if you are *not* making this claim, then you cannot
> legitimately restrict Linus' authority to the kernel.
Linux is the kernel; the kernel is Linux.
--
http://www.whereismyhead.com/clark/
Clark S. Cox, III
clar...@yahoo.com
Dan Pop
Since it doesn't contradict me... :-)
Dan Pop
Wrong example, because casting the return value of malloc is *technically*
correct :-)
You seem to badly confuse things where what is correct can be proved or
demonstrated irrefutably and things where this is not possible.
The shape of the Earth can be determined, the suitablility of a C
construct can be assessed based on the formal definition of the language.
However, the answer to the question "what is the most popular Cola-style
beverage" can be obtained *only* if there is a majority view. On such an
issue, the majority view *cannot* be wrong.
In the absence of any international standard defining the exact meaning(s)
of "Linux", the only authority of any relevance is the majority view.
>>>You are attempting, fallaciously IMO, to counter appeal to authority
>>>with appeal to majority.
>>
>>To what "authority"???
>
>To the authority of the comp.os.linux.misc FAQ. You may not accept it
>as an authority, but that is nevertheless the capacity in which it is
>being presented.
Where does this FAQ derive its authority from?
>By the way, your ques key appears to be bouncing.
It's only an illusion :-)
Paul D. Boyle
: Debian, RedHat, SuSE, Slackware (if that one still exists),
: LST or any of the countless distributions. As a matter of fact
: when Torvalds started working on this project (which was
: more or less for training on C programming and i find a kernel
: to be a very bad training project
<offtopic>
Actually, I think it started as writing a task swapper exercise.
Then he wanted a version of Unix for his new 386, but there was no
version which suited him, so he started writing his own.
: after a while he sent
: what he had so far to some newsgroups (I don't remember which)
comp.os.minix
</offtopic>
Paul
--
Paul D. Boyle
bo...@laue.chem.ncsu.edu
North Carolina State University
http://laue.chem.ncsu.edu/web/xray.welcome.html
Mark McIntyre
>In <3a87d984....@news.ntlworld.com> DiGGr _spamtrap_ <Di...@bigfoot.com> writes:
>
>>As Nitpicker Extrordinaire to the ISO C Standard, you of all people
>>should know that. *Many people* cast the return value from malloc().
>>That doesn't mean it's right.
>
>Wrong example, because casting the return value of malloc is *technically*
>correct :-)
You're going to have to explain that pretty well. How can it be
technically correct to do something utterly unnecessary, and which
hides a bug ?
Dan Pop
>Z (Zoran....@daimlerchrysler.com) wrote:
>: Debian, RedHat, SuSE, Slackware (if that one still exists),
>: LST or any of the countless distributions. As a matter of fact
>: when Torvalds started working on this project (which was
>: more or less for training on C programming and i find a kernel
>: to be a very bad training project
>
><offtopic>
>Actually, I think it started as writing a task swapper exercise.
Correct.
>Then he wanted a version of Unix for his new 386, but there was no
>version which suited him, so he started writing his own.
What he actually wanted was to learn how to program his new 386 and how
a kernel works, so he started to expand his task swapper into a proper
kernel. Because of this, the first implementation was closely tied to
the 386 architecture. He later repented and ported it to the Alpha
architecture, properly isolating all the architecture dependencies in the
process.
The feature that accounts (to a large extent) for Linux' success over
its considerably more mature "competitor" at the time (early nineties),
386BSD, was not actually the kernel, it was shared libraries
(implemented as a quick and dirty hack, BTW). Having shared libraries,
its executables were quite compact and a Linux installation could
easily fit on the hard disk of a typical 386 PC at the time. IIRC,
Linus' PC had a 60 MB disk. The statically linked executables of
386BSD required a considerably larger disk, which most of the people
who wanted to run Unix on their PC's (mainly CS students) could not
afford.
Dan Pop
That's easy: such a cast is allowed by the standard and it doesn't break
any *correct* program.
It's stylistically wrong (IMHO) to do it, but so are the casts that so
many people put in their code to shut up the silly warnings about possible
loss of precision issued by Borland and Microsoft compilers.
