Slashdot Mirror

Why don't they just make their own distro using something such as Linux From Scratch? THey have enough money to create an entire Linux division in their corporation.

Real men use http://www.linuxfromscratch.org/

After all, X.org is an essential part to all big name 'distros'.

X is an important and widely used piece of free software, but it is not as essential as GNU and Linux are. Many people run GNU/Linux without X. Every distro worth naming has an X package, but its installation in many cases is optional. Debian only installs X if asked. Ubuntu provides a nice convenient way for an server admin to opt out of X and graphical programs at the start of its installation. Who provides a choice during install of C library, shell, and coreutils? Not even the Linux From Scratch people raise this possibility.

Go ahead and refer to the operating system as GNU/Linux/X if you really value X's existence that much. That would be a lot better than simply calling it Linux, as you would at least be including the name of the most significant component, GNU.

why should the operating system be named after optional and replaceable stuff sitting on top of the operating system. Why name my automobile after the brand of tires it uses?

I do not accept your comparison of the GNU C library, BASH, and GNU coreutils to tires. These components are quite important, and not as interchangeable with alternatives as you may think. Replace them with dietlibc and busybox and will you have a new operating system, as distinct and incompatible as Solaris, FreeBSD, GNU/HURD, and GNU/Linux are to each other.

There is another point worth adding. If the significance of GNU's contribution was the only issue, people like myself would not bother engaging in this argument. There is a second major reason for why we call the system GNU/Linux, I discuss it elsewhere in the thread.

You have absolutely no idea what you're talking about! How on earth did this get modded up!?

There is only the Linux kernel... and no you don't have to develop a driver for multiple versions of Linux! That's nothing short of absolute lies!

If it's one thing I hate it's an anti-Linux zealot that doesn't even know what they're talking about. At least take the time to learn about what you preach against.

A point by point rebuttal of everything you said:

But first let me point out that I've actually written device drivers for both Windows and Linux, I am an open source software author, and I've played parts in writing large applications for big Windows shops. I run and use Linux and Windows on a daily basis... something you have obviously never done.

So here goes...

Windows moves *slower*. When you're writing drivers, slower is demonstrably a good thing.

Windows does not move slower than Linux. The driver API changed significantly with NT, then with 2000. It's been largely stable since then, but there are still continuous changes. It's a complete misnomer to suggest otherwise.

By the same token, the Linux API isn't as unstable as "keeping the API open" suggests. There are many drivers available in the kernel that have been there for... a LONG time. Most of them were ported to 2.6 with no trouble at all.

As a person who has written device drivers I can tell you that writing and maintaining a Linux driver is significantly easier. The docs and community support is all there, and everything makes sense. It's pretty much the opposite when it comes to Windows driver development.

Trying to maintain a driver for Linux would require constant attention.

Simply not true. And the beautiful part about Linux is that even if a driver does need updating, there's a significant chance that if the driver is used by enough people, some person will just fix it on their own. But let me just reiterate that this is completely untrue in most cases. At least not any more than it's true for Windows.

Plus, Linus' kernel isn't stable. He just waves his hand in the air and announces that 'the distros' will have to make Linux actually work. That means that now we have Red Hat's kernel, Suse's kernel, Mandrake's kernel, Debian's kernel...

I'm sighing right now. Why... where do these idiots come from? And how do they get modded up!? Linux is a kernel. It's not an operating system. Nor is Red Hat, Ubuntu, Gentoo, etc... they're distributions of an OS that uses Linux as its kernel.

I've built Linux From Scratch a few times, so I'm painfully / joyfully aware of what this actually means. You're obviously confused about this point so I'll explain it to you.

Basically no matter what distro of Linux you use... you are using your own customized version of a Linux based OS. It may not seem like it when you've first installed it, but it's still true. By the time you get to know what you're doing your OS is probably inherently different than even some other person using the same base distro. You've installed different packages, maybe compiled your own apps and installed them wherever you feel like it. Customized start up scripts, etc.

Whether or not you see that as a benefit is up to you. But let me tell it is a great benefit, and that's what makes Linux so great! That's why there are so many flavors (and no there's not just 5, there are literally hundreds). Choice is what makes it so great.

