Re: What is up with Redhat 7.0?

From: Kristofer T. Karas (ktk@bigfoot.com)
Date: Mon Oct 02 2000 - 02:24:29 EST


Alan Cox wrote:

> I don't see your point except as 'never change anything'. I got bored of
> libc2 a while back. I prefer change

Reasonable coordinated change, or reckless change?

How about shipping a new distribution with a 2.3.x kernel patched in-house to
get rid of most of the oopsen? This is not conceptually all that different
than shipping a distribution with a patched version of a not-yet-released
compiler. You are trading innovation for stability; innovation is only useful
to people if it offers sufficient stability and permanency to get the job done.

Change that propagates smoothly through acceptable channels is no doubt a good
thing. The Linux Kernel enjoys a high rate of change and attempts to appeal to
a wide audience; for this it has achieved excellent name recognition
world-wide. Despite that, with the possible exception of early versions of
0.x, the kernel has always offered a stable branch that changed comparatively
little versus the development branch; this gave library and application
developers a chance to sync up, resulting in a kernel that was actually useful
to somebody other than a kernel developer. Had Linus, instead, kept only a
development kernel that changed hourly, never stopping long enough to get the
drivers updated to the structures in the core, then the fragmentation in users
of that kernel would have had a serious negative impact on the viability and
popularity of Linux to the computing world at large. Change that is
fragmented, leaving discontinuities, has high entropy and a general lack of
solidity.

Granted, the customized software RedHat ships that results in binary
incompatibility is a slightly different concept than degree of change. But the
end result is similar. A coordinated effort between RedHat and the other
distribution builders to achieve a common framework for advancement would not
hinder development; to the contrary, the common ground achieved from such a
coordinated effort would make Linux that much stronger.

Kris

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sat Oct 07 2000 - 21:00:09 EST