The best argument in favour of casting malloc is that it allows
portability between C and C++ compilers (if the code is written with such
portability in mind). If the coding conventions *imposed* by your
company require C++ compatibility you have to choose between doing it
or finding another job...
Mark McIntyre
<who cares>
If you guys will promise not to change the subject line again, we'dall
be bery happy. I'm getting fed up adjusting my killfile to remove this
stupidly offtopic thread.
--
CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
DiGGr _spamtrap_
[...]
>>>>In the absence of such, this whole argument centres around the fact that
>>>>you have chosen to use the term "Linux" in a particular way simply
>>>>because a majority of others do.
>>>
>>>Isn't this the most sensible thing to do, rather than trying to impose
>>>a minority view?
>>
>>Generally, no. In ideas, as in many other things, it is quality not
>>quantity that is important.
>>
>>As Nitpicker Extrordinaire to the ISO C Standard, you of all people
>>should know that. *Many people* cast the return value from malloc().
>>That doesn't mean it's right.
>
>Wrong example, because casting the return value of malloc is *technically*
>correct :-)
I didn't say it wasn't technically correct, I said it wasn't "right".
On the basis of past postings in clc it seems reasonable to suppose that
most of the regular posters will point out that casting malloc(), even
though strictly correct, is unnecessary and error-prone. Is it not your
principle to accept the majority view?
>You seem to badly confuse things where what is correct can be proved or
>demonstrated irrefutably and things where this is not possible.
Now you're just trying to confuse the issue.
The *only* reason that the definition of "Linux" cannot be established
is because you flatly refuse to entertain any evidence that contradicts
your prejudice.
>The shape of the Earth can be determined, the suitablility of a C
>construct can be assessed based on the formal definition of the language.
The shape of the Earth can only be determined if you accept scientific
determination of the shape of the Earth. If you're a flat-Earth fundie,
all bets are off. If *most people* are flat-Earth fundies, then the
Earth is flat - by definition, according to your principle.
Similarly, there is no law which forces a compiler to be C90 or C99
compliant. Any compiler vendor is perfectly at liberty to reject the
ISO Standard definition of C. I respectfully remind you that the
Standard prevails by popular consent, not because it is inevitable.
Evidently, acceptance of your supposedly "irrefutable" facts is more
subject to popular consent than you had imagined.
>However, the answer to the question "what is the most popular Cola-style
>beverage" can be obtained *only* if there is a majority view. On such an
>issue, the majority view *cannot* be wrong.
Fallacious. You are assuming your conclusion.
We were discussing the *correct* definition of Linux, not the popular
definition. It is only you who assumes the two are synonymous.
>In the absence of any international standard defining the exact meaning(s)
>of "Linux", the only authority of any relevance is the majority view.
No, the only authority of any relevance is the *informed* view.
The *majority* of Earth's 6 billions have almost certainly never heard
of Linux, and no doubt care less about what it might be.
Besides, how can there be an International Standard definition of
"Linux" when nobody can agree what it is?
>>>>You are attempting, fallaciously IMO, to counter appeal to authority
>>>>with appeal to majority.
>>>
>>>To what "authority"???
>>
>>To the authority of the comp.os.linux.misc FAQ. You may not accept it
>>as an authority, but that is nevertheless the capacity in which it is
>>being presented.
>
>Where does this FAQ derive its authority from?
I really have no idea, since I haven't read it.
Please note that all I said was that it was presented as an authority.
I didn't say it was an authority. (However, I find it difficult to
imagine that it is any less authoritative than any other reasonably
informed source.)
I also find it difficult to imagine that this thread is of any interest
or relevance to anyone any more.
I have shown your position to be logically untenable. (I notice you
snipped that bit. :) So unless you have something to add that doesn't
essentially boil down to "everyone who disagrees with the majority is
wrong", I suggest that we call it a day.
Dik T. Winter
> You see this is what the whole discussion is about, Solaris, FreeBSD, IRIX
> etc, have a kernel, and they also have there own C librairy, there own init
> process, and there own base software packages like adduser, ps , cd and
> such, linux is not like this, the only thing linus travolds and the people
> who develop linux around the world maintain is the kernel,
Eh? I know somebody who did work on the Linux fsck.
