[linux-audio-user] Linux and Standards

tim hall tech at glastonburymusic.org.uk
Tue Nov 2 08:14:48 EST 2004


Last Tuesday 02 November 2004 03:00, Jack O'Quin was like:
> Jan Depner <eviltwin69 at cableone.net> writes:
> > On Mon, 2004-11-01 at 15:15, lau at lupulin.net wrote:
> > > On Mon, Nov 01, 2004 at 02:04:36PM -0600, Jan Depner wrote:
> > > > No disrespect intended to Richard Stallman and the GNU crowd.
> > > > The OS wouldn't exist without those tools but the tools are not
> > > > part of the OS.  They are merely applications that are bundled
> > > > in with the distribution.
> > > >
> > > > Given the more widely accepted definition of an operating system
> > > > I think it is perfectly acceptable to speak of Linux as a
> > > > standard.
> > >
> > > This is a gray area, but I think that you cannot just say that the gnu
> > > tools are _not_ a part of the operating system.
> > >
> > > Would you say that the startup scripts are _not_ a part of the OS ?
> > > All the startup scripts that I've seen rely are parts of gnu coreutils.
> > >
> > > I think that qualifies as being _part_ of the OS.
> >
> > Nope.  A startup script is just a startup script.  Grub is not part
> > of the operating system either.  The OS is, by definition, the
> > kernel.  An interesting thing to consider is RTLinux.  Linux is
> > *not* the OS in RTLinux.  The RT microkernel is the OS.  Linux is
> > merely the idle process.  I guess you could say it's part of the OS
> > since it is in the inner loop so to speak.
>
> I suppose reasonable people could disagree about the definition of
> "Operating System".  I know what it means to me, and that's a heck of
> a lot more than my friend Jan's minimalist definition.
>
> In the commercial world, when you buy Solaris from Sun, AIX from IBM,
> or Windows XP from M$, they call what they sell you an "operating
> system".  The kernel and all its device drivers represent about 5 or
> 10 percent of those products.  The rest is all the shells, utilities,
> install programs, startup scripts, etc. that are needed to actually
> "operate the system".  Without an OS, your computer is useless bare
> metal and silicon.
>
> It is useful to make a distinction between "applications" which are
> programs users actually *want* to run and "system utilities" which are
> all the programs needed to install and run those applications.  The
> system utilities are traditionally part of the OS.
>
> There are some grey areas in the realm of development tools.  IBM
> sells their compilers separately (not all users are programmers), but
> the linker is part of the OS, because some applications run a
> link-edit step when installing.  In an open source world, compilers
> *are* needed for installing many applications.  So, while the IBM
> compiler is not part of AIX, GCC arguably *is* part of GNU/Linux.
>
> In traditional Unix, the basic development tools were included in the
> OS.  (Anyone remember the Portable C Compiler?)  Thompson and Richie
> invented the "nut metaphor" for talking about this stuff.  The
> "kernel" was the seed in the middle of the nut, while the "shell" was
> the outer wrapping seen by users on their teletype machines.  Even in
> 1969, Unix was a great deal more than the kernel.  The scope of
> operating systems has grown considerably since then.
>
> I think Richard Stallman is right to say that "Linux" is only the
> kernel and that most of the GNU utilities are needed to make a
> complete OS.  And, I agree that GNU/Linux is a reasonable name for
> that OS.
>
> By this definition, Debian is much more than just an OS.  It also
> contains tens of thousands of applications.  So do Red Hat and the
> other distributions.  This is much of the power of open source
> software.

Thanks for clarifying. I like this explanation best :)
I fully realise I'm being a bit of a Debian/GNU/DeMuDi chauvinist here.

Respect to all

tim hall



More information about the Linux-audio-user mailing list