Imagine a world with 5 automobiles that were supposed to fit everyone.

Anyway... getting back to the point. So you've got all these infinite numbers and possibilities of Linux based OSes out there. Driver hell? I don't think so. This doesn't mean the kernel is any different and it doesn't mean writing a device driver for Linux has to be re-done for every OS, distro, or any other such nonsense.

it means any commercial entity has to develop separate driver

In the open source world, "suggestions" come in the form of "source code"

Note that there is now support for other hashes in some form (either patches or built-in) for linux, BSD and solaris. Just as $1$ means "md5", there is $2$ or $2a$ for "blowfish", and a few other less common ones.

e.g. http://www.linuxfromscratch.org/hints/downloads/fi les/blowfish-passwords.txt ,
http://web.cps.msu.edu/cgi-bin/man2html?crypt.conf ?4?/usr/man.
(I think the linux link there is thoroughly obsolete and the functionality has started to appear "upstream", but I can't find the up-to-date reference).

It's mostly doable already. linuxfromscratch lists one possible way. I suppose one big problem for other distros is that a generic package manager which can handle any package will require root permissions, because some programs have a legitimate reason to be installed setuid root. For obvious reasons, non-root users can't install setuid root programs.

In fact I'd be amazed if you can even build glibc from the sources yourself, its build system is notiorusly arcane and undocumented - often requiring CVS builds of binutils/gcc, at some points it's even required custom patches not available to the public.


LFS manages to compile glibc just fine.

I switched all our servers from redhat to slackware about three years ago. Having tried many distros I found slackware was the most stripped down yet contained all the basic stuff you need for a server environment (sendmail, IMAP, POP, apache, etc.). The main reason I switched was library conflicts. Most other distros come with too much stuff and as you install more stuff on top you inevitably run into probelms. Slackware is nice coz it's bare bones.

I've had way more problems with config programs and package managers on other distros than I've had learning to rebuild the kernel, compile software, and edit config files on slackware. Plus I've learned a hell of a lot along the way.

I also run slackware on a laptop as my sole OS. If you're a slackware fan like me you might also want to check out Linux from scratch

Happy hacking:)

As an 'expert' system administrator (albeit unpaid) I have four servers. One is running Microsoft Windows Small Business Server 2003, one is running Microsoft Window Server 2003, one is running Ubuntu Linux 5.10 (Server), and the other is running Apple OS X Server (10.4).

I can tell you now that when I first started my company, although I was a major advocate of Linux, I soon found that I did not have the time to maintain a then Gentoo or custom LFS distribution, Debian was far too heavy to pick up, and Slackware felt a little dated. So I took a look at Microsoft Windows Small Business Server 2003, liked what I saw, and bought a Dell PowerEdge 400SC with an OEM install.

At first Small Business Server was a breath of fresh air. It was easy to maintain, with a full complement of features, having been bundled with Microsoft Exchange, Microsoft SQL Server, and Window Sharepoint Services. I actually enjoyed - yes, enjoyed - using it.

Until backup stated to fail. Until my tape drive disappeared. Until the sharepoint website database got corrupted. Until exchange monitoring failed. Until the POP connector started to thrash the CPU. Until the Windows Update website failed to check for updates.

These things happened. I'm not saying that they wouldn't happed with another system, but that is not the point, since they happened to me, and that caused me grief, and time, and money to resolve. I ended up trying to build a new system based on Microsoft Windows Server 2003, since I already had Microsoft specific data (files and tables), but this proved even more difficult to maintain.

I struggled for eighteen months, and then decided to build an Ubuntu 5.10 server. I use Ubuntu on one of my laptop, and had gently learnt the apt- way, and liked it. I set up a server with similar features to the Small Business Server, using Postfix, MySQL, and Plone, and even went some ways to transferring my sharepoint data. It works. It hasn't failed yet.

I bet the guys who took part in the survey only set up a server, installed some applications, and patched it. I bet they didn't try running a business for 18-months, just to see what it was really like.