--
dik t. winter, cwi, kruislaan 413, 1098 sj amsterdam, nederland, +31205924131
home: bovenover 215, 1025 jn amsterdam, nederland; http://www.cwi.nl/~dik/
Z
> In article <20010209224107...@ng-fe1.aol.com> globall...@aol.com (GLOBAL LINK 2001) writes:
> > You see this is what the whole discussion is about, Solaris, FreeBSD, IRIX
> > etc, have a kernel, and they also have there own C librairy, there own init
> > process, and there own base software packages like adduser, ps , cd and
> > such, linux is not like this, the only thing linus travolds and the people
> > who develop linux around the world maintain is the kernel,
>
> Eh? I know somebody who did work on the Linux fsck.
and I heard of somebaody who knows sombody who sometime ago used
to know somebody who knew somebody who was working on special
Linux version of a calculator program. There is no point in
that.
On the other hand there is lot of code written for linux. So
GLOBAL LINK 2001 seams to be missing the point too.
--
Z (Zoran....@daimlerchrysler.com)
"LISP is worth learning for the profound enlightenment experience
you will have when you finally get it; that experience will make you
a better programmer for the rest of your days." -- Eric S. Raymond
Z
> The proper
> name of the OS is GNU/Linux. Look at
> <http://www.gnu.org/gnu/why-gnu-linux.html>
That is only if you're refering to GNU but if not than there
is no point in that. Go to RedHat or SuSE and find out how they
call it.
Matthew Palmer
>>> ues12:~/tmp 17> uname -s
>>> Linux
>
>>So you see Linux is just a kernel, you don't get an operating system until you
>>add a C librairy etc and other base packages
>
>The Linux kernel plus these other compnents is called the Linux operating
>system. Which eactly what I've proved by showing the output of uname -s.
And I'll prove differently:
mjp16@inductor:~$ uname -s
GNU/Linux
Oops...
--
-----------------------------------------------------------------------
#include <disclaimer.h>
Matthew Palmer
mj...@ieee.uow.edu.au
Clark S. Cox, III
> Once upon a while "Clark S. Cox, III" <clar...@yahoo.com> wrote:
> > The proper
> > name of the OS is GNU/Linux. Look at
> > <http://www.gnu.org/gnu/why-gnu-linux.html>
>
>
> That is only if you're refering to GNU but if not than there
> is no point in that. Go to RedHat or SuSE and find out how they
> call it.
RedHat and SuSE are both GNU (just look at any of the included
utilities like gnutar, gcc, etc.), in fact every distro of Linux (that
I've ever seen) uses GNU software for every "UNIX-like" app. So GNU's
information is valid for all Linux distros.
If someone finds a Linux distro that isn't based on GNU then I hope
they come forward and say so.
--
http://www.whereismyhead.com/clark/
Dan Pop
>I have shown your position to be logically untenable.
You *think* you have. Big difference!
Dan Pop
>Z <Zoran....@daimlerchrysler.com> wrote:
>
>> Once upon a while "Clark S. Cox, III" <clar...@yahoo.com> wrote:
>> > The proper
>> > name of the OS is GNU/Linux. Look at
>> > <http://www.gnu.org/gnu/why-gnu-linux.html>
>>
>>
>> That is only if you're refering to GNU but if not than there
>> is no point in that. Go to RedHat or SuSE and find out how they
>> call it.
>
> RedHat and SuSE are both GNU (just look at any of the included
>utilities like gnutar, gcc, etc.), in fact every distro of Linux (that
>I've ever seen) uses GNU software for every "UNIX-like" app. So GNU's
>information is valid for all Linux distros.
What does uname -s display on RedHat and SuSE?
> If someone finds a Linux distro that isn't based on GNU then I hope
>they come forward and say so.
If someone finds a Linux distro that is based *exclusively* on the Linux
kernel and GNU software, I'm all ears.
Jan Schaumann
>
> What does uname -s display on RedHat and SuSE?
with RH it's "Linux", but so it is with Debian.
>
> > If someone finds a Linux distro that isn't based on GNU then I hope
> >they come forward and say so.
