Re: [SLUG] Kernel versioning

From: Paul Braman (aeon@tampabay.rr.com)
Date: Wed Oct 31 2001 - 22:56:23 EST


On Wed, 31 Oct 2001, Russell Hires wrote:

> That wasn't helpful. :-( In general I'm questioning the way the
> versioning worked in the past vs how it's working now. Derek answered
> pretty well my question. From that answer I get the idea that Linus
> knew that a larger group of people wouldn't test the kernel out
> because it carried the stigma of being "experimental" or "unstable."
> So he played a semantic trick on people to get more people to put it
> through its paces.

The Linux kernel is just like any other piece of software. When you are
done doing all tweaking and debugging in an experimental sense you have to
"chop it off" and call it stable.

The kernel's first few stable revisions are never really stable like you
might think and that's why we go through so many versions of them right up
front. The same thing happened with 2.2.x when it came out.

In the case of 2.4.x there had been a good long stretch in experimental
mode and it was just time to stop experimenting and start stabilizing.
That's what happened. It's no trick and no one's trying to fool anyone.

Look at Red Hat for a parallel example. Whenever they release a new
version of their software package it's pretty much assumed that you should
really wait until x.2 to upgrade because x.0 will be unstable and x.1 will
be rushed out simply to fix bugs. It won't be until x.2 that you get a
nice system worth installing.

It's all a numbers game but if you understand the rules you can learn to
be happy.

Paul Braman
aeon@tampabay.rr.com



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 16:07:06 EDT