My name is Robert Cox , I'm a Linux newbie. I've got some windows experience. I've built a few computers, I've installed a few Linux distributions and many windows versions.
My main goal is to learn how linux works.
This is to be the documentation of my efforts to build Linux from Scratch.
As I am a newbie, without a technical background. I feel that my point of view will be from a new " angle" the dummy "angle" LOL. Therefore I feel that my input would be very useful to any newbie attempting this endevor.
First and foremost , I will write this blog as I read the information given and document all my problems and errors that I face and how they were resolved.
Let me just say that I believe my biggest test will be in learning to understand the bash shell and the language it uses and reading and interpreting the errors that are displayed during configuring, compiling and installing the different programs in LFS.
Next blog - Prerequisites, Since I'm a newbie I'm not even ready to start yet I have to drop back and get myself ready to go.
I recently read about Debian changing from the GNU C Library (GLIBC) library to the new Embedded GLIBC (EGLIBC) library. This may be the beginning of a sweeping change similar to the GCC vs. EGCS or XFree86 vs. Xorg changes in the past.
The source of the change is the controversial nature of the lead maintainer, but the story is as old as FOSS itself. The ability to fork a project exists to protect the users of software from having their rights hijacked by the developers. This is one of the most important advantages of FOSS over most other development philosophies. The user should never have to beg for bugs to be fixed, especially when there are large groups of users doing the begging.
I do wonder how Red Hat will handle this. I know that Red Hat is still considered the most commercially viable Linux, but one of their employee (or at least someone with a redhat.com email address) has created enough problems that a large, generally conservative and GNU-friendly project like Debian is willing to risk a fork of a core GNU library.
I have played with Linux for a very long time now. I think it was about 1996 i started using it fulltime as my home desktop. 2001 i got a job as an admin managing about 400 users on 70 old desktops against various Linux servers.
Now i manage about 600 computers and 1400 users with mostly Linux on the backend and windows on the desktops. This is my wishlist as and admin after having worked with K12ltsp, SUSE, RedHat, Ubuntu desktops, servers and Windows.
This is my personal wishlist:
Profil/policy handling in Linux is really pretty straight forward. What i feel a lack for is more work on Sabayon which from my viewpoint is much better than anything else on the market right now. Simpler use of Sabayon and more work on getting it setup correctly for getting profiles from a remote server would make policy handling in Linux much easier than in other OS.
A better simpler way of sharing files between a linux server and a linux client. Right now all work seems to be on making it easier to connect to a Windows world and nothing at all in making it easier to use Linux. This is a big drawback that makes it much less interesting running Linux desktops.
More work on integrating those stuff with LDAP would go miles for making a Linux desktop very compelling in a bigger network.
Alltogether i feel most companies concentrate on managing Windows boxes from Linux instead of making the combination of Linux servers and desktops compelling. I think thats a big mistake of both RedHat and Novell. Admins like me already have a really tight schedule but often pretty good influence on the spending budget. If im a pure Windows admin and my boss asks me to trial Linux on the desktop i will go bonkers from trying to setup NFS, LDAP, /etc/skel and whatnot. I can make this happen by myself but im very sure most admins cant and dont want to either.
Some commercial products exists but the ones i have tried has been buggy or only supports one single Linux distribution etc.
Dont know if i make any sense whatsoever here but there you got it. Its a pretty short list compared to the one i have for the Windows boxes i manage, that list is a mile long.
Well actually he's not that old, 43 to be precise, but he is in a very poor situation currently. Here's a little background info. Gary hacked into the US governments computers in April 2001, including the US Navy, Army, Air Force, Department of Defence, and NASA. He wasn't caught until September 2001, after the 9/11 incident when the computers went down causing the government to suspect a terrorist attack. Gary even went so far as to write on one of the PC's:
US foreign policy is akin to government-sponsored terrorism these days... It was not a mistake that there was a huge security stand-down on September 11 last year... I am SOLO. I will continue to disrupt at the highest levels.