>
> If someone finds a Linux distro that is based *exclusively* on the Linux
> kernel and GNU software, I'm all ears.
Well, Debian is quite close. that's why it's called Debian GNU/Linux.
-Jan
--
Jan Schaumann <http://www.netmeister.org>
$ drink <bottle; opener
bottle: cannot open
opener: not found
Sascha Bohnenkamp
> Like most operating systems, it has a kernel. But the kernel is part
> of the OS, not the OS itself.
after which definition?
imho the OS is the programm which manages the hardware for applications,
e.g.
devices, filesystem, network, memory, task-management etc. which is what
the kernel does ... kernel == OS.
the rest are applications running on it.
Sascha Bohnenkamp
> And I'll prove differently:
>
> mjp16@inductor:~$ uname -s
> GNU/Linux
if you hack it it got do
# uname -s
not windos
you have the sources ... those messages mean nothing
Dan Pop
>Dan Pop wrote:
>
>> Like most operating systems, it has a kernel. But the kernel is part
>> of the OS, not the OS itself.
>after which definition?
>
>imho the OS is the programm which manages the hardware for applications,
>e.g.
>devices, filesystem, network, memory, task-management etc. which is what
>the kernel does ... kernel == OS.
Your opinion may be humble, but it is dead wrong. Any idea where the
terms of "kernel" and "shell" come from?
If you don't, I'll tell you: one is the kernel of the operating system,
the other is the shell of the operating system. Figuring out what
components of the operating system are between them is left as an
exercise to the reader.
>the rest are applications running on it.
OK, let's say I give you a kernel, an X server, Netscape, xtetris (all
statically linked) and... nothing else. With these components, build
a system that boots, starts the X server and runs Netscape and
xtetris.
Dan Pop
Unless they come from the unhacked sources.
Z
> OK, let's say I give you a kernel, an X server, Netscape, xtetris (all
> statically linked) and... nothing else. With these components, build
> a system that boots, starts the X server and runs Netscape and
> xtetris.
>
Many applications depend on other applications/libraries. What
is the point here?
Anyways, don't you think it's time to drop this silly thread or
take it to comp.os.linux.advocacy for instance? This has by no
means been a discussion about the C language nor has it been a
discussion about C's environment.
Matthew Palmer
Exactly. Thankyou for stating my point.
Dan Pop
>Once upon a while "Dan Pop" <Dan...@cern.ch> wrote:
>
>> OK, let's say I give you a kernel, an X server, Netscape, xtetris (all
>> statically linked) and... nothing else. With these components, build
>> a system that boots, starts the X server and runs Netscape and
>> xtetris.
>>
>
>Many applications depend on other applications/libraries. What
>is the point here?
If you read again what I wrote you'll see that I've constructed a
scenario where no other libraries/applications are needed (assuming
that the previous poster was right). Well, Netscape won't be able to
use plugins and start external applications, but it could still
function as a Web browser.
I thought that the point was quite clear: the kernel is NOT the OS.
Apparently, not clear enough for you...
>Anyways, don't you think it's time to drop this silly thread or
>take it to comp.os.linux.advocacy for instance? This has by no
>means been a discussion about the C language nor has it been a
>discussion about C's environment.
Then, why did YOU keep contributing to this thread? ;-)
Dan Pop
>Sascha Bohnenkamp is of the opinion:
>>> And I'll prove differently:
>>>
>>> mjp16@inductor:~$ uname -s
>>> GNU/Linux
>>
>>if you hack it it got do
>>
>># uname -s
>>not windos
>>
>>you have the sources ... those messages mean nothing
>
>Exactly. Thankyou for stating my point.
Many configuration scripts *rely* on the output of uname -s when choosing
the right parameters for the given platform. Tampering with it is a quite
efficient way of shooting yourself in the foot.
Sascha Bohnenkamp
> I thought that the point was quite clear: the kernel is NOT the OS.
> Apparently, not clear enough for you...
there exist many embedded-os where the apps run directly ontop the
kernel.
(muC/OS for example)
well and afaik the shell is a user-interaface to the os.
e.g. running on the kernel and making is usable by human-beings.
Sascha Bohnenkamp
> Unless they come from the unhacked sources.