Results 21 to 24 of 24
well, starting from scratch would be quite an achievement but maybe I should tone it down even more from a new os to a new DE. If you would both ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
- 03-30-2012 #21
- Join Date
- Mar 2012
- 04-03-2012 #22
I think if your willing it is a great project and motivator to learn Linux or any other unix/unix-like OS.
Honestly, i am no Linux expert. In fact i am rather shamefuly admitting that last night was the first time i have compiled a customized kernel.
3.2.13 from kernel.org On Scientific Linux Fermi. Its motivated me to want to create another one but aiming for higher performance. I have noticed the Oracle's UEK kernel has better performance, so i want to see if i can take what they have done and make it better still, well at least their configurations.
I would start there, it is honestly really simple. The read-me that comes with the kernel source is all you need. But maybe a list of what the different kernel components do would be useful if you really want to optimize the kernel. Also it can be alot smaller than the default kernel and isn't limited by what Red Hat or Debian implement. For example my customized kernel could read and write to all the different Linux Filesystems (XFS, JFS, Ext2-4, ReiserFS, etc) without the userspace modules. Which as you know red hat isn't capable of by default (except for XFS & Ext2-4) unless you install those modules, which even then is limited for some of them.
I am also looking at creating my own RPM repo. To go with the eventual customized kernel (aimed at high-performance).
If you get my drift... Take an exisitng distro preferably one that you like and feel you could improve or add too, and slowly add your own signature to it. A customized kernel, your own repo. Once you've done that, you have already got the building blocks for a new distro, and can take it further. i.e. making a new Distro logo, renaming the current one.
Last edited by SL6-A1000; 04-03-2012 at 03:16 AM.
- 04-03-2012 #23
I think this is a lofty goal but ultimately I think you need to play around with a lot of Distros and DE's before deciding that a new one is necessary. I mean just to name a small list of what should be tired:
E17 (I use this one and I think it achieves all of your goals)
I know that if the plan is to make a DE from scratch.....E17 has been in development for 12 years and has a whole list of incredibly advanced devs on the team and STILL they call it a pre-final release version. I think what you may want to do first is:
Decide a base distro -- strip it down completely
Decide a base Desktop Environment -- strip it down
Then start going through the code and altering it to suit your needs. You'll see hundreds if not thousands of pages of code to make a DE.
This is exactly how cinammon is being developed by the Mint developers (they don't like Gnome-Shell so they're doing their own remake on it). Again this is a well established group of developers who after a few months of trying to make their own DE from scratch said screw it because it's just too much work (years - plural) for a team of well established devs
Maybe try joining some dev mailing lists and pitching more concrete ideas other than "sleek" and whatever else you're looking for.
Now if you're trying to just package a new set of packages for the base install of a distro, that can be done with a smaller team and requires a lot less work. Bodhi is an example of this but still requires substantial programming skills and a fundamental understanding of DE's, Distros, Package Managers, etc....Bodhi 1.3 & Bodhi 1.4 using E17
Dell Studio 17, Intel Graphics card, 4 gigs of RAM, E17
"The beauty in life can only be found by moving past the materialism which defines human nature and into the higher realm of thought and knowledge"
- 04-03-2012 #24
Slightly OT, but somehow relevant
No, this is not an Aprilīs Fool.
Yes, this is a commercial game about creating a linux distribution. It costs 4$ to startYou must always face the curtain with a bow.