Now he is facing a possible 70 years in US prison. He has the support of 80 MP's, as well as musicians such as David Gilmour (Pink Floyd), The Rolling Stones, U2 and David Bowie all vouching for him not to be extradited to the US. The US claims he cost them approximately $700 000(474 000GBP, 836 000CDN) to find the culprit. Now that you have the background of this...
In my honest opinion, I think that Gary should be charged, however not in the way the US wants. Should he be tried by the US? Yes, but it should be in the UK, where the crime was commited and his homeland. There's no way the US would extradite one of their own to Japan if Japan layed chargeson an American. So, what do you think. Should he be extradited? Charged in the UK? All charges dropped?
Quite often distros, kernels, drivers and applications are defined as "unstable". Yet, in most cases when using technology tagged as "unstable" it works fine and with no reason for concern whatsoever.
3 basic reasons for the use of "unstable"?
What's the definition of "unstable"?
- It's indeed not yet suitable for production environments
- It's a label used by the project to idemnify themselves
- It's used to demote the value and importance of a project
- Meltdown and complete reinstallation of your system?
- Loosing everything on your system (Aka "eats your hamster)?
- X crashes and you have to restart?
- Unable to log in using X and GUI/DE?
- The Desktop Environment freezes?
- Application crashes and you loose unsaved work/settings?
Are there "unstable" mainstream distros?
The mainstream and widely used distros have different focuses. Fedora expresses themselves as experimental, Ubuntu are for everyone, Debian is rock solid and Arch are first with the latest.
At present I'm mainly using Arch with Kernel 2.6.29.x, Ext4, Xorg 1.6, Nvidia 180.44, KDE 4.2.2 (built for Qt 4.4.x and compiled with Qt 4.5) with Qt 4.5.1. I've just added Qt Creator 1.1.
I would believe it's fair to state that this is a rather "early adaptor" setup. Yet I do not have trouble with it. It doesn't eat my hamster, X doesn't crash, it's fast, lean and runs well.
I have one issue though:
Not all plasmoid that are developed for KDE 4.1 and Qt 4.4.x.works fine with the KDE 4.2.2 / Qt 4.5 and Qt 4.5.1 combo. Does that mean "unstable"? I don't think so. It's simply me - using packages that are labeled experimental and testing. I just have to be a bit more careful when selecting plasmoids, that's all. But "unstable"? Don't think so...
It runs indeed well. No issues has been experienced so far.
OpenSuse 11 and 11.1:
I installed those and used them extensivly from day 1, without having any difficulties.
Used it with Ubuntu 9.04 beta without any difficulties.
Used it as main DE since 4.0.85 (KDE 4.1 Alpha/Beta). No trouble since KDE 4.0.9x.
Where are the unstable distros and desktop environments?
My experience is that the term "unstable" is somewhat abused within the context of "desktop distros". Debian, CentOS desktop and SLES 10/11 must indeed be very very good to be classified as more stable than a standard setup of a standard, mainstream distribution.
That's my opinion anyway ;)
Im no coder by any means possible and that nags me from time to time. In my work as a system admin i often get tremendous help from various open source tools and applications. I have tried to go into some ptojects and start coding but i frankly just dont got the nack for it no matter how hard i try. My brain just isnt wired the right way for the type of logic involved in coding.
But, that doesnt mean i cant help or that other users cant even if they are terrible at coding. There are numerous projects that can benefit greatly from non-programmers help. Graphics, layout, testing, triaging and much more are things we users can help out with.
I have found my thing now, translating. Even if its pretty tedius, boring and repetitive i still find it challenging. Best of all for me is if someone have use from my work because if someone can have use from one of my hobbies i do for fun, then its a double reward.
Well, things look to be shaping up quite nicely around here. Good articles (such as Brian Proffitt's Linux is Everywhere), slick design, and decent member services (like this blog).