I must say that we recently purchased an Apple PowerMac, and were so impressed we are now looking at completely switching, hence the OS X Server. It is a dream to install and configure, but we are going to run it for several months until we are satisfied that it can do the job.

If I wanted to build a Linux system that would have the same functionality as this, how much effort would go into that sort of thing?

For one computer, not as much as you'd think. http://www.linuxfromscratch.org/
Getting it to work on the thousands of variations out there takes a little more effort.

I've found that the best way is to solve some particular problem.

Very true indeed. However, to fully understand linux and the other unices, it is an invaluable lesson to build your own system. Yes, that's right, the tedious exhausting Linux From Scratch method (or even the FreeBSD equivalent, for the daring). It need not be accomplished all at once; one or two spare hours at a time is quite enough. Even if you do not stick with your eventual creation, you will have a newfound grasp and appreciation of the function of the system as a whole - from the hardware and kernel level on up to the gui. Everyone from novice to system administrators can benefit, and no matter whether its a crash course or review.

I'm relatively new to unix-OS's. I tried almost a dozen Linux distros and ended up getting annoyed with the sheer amount of 'choice'. (Off on a tangent I reckon Windows Vista with umpteen versions could suffer in the same manner).

I'm used to picking up an OS CD, installing it and having a base to work with that's minimal. No matter what distro of Linux I try I feel like i'm using somebody elses machine, not mine. With FreeBSD I started from scratch installed X, a desktop environment (XFCE) and all my apps. I felt it was a good comprimise between a a Linux distro and Linux from Scratch. The FreeBSD handbook guided me through a kernel recompile with no hassle and everything I needed to get going in the first 2 days. To me this was alot less confusing and worthwhile (in terms of actually learning something) than all the Linux distributions I tried. It's a "Here's how you get going with FBSD" experience as opposed to "Here's your Linux machine, enjoy!" one. It wasn't as satisfying going into the plethora of open source software out there without that.

Sure there probably are Linux distro's that have package managers as good ad FBSD's ports tree, and FBSD of course owes alot to Linux, but all the time FreeBSD runs everything I try labelled Linux that I try and all the time the OS gives me no hassle, i'm going to keep using it.

FreeBSD 6.0 is on the way with improved wireless support (apparently) and after the other day's post of FUSE (file system in user space) on Linux I looked it up for FreeBSD..and hey it has FUSE to (i haven't tried it).

I'd say my FreeBSD/Windows usage is about 70/30 atm, and increasing.

I have parts for several 1+ GHZ machines that I am in the process of putting together. They will all run Linux (probably my own flavor of Linux From Scratch)

This leads into my next point: why spend money on a machine that will not provide exactly what you want? Several years ago I decided not to buy retail machines again, opting instead to build any new machine I needed. I put together a P4 2.4 GHZ Intel machine with 512mb Ram, Nvidia GeForce 5600 video with 256mb onboard ram using some existing drives (cdrom, HD) extracted from retired machines - for $550. It had exactly what I wanted and it screams. This motherboard is further upgradable - I can add more and faster RAM, video card, and serial ATA drives when bottlenecks become a problem. If I chose AMD instead, it would have saved even more money.

While doing it yourself might cost more in some respects (time, and sometimes money), if done correctly it extends the useful life of the PC as a front-line workstation or server beyond that of a retail machine, provides build quality you can trust (after all you should trust how you put the machine together), and does not include the Microsoft tax. If you are real frugal (buying used parts instead of the latest technology, for example) you can actually save money and get more machine for the same price.

My goal is a googleplex of processing power in my own beowulf cluster of machines that I have built myself. :)

Linux From Scratch, of course!

Hi Charlie,
1. First of all, grab Knoppix, burn it to a CD, and spend a few days (or a few weeks, whatever you feel comfortable with) playing with it. The base CD doesn't install anything to the hard drive, so you can't harm anything. This is a completely safe, non-intimidating way for you to initially get your feet wet.

2. While you're exploring Knoppix, there are a few things to read which will really help you. This will give you a very good introduction to Linux, in terms of a little history of the system, how to begin using it, and how some basic things work. Here is another in-depth document about using Linux, leading on from the previous one.

