Re: GNU-Linux vs Linux naming [was RE: LAM/MPI Parallel processing]

["Stephen J. Turnbull" <turnbull@example.com>]

>>>>> "Scott" == Scott M Stone <sstone@example.com> writes:

    Scott> linux kernel isn't useless without the GNU tools.  You
    Scott> could just as easily compile the BSD tools on it.  Using a
    Scott> cross-compiler, of course, on a Sun machine running Sun's C
    Scott> compiler, to avoid using gcc :)

No, no, no ... you're missing the point.  We're talking about what
existing distros distribute---they're the ones the whole Lignux ->
GNU/Linux propaganda campaign was aimed at.  There are existing
embedded Linux distros, and while they have to do some hoop jumping
for each embedded app, that's the nature of embedded systems---WinCE
(what an appropriate acronym!!), EROS, whatever, is the same.

But there is no installable BSD/Linux system, and I'm sorry, on Linux
you can't avoid GCC (and GNU Make for that matter).  You have to be
able to rebuild the kernel[1]; QED.

Of course all of these things _can_ be done with sufficient hoop-
jumping, and you know what?  Stallman would cheer and wouldn't care
what you called them.  It's unlikely[2] that he'd lighten up on the
subset of GNU/Linux systems after all that water under the bridge.
But had there been multiple Linux systems (GNU/Linux, BSD/Linux,
newlib/Linux, Slowaris/Linux, AIX-Linux, HPlux, ...) from the start, I
doubt he could have gotten so exercised about GNU/Linux.  But as long
as all of those are hypotheticals, the only Linux is GNU/Linux.

The point is that due to rms and the GNU Project, there was a whole
set of OS utilities ready to run once the kernel was running.  Did
Linus and his coworkers cross-compile from Suns and use BSD libc?[3]
They did not; they dropped the kernel into a GNU system with a
somewhat modified (shut up, Chris, we know how you feel about HJ :)
GNU libc.

And that, my friends, is what I like about Open Source.

RMS of course has been rather miserly about giving credit in reverse.
Eg, it just occurred to me that the existence of the Linux kernel is
probably the main reason GNU libc v.2 actually got out the door.  A
libc is a tar pit just like a kernel, and we all know the amazing speed
at which the HURD has developed.[4]  (This is just a guess.  But in 1995
DJ Delorie told me GNU libc was not ready for prime time, so it wasn't
worth his while to port it to DJGPP.  He used BSD sources as the main
ingredient.)

Still there is a lesson here that open source ignores at its peril:
we're all in this together, and it pays to be polite.  We need to
remember our roots.  For Linux systems, they are gnee-deep in the
GNUpla.[5]


Footnotes: 
[1]  Adopting Steve Baur's definition of a complete system.

[2]  I have crossed swords with the man, and if you think I have
trouble backing down gracefully, you should see this guy.

[3]  Not a hypothetical; DJGPP did exactly that.

[4]  And the HURD is not the only kernel that wasn't born until it was
obsolete.  No question about it, Linus did an amazing thing, both in
terms of a virtuoso individual performance, and in terms of turning a
hurd[sic] of hackers into a kernel development machine.  It is a shame
it seems to have turned back into a hurd suitable for producing pre-
obsolete monstrosities....

[5]  Apologies to Ms. Slick only.


-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."