The DistributionsCentral section, though, could be better. The news and blog feeds are good for those of us who already know about the distribution scene, but there's not much there for someone who is just coming into the world of Linux and wants to know what distro to use. My solution: let's find out what distributions the members of Linux.com use, and why, and include this info on the entries in the distro listing. I've put this idea up on ideaforge here.
Similarly, the directory is a good idea but it reminds me a little too much of the old web directories that you used to see before google came along: lots of categories and items, but few cues to let you make sense of it all. Let's make this thing more social, say something along the lines of the idea proposed by voxel.
A while back, my father (knowing I'm interested in these things) sent me a link to this article, in which the author supplies a critique of the windows 7 user interface. Here's my response, which I thought I'd share:
Good article, the author raises some interesting points. I definitely agree with him where he says that interface design seems to be stuck in the '80s. One thing I disagree with, though, is his optimism about Microsoft's ability to resolve these issues.
The problem is, you have a small group of 'specialists' trying to guess what people want, and come up with a single solution that is good for all -- power users and casual users, and perhaps more importantly, those who have invested a lot of time learning the old ways of doing things and those who are willing to experiment with new (and hopefully better) approaches.
Of course, the solution to this would be to have a selection of different interfaces designed for different types of user, but I don't see Microsoft adopting this approach anytime soon. On the other hand, this is exactly how the Linux world works, and is actually one of the main reasons I'm into Linux. For a start, you have the two main desktop environments, KDE and GNOME, both of which offer fairly conventional user interfaces, but tailored for slightly different sets of users. Then there are some less common environments, like Mezzo and Etoile, which have their own ideas about the user interface.
The great thing about this system is that it works on the 'survival of the fittest' principle -- the success of a given project depends on how well they serve the purposes of their users. If too many projects exist in a given niche, the weaker ones will die out, and if there is a niche that is not being addressed, projects will spring up to fill it. In fact, this applies across all applications in the free/open source world, not just desktop environments (remember how Firefox supplanted the Mozilla suite?).
Well, currently I just installed Jaunty, and it seems to be okay. I'll start off with the installation.
It was rather straightforward. I used the install option rather than booting to the LiveCD desktop. First thing you'll notice is the installer is full screen. As well, they also fixed the partitioner. It now shows the disk properly, and the more partitions you make it works correctly, rather than going off screen. I chose Ext4 as the default FS, and it's okay, but more on that. It took a grand total of 15 minutes to install, the fastest I've ever seen it yet.
The first boot was fast, as are all the subsequent boots. My 1.6Ghz, 512Mb Ram craptop boots in 30 seconds flat now. The login screen is very nice too, revamped and more modern.
I started by moving my backups onto my drive -- 115GB of video, music, and games. It worked fine. I installed a list of applications that I find standard -- mplayer, smplayer, gcc, build-essential, and of course, armagetron. They plus 300+ other MB of download files installed correctly.
The menus are snappier, gnome notification is much more professional, and everything is generally more responsive. After I added all the repos and did 'sudo apt-get update', then ran hardware drivers, it picked up my broadcom just fine. Everything just works.
However, there is a problem. Ext4. Don't install this if you want a stable computer. Removing things to your trash is safe, but don't be doing anything essential if you're emptying it. Deleting files can hard-lock your pc. Even "<ctl><alt>REISUB" doesn't work. It requires a hard boot. Stick with Ext3, but if you were unlucky enough to install 4, there will be a patch in about a month, so it's not forever.
The update manager went through some changes. Only once a week it will notify you about an update for files, and it will notify you immediatly if it's security related. However, if you close it when it's security related, it comes back right away -- very annoying. This is changeable, but I haven't gotten around to it.
Overall, besides Ext4 and the UpdateManager, I haven't many complaints about 9.04. Installs quick, apps just work, and everything seems to be nice and dandy. I recommend it to anyone. Good job Canonical.