3. Once you've gone through those two, (take as much time as you need) this, written by the same man as the introduction, will introduce you to the Bash shell, the textual command interpreter where as a sysadmin in particular you'll likely be spending a lot of your time. This will ease you into scripting in what I think will be a very non-intimidating way. You will be able to try out all of these exercises with the Knoppix CD, and again, because the CD doesn't install anything to the hard drive, you needn't worry about destroying your existing system's contents while you learn. This is another book on Bash scripting which to a degree follows on from that one, and will go into somewhat greater depth. Both of these should lead to you feeling very comfortable writing shell scripts and moving around to a degree on the system.

4. Here is where we get to some meat. This document goes into compiling and installing generic Linux/UNIX software, and offers some basic applications and examples. Once you've gone through this, coupled with the material above, you should now have sufficient understanding to be able to compile and install at least a basic application yourself.

5. The Pocket Linux Guide will take you step by step through the process of learning to make a small, bootable Linux system on two floppy disks. Although compiling a basic custom kernel is part of this process, the Guide contains a link to another document which explains very clearly how to do this, and given the background you will have received from the previous documents, this should not be difficult.

6. Once you have completed the Pocket Linux Guide, you will then be ready to proceed to this site, which is the homepage of the Linux From Scratch Project. Here you will be able to read an HTML-formatted book which will give you the necessary information to successfully build an entire base Linux system of your own, and a more pure boot CD than Knoppix to initially build it upon. The Linux From Scratch Project also has a sequel book, Beyond Linux From Scratch, which describes how to install, among other things, a full graphical user interface with the X Windows system.

7. After you have completed all of this, although it is not crucial, I thoroughly recommend reading this book during idle moments. (It's still a good mealtime accompaniment for me) It will give you a detailed knowledge of the history and philosophy behind the UNIX operating system in general, which I am sure you will find enormously useful.

(Slashdot flamers, start your engines. I'm aware I'm likely to get a ton of abuse from corporate droids in particular about how I'm not telling him to find hand-holding tech support/this isn't practical advice, blah blah blah. For anyone considering res

Built a Hylafax http://hylafax.org/ system on top of the latest v6.1, LFS http://linuxfromscratch.org./
Details:
3GHz Intel Pentium 4 Processor, 1Gb RAM
11,878.40 BogoMIPS Total, 250Gb Hard Drive
    GCC 3.4.3
    Samba 3.0.14a
    HylaFAX 4.2.1

Gotta say it's way ahead of expectations.
I won't touch another distro now for my mission critical.
Although, Knoppix, http://www.knoppix.org/ and Ubuntoo, http://www.ubuntulinux.org/ are great "insert CD and run" distros, for workstations.
Working with SlackWare seems effortless also, http://www.slackware.org/.

Was fortunate enough to meet the fine gent who started the LFS project: Gerard Beekmans
Highly recommended support for the project, even if it's just $5 for a beer via donations :->, http://www.linuxfromscratch.org/hints/contribute.h tml or a much needed "hints" writeup, http://www.linuxfromscratch.org/hints/.

Built a Hylafax http://hylafax.org/ system on top of the latest v6.1, LFS http://linuxfromscratch.org./
Details:
3GHz Intel Pentium 4 Processor, 1Gb RAM
11,878.40 BogoMIPS Total, 250Gb Hard Drive
    GCC 3.4.3
    Samba 3.0.14a
    HylaFAX 4.2.1

Gotta say it's way ahead of expectations.
I won't touch another distro now for my mission critical.
Although, Knoppix, http://www.knoppix.org/ and Ubuntoo, http://www.ubuntulinux.org/ are great "insert CD and run" distros, for workstations.
Working with SlackWare seems effortless also, http://www.slackware.org/.

Was fortunate enough to meet the fine gent who started the LFS project: Gerard Beekmans
Highly recommended support for the project, even if it's just $5 for a beer via donations :->, http://www.linuxfromscratch.org/hints/contribute.h tml or a much needed "hints" writeup, http://www.linuxfromscratch.org/hints/.

