The new Linux kernel development model...

Submitted by Jeremy
on July 22, 2004 - 8:38am
...is a wonderful idea that will result in a better kernel.
14% (230 votes)
...scares me.
21% (339 votes)
...must be a good idea, though I don't understand how.
12% (188 votes)
...is the beginning of the end.
9% (151 votes)
...is not much different than what we've always had.
10% (161 votes)
...is evolving, and I'll be interested to see where this leads.
34% (563 votes)
Total votes: 1632

sorry i have read the articl

Anonymous
on
July 22, 2004 - 11:46am

sorry
i have read the article
but i dont clearly understand what's the new development model :/
could someone just explain in a few sentences?

until now "new things" (exper

Anonymous
on
July 22, 2004 - 12:07pm

until now "new things" (experimental/unstable/untested code) in the kernel went to the developing trees which end in an odd numer (2.3, 2.5) now andrew and linus thought that their teamwork was nice that there's no need to open a 2.7 tree now and that they'll be moving this "unstable" code into andrew's -mm tree instead of 2.7. That leads to the point, that unstable code is much quicker in the main tree and that distributions will possibly have to revide and clean up/fix the code. it's said that they're doing this anyway..so it wouldn't be a big change. No it's sceptical if this new model will make the stable tree (2.6 currently) less stable what would not be welcome to many people and is appreciated by others (mainly developers i'd say). i think that this will speed up development as things will be unite and not in completely different trees, but we'll have to keep watching the stable tree a bit closer to keep it stable the way it is..time will show..

thanks

Anonymous
on
July 23, 2004 - 4:19am

thanks

let's see

wouter
on
July 22, 2004 - 3:34pm

I feel a bit sceptical, since I appreciate stability more than features or speed (from a kernel). But, let's see what happens... I'm sure some of the core developers know better than I what they are doing.

As long as it doesn't result in many machines crashing, which would be definitely a bad sign for the whole of linux. How will we know which kernels are very stable and can be used for production systems?

stable kernels

Anonymous
on
July 30, 2004 - 7:06am

If you are running a production system stick with 2.4.x series than to try 2.6.x series.

2.6.x series is quite new, it's been less than year since it was released, remember what happen to early 2.4.x releases.

2.6.x *is* stable

Anonymous
on
August 19, 2004 - 9:35am

I've been using Linux since pre-1.0. I was skeptical, but when people say that 2.6.x is the most stable "stable" kernel yet, it is true.

I think this success can be attributed to several things, the following being some of the more important:

  1. Linus's improved working conditions (OSDL, plus his use of bitkeeper and various automated tools).
  2. Linus's work on sparse, the Stanford checker, etc. doing automated code review.
  3. Andrew Morton (and others) getting better at feeding Linus the good bits.
  4. Automated testing (by OSDL and others) doing a good job of regression testing.

aanyhow

Anonymous
on
August 6, 2004 - 2:54am

I like lard !

Why change now? We had an easy to follow scheme...

Anonymous
on
August 16, 2004 - 10:05am

How will we know which kernels are very stable and can be used for production systems?"

Good question. I think the kernel versioning scheme should have stayed the same. Keep progress going in odd numbered series, but keep the even numbered series for bug & security fixes, not ongoing development.

This contradicts the way things have been done, and what we've been able to rely on. This change will complicate things unnecessarily. :-[

~ www.whyaskwhy.org (currently down)

Don't really care

Anonymous
on
July 25, 2004 - 5:27pm

I don't really care what development model they want to use as long as the future kernels keep improving.

some dependencies

Anonymous
on
August 6, 2004 - 3:54am

Excuse me, for my bad english - it's too bad

I don't like some dependencies.
In kernel-2.6.7, I would to built system without support for mouse, but before compilation 'make dep' - write error and compile support for mouse.

In the future, I will see, that my graphic card coludnt be installed because I couldnt installed my sound card properly. (Like MS Windowss)

Thanks for your time Ales ( avelebil@avprog.cz )

will never cease

Anonymous
on
August 16, 2004 - 3:26am

will never cease

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.