login
Header Space

 
 

Objective evaluation of Lunux Distributions

December 16, 2008 - 3:16pm
Submitted by cushioncritter on December 16, 2008 - 3:16pm.
Linux

I am trying to select a new distro, having used Debian unstable for the last 7 years.

Since the global financial meltdown, I have the following observations of Debian:

1) While kernel.org is now 2.6.27.9, Debian unstable has still no 2.6.27 package (the onetime Debian kernel maintainer Herbert Xu quit sometime back over Debian not adhering to a "one China policy"). Debian derivatives such as Ubuntu do have 2.6.27 packages -- but if the Debian core is rotten, how can anything derived from it prosper?

2) Distrowatch reports xorg-server at 1.4.2 versus a current version of 1.5.3. Perhaps this explains why Google Earth crashes with most video drivers, although Google has not issued any updates for Linux Earth in a very long time.

3) Today I see updated openoffice.org-2.4.1 Debian packages. Why is time wasted on updating this old, buggy version when the new 3.0 version is so much improved, loading faster than MS Office?

4) Many Debian packages are built erroneously, for example package-nox which is intended for use without X-windows (using SDL, directfb, etc.) has been built by the Debian maintainers with many xorg/x11 dependencies. So I end up having to rebuild from source for my embedded system.

I think general criteria for selecting a new distro. should be:

1) Most packages should be very recent to the upstream versions, for example, xorg-server would be at most one version behind the X.org site. In this way, Linux users would be reporting useful bugs, rather than ancient bugs no one cares about anymore. This violates the Redhat/Suse/Ubuntu "support" concept of fixng an "Enterprise" version with badly outdated packages one at a time using outsourced support centers.

2) Linux users should be encouraged to build the latest upstream packages that are important to them from source and learn a little about Makefiles instead of being cripples requiring IDE's like MS Visual tools to write/build any code.

3) Packages should match the hardware exactly, i.e. the entire distro should have 486, 586, p3, p4, k6, k8, ... versions. (Already possible with a totally source-based distro.)

4) Non-technical, emotional extraneous input such as "I just like distro XYX" without technical justification must stop. One must be willing to move on to improved distributions (or create your own).

5) Allowing education institutions to treat Linux as something akin to a skilled trade such as plumbing must stop, and courses for using Linux as an OS rather than to get a low-paying, stressful "SysAdmin" skilled trades job with a MSCE/A+-type Suse/ Redhat certification, etc. must stop.

6) Non-technical input from Windows-level users such as "Distro XYZ has better, more updated drivers" must be corrected rather than infinitely repeated. Linux users must be gently chastised that if they do not understand something, they should not make an ignorant comment.

7) We must ostracize those within the Linux community whose behavior, some of which is outlined above, keeps us down at the 1% adoption level.

8) We must challenge the concept that "Windows has feature XYZ" as a reason why Linux needs it also.

9) Linux users must be more united and work together. Attendance at LUG's falls, users do not assist each other in group promotions of Linux to increase adoption, and tune out "political" messages that monopolies are evil (i.e. 'please remove me from the LUG mailing list, this thread is consuming excessive bandwidth ...), all the time hoping running a few GNU utilities under Windows Vista will someday lead somewhere. Helping to enlarge and spread Linux will lead to more and better jobs instead of trying to build a Linux system only you understand for your customer.

gentoo is (for me) the way

December 17, 2008 - 3:36am
some random hack (not verified)

gentoo is (for me) the way to go.
I have been a gentooist for several years now, with many installations (servers, desktops, laptops), and it is precisely the kind of system I like.

The pros (quick non-exhaustive list):
* Complete control of your system: source code for everything (ok, you can still install binary packages if you want, but you have to do that explicitly), compiled to your specs with no fat.
* Superb Performance: direct consequence of point 1: you compile everything from source, you can tweak it to your needs.
* Excellent package manager: portage gives you everything you want from a package manager and more. think bsd ports on steroids, with excellent control and mixing of stable and unstable code, trivial recompilation, and so on.
* Excellent package database: in general very up-to-date with new packages, and trivial local package creation with portage overlays.
* Extensive documentation to get you started

The cons:
* Not for newbies! But you are not, so this is not an issue for you.
* Installation is in manual mode and can take a while to get the initial system running with all the compilation. There are faster ways to get going, if you are willing to live with the precompiled initial image (you will recompile everything eventually anyway, so why not do it from the start?). Also, nowadays the live cd contains a graphical installer that is supposedly friendlier, but I haven't used it so I can't comment (I always install pure manual, starting with the minimal image)

Is the best... I like too

December 28, 2008 - 11:37am

Is the best...

I like too install it from minimal, is recomended... install it from GUI... idk, i dont like it.

If you are a bit newbie, maybe you will fail 3 or 4 times your instalation.

And... is better if you DONT use genkernel =) Makes a bit speed diference.

There's so much things i

December 21, 2008 - 12:59pm
Anonymous (not verified)

There's so much things i don't understand in your post.

First you ask about a new distro, which is a common request in our community, then you go on with criticizing many things at many different levels in a strange rant about what's wrong in debian/FOSS.

This post very much reminds me of post by new linux users (mostly coming from that other OS), encompassing the FOSS landscape and posting a "OMG! I understand all that's wrong in FOSS development and communities, let me explain itto them." type of rant, which most of the time is, even though mixed with good sentiments, a mark of gross misunderstanding of that same landscape/communities and realities behind them.

All you seem to be saying is "i want the latest versions of software X" (which is perfectly fine). Then again you're using debian, and criticizing them for having unstable packages/bugs in sid (yeah, yeah, they really suck).

Why not just go on with whatever uptodate distro fitting your needs ?

== General Criterias ==

Who are those criterias for exactly ? Are they guidelines for the FOSS community ? I, for one, completely disagrees with most of these, even as a convinced FOSS advocate, user and developper.

Now, i don't know if you refer to debian as "those within the Linux community whose behavior [...] keeps us down at the 1% adoption level.", but realizing debian doesn't fit your needs anymore shouldn't make you turn your back on a community which has achieved so much. Really, much kudos to them. So, it's really unuseful to come here and post such things that "Many (???) Debian packages are built erroneously".

PS: I'd like to apologize for the tone of this post, I didn't want to sound to harsh to you. I've just the same kind of rant so many times. :-) Of course, feel free to explain your point further if you think i missed the point altoghether.

Comment viewing options

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