🚀 go-pugleaf

Antti-Juhani Kaijanaho <gaia+news@iki.fi> writes:

> [Moving to gnu.misc.discuss.]
> 
> David Kastrup <dak@mailhost.neuroinformatik.ruhr-uni-bochum.de> writes:
> 
> > No, it's the architect of the assembly that determines its name.
> 
> I have small problems with that statement, but I'll accept it as your
> position.  But in light of that position, I find your statement below
> a little odd:
> 
> > For that reason I call Linux Linux
> 
> I'm assuming you are not referring to the Linux kernel but to the
> system that has evolved around it.  You know, Linus Torvalds did not
> architect the system.  He architected and implemented the original
> kernel.  The GNU system sans a kernel was already there when he
> started Linux.

No it wasn't.  A lot of parts for a GNU system were there, but never
yet before had a coherent system been assembled from those.  For that
reason, a lot of glue was missing (for example, system utilities), and
a lot of work had to go into the libraries as well to make them work
properly as stand-alone libraries.  This lead to the libc5 code
diversions between the GNU variant of the C libraries and the Linux
variant.

Fortunately, by now we again have a coherent library (libc6), but to
say that this was how the things started out is preposterous.

The first GNU systems were pioneered with Linux.  Talking about an
"existing GNU system" before Linux is simply belying the facts.  A
large ratio of the parts existed, and they were in part *very*
well-made and versatile parts, but they are parts.  For example, they
can find their files and other stuff at arbitrary locations depending
on the system they are compiled for.  There was no GNU files system
standard around at that time and still isn't.  There is a Linux file
system standard.

The GNU utilities can be made to fit a large variety of different
systems, but they do not come with a "GNU flavour".  They form an
excellent example for portable engineering techniques, but calling
them a complete system instead of parts had been wrong before the
advent of Linux.

Good parts are very necessary for building a system, but the presence
of them does not mean that a coherent system has already been made.


-- 
David Kastrup                                     Phone: +49-234-700-5570
Email: dak@neuroinformatik.ruhr-uni-bochum.de       Fax: +49-234-709-4209
Institut f�r Neuroinformatik, Universit�tsstr. 150, 44780 Bochum, Germany