that old GNU/Linux argument

Patrick O'Callaghan pocallaghan at
Fri Jul 18 14:18:23 UTC 2008

On Fri, Jul 18, 2008 at 10:33 PM, Alexandre Oliva <aoliva at> wrote:

> On Jul 17, 2008, "Patrick O'Callaghan" <pocallaghan at> wrote:
> > On Thu, 2008-07-17 at 23:38 +0200, Björn Persson wrote:
> >> programs. Therefore, contrary to what you wrote, GNU does fit
> Alexandre's
> >> definition of "operating system".
> > Better tell him that, because he's not using the term in the way you
> > describe. This isn't a question of right or wrong, it's one of
> > consistency. If you check back on this thread you'll see that Alexandre
> > repeatedly refers to GNU/Linux as (paraphrasing here) "the GNU Operating
> > System plus the Linux kernel", which IIRC is in line with what the FSF
> > says when promoting GNU/Linux as a name. If "the GNU Operating System"
> > is complete, i.e. already has a kernel, then where does Linux fit in
> > this scheme? By this token, you should be referring to the Operating
> > System as GNU/Hurd, not as GNU.
> If this was arithmetics or formal logic, you'd be absolutely correct,
> and to be pedantically right we'd have to say GNU-Hurd+Linux.
> However common sense kicks in: if you replace or remove any single
> component of the large collection of programs that together amount to
> an operating system, this won't make enough of a different to make it
> a different operating system.  So, GNU-Hurd is still the GNU operating
> system.

So Operating System - kernel = Operating System? I think we're talking past
each other here. Clearly natural language is not the same as formal logic,
but that's no reason not to look for clarity. We're talking about
definitions here. In everyday usage we can be more relaxed (in fact I would
characterize the anti-GNU/Linux position being exactly that: people use the
terminology they find convenient and aren't too worried about exactness).

> And if you add Linux or the OpenSolaris kernel or the FreeBSD
> kernel, it's still the GNU operating system, running on top of a
> different kernel.
> Now compare this with the so common situation in which people take the
> name of the kernel for the name of the operating system.  You replace
> a single program, the kernel, that amounts to a small fraction of the
> whole; adjust another component, say libc, and leave everything else
> alone.  Would you say the name of the operating system should change
> in this case?

Replacing the kernel is not remotely on the same level as replacing some
random program, and counting lines of code is no way to assign importance.
For the that matter, replacing (or just removing) X, Gnome and/or KDE woul
reduce the code count by a huge amount, but there's no doubt in my mind that
we'd still call the result Linux (or GNU/Linux according to preference).

> > OTOH (and this is something I haven't raised before), what people in the
> > great majority *do* say is Linux, not GNU/Linux.
> Ad populum?  The great majority thinks Windows is part of the
> computer, but that so many people make this mistake doesn't make it
> right.  Oh, and they've never heard of GNU or Linux.

On that basis the majority of the world's population have never made a phone
call, let alone know what an OS is, or even what a computer is, so naturally
that's not the population I refer to. "The majority" means "the majority of
people who *do* know what these things are".

> > You have every right to object to that on principle, but forgive my
> > scepticism that it will actually change anything.
> If it weren't for the very man who actually started asking people to
> give credit to the project he started to give people freedom while
> using computers to start this project, nothing would have changed
> either, and we might very well find out we wouldn't have any Free
> operating system to use.

I think you're confusing two things. RMS defined Free Software (he didn't
invent it but he did formalize the idea) and deserves every credit for that.
The success of the idea is in part due to people signing up to the principle
because it appeals to them, and in part because it has clear economic
advantages. OTOH pushing a *name* for something a) appeals to far fewer
people, especially those who have become used to a different name, and b)
has no economic impact whatsoever. Add to that the fact that the proposed
name is more awkward to say and write than the preferred name, and I'd say
it's an uphill struggle.

> Be skeptic as much as you want, but he has
> already succeeded once at something that many thought impossible.
> > But it's not going to change.
> Everything eventually changes.

Time will tell.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the fedora-list mailing list