Built a Hylafax http://hylafax.org/ system on top of the latest v6.1, LFS http://linuxfromscratch.org./
Details:
3GHz Intel Pentium 4 Processor, 1Gb RAM
11,878.40 BogoMIPS Total, 250Gb Hard Drive
    GCC 3.4.3
    Samba 3.0.14a
    HylaFAX 4.2.1

Gotta say it's way ahead of expectations.
I won't touch another distro now for my mission critical.
Although, Knoppix, http://www.knoppix.org/ and Ubuntoo, http://www.ubuntulinux.org/ are great "insert CD and run" distros, for workstations.
Working with SlackWare seems effortless also, http://www.slackware.org/.

Was fortunate enough to meet the fine gent who started the LFS project: Gerard Beekmans
Highly recommended support for the project, even if it's just $5 for a beer via donations :->, http://www.linuxfromscratch.org/hints/contribute.h tml or a much needed "hints" writeup, http://www.linuxfromscratch.org/hints/.

http://www.linuxfromscratch.org/

You could build and maintain your own distro - then you control upgrading it so you don't run into these problems.

At least with Linux you have options. When using Windows you are stuck with whatever Redmond thinks you should have.

The BLFS book has excellent sets of instructions for compiling software from scratch, particularly MPlayer. works perfectly for me, and I only really have problems with badly encoded quicktime files, check it out and let me know if you have any success.

The BLFS book has excellent sets of instructions for compiling software from scratch, particularly MPlayer. works perfectly for me, and I only really have problems with badly encoded quicktime files, check it out and let me know if you have any success.

That's exactly what I was thinking. Me thinks he should go read Linux From Scratch, then follow up with a few books on Linux Design and Programming.

>It's so bad that we have to depend on big commercial
>vendors to put together reliable Linux
>configurations for us.

No, you don't have to.

Think Geek: (Link already on Slashdot)
Amazon
eBay
Froogle
Graigs list
Koamart: For ramen.
Bevrages direct for caffine.
Ramen Depot also for raman but I prefer Koamart
Cheap Bytes for Linux related stuff (books, Linux distros, The Linux from Scratch book. Everyone must buy the LFS book. Forget Debian LFS is god
Easy Linux CDs Primarly linux trainning matereal.

Thanks Slackware.

Slackware Linux 10.1 & Slackware Essentials, 2nd $60
Got Slack? T-Shirt - Extra Large $16

Support for Linux, excellent educational material and the most efficient way to run a server, $priceless$

Its a bit expensive, compared to http://linuxfromscratch.org/
but worthy nonetheless.

I'll join my voice to the ones praising Linux From Scratch. It's an amazing resource for learning how a Linux system is built.

We used it as a reference when we built the first full version of GoboLinux -- essentially following the steps of the book and adding our modifications (configure and makefile flags) to build the new directory structure, to make our "/usr"-less distro. :)

To this day, I refer to their build instructions every now and then. They also contain a good collection of security patches, so if you're into compiling your packages by hand, drop by at their site and see if they suggest any additional patches. LFS covers the basic system and Beyond LFS covers the additional stuff (KDE, GNOME, etc.).

I'll join my voice to the ones praising Linux From Scratch. It's an amazing resource for learning how a Linux system is built.

We used it as a reference when we built the first full version of GoboLinux -- essentially following the steps of the book and adding our modifications (configure and makefile flags) to build the new directory structure, to make our "/usr"-less distro. :)

To this day, I refer to their build instructions every now and then. They also contain a good collection of security patches, so if you're into compiling your packages by hand, drop by at their site and see if they suggest any additional patches. LFS covers the basic system and Beyond LFS covers the additional stuff (KDE, GNOME, etc.).

lived from a LFS distro: http://lfs.osuosl.org/hints/downloads/files/2.6-ud ev-nptl-bootcd.txt package management: http://linuxfromscratch.org/~gerard/log-install http://linuxfromscratch.org/~gerard/pkgdel and http://lfs.osuosl.org/hints/downloads/files/rpm.tx t reinstalling LFS (for toolchain updates): http://lfs.osuosl.org/alfs

lived from a LFS distro: http://lfs.osuosl.org/hints/downloads/files/2.6-ud ev-nptl-bootcd.txt package management: http://linuxfromscratch.