[clug] Kernels are easy, ls is the hard part!

Hugh Fisher hugh.fisher at anu.edu.au
Thu Sep 13 07:30:11 GMT 2007

Richard Stallman continues to inflate himself in this
online interview:

> In 1992, the GNU system was complete except for the kernel. (Our own
> kernel project, started in 1990, was going slowly.) In February 1992,
> Linus Torvalds changed the license of Linux, making it free software.

> The kernel Linux filled the last major gap in GNU; the combination,
> GNU/Linux, was the first free operating system that could run on a
> PC. The system started out as GNU with Linux added. Please don't call
> it "Linux;" if you do that, you give the principal developer none of
> the credit. Please call it "GNU/Linux" and give us equal mention.

The arrogance of this takes my breath away.

You may have noticed that this message comes to you courtesy
of the IETF developed TCP/IP protocol suite. The IETF was
insisting on open standards and freely available code a decade
before RMS had his traumatic experience with a printer driver.
The Internet exists largely due to Berkeley System Labs and
their BSD licensed TCP/IP stack, which became the basis for
TCP/IP in Windows, MacOS, and just about every printer/router
manufacturer in the world.

Most likely you are reading this on a CRT/LCD monitor with
X Windows running underneath. X Windows was developed at MIT
and released as an open protocol standard and open source
implementation. Not so widely adopted as TCP/IP, but still
the standard for every Unix-like system and ported to Windows,
MacOS, VMS, and elsewhere.

To read the article, you'll fire up a web browser and connect
to the pcworld web server. The HTTP protocol started at the
IETF of course, the first widespread web browser was Mosaic,
which led to Netscape. And the archetypical web server is
Apache, an open source project and license.

While all these developments were going on, exactly what was
the FSF doing? They were writing implementations of the Unix
command line environment as defined by the POSIX specs. Bell
Labs, BSD, and others had done the initial design; and the
POSIX committees did the hard grind of agreeing on details
and resolving conflicts. The FSF then handed out man pages
for the POSIX spec to volunteers and told them to "write this".

What about gcc? I hear some people cry. Well yes, that was a
big project. But, you know, there were C compilers already.
Even open source ones.

And of course the one thing the FSF couldn't do was the bit
that was absolutely essential, a kernel. Proprietary companies
developed kernels in the 80s: Apple, Sun, DEC. Professors at
universities developed kernels: Minix, Mach. A very bright
student developed a kernel, which we all know about. About
the only software development group which proved incapable
of developing a kernel was the FSF!

If someone removed TCP/IP from your Linux system, perhaps
replacing it with AppleTalk or IPX, you'd notice and complain
very quickly. Ditto if X Windows were taken away. (And for
those of you reading this in a text environment, SSH was
written by BSD developers, not the FSF.) But if your CLI
environment was replaced by the BSD equivalents I think it
would be a while before you noticed, and if your kernel was
recompiled with the Intel C compiler (for x86, IBM if PPC,
etc) I doubt you'd notice at all unless you happen to be an
actual kernel developer.

The FSF contribution to Linux is important, because everyone
who contributes is important. But the GNU bits are neither
original nor difficult to replace, and any claims to be the
"principal developer" of Linux are laughable. Calling the
system "GNU/Linux" overstates their contribution and ignores
the good work of too many others.


More information about the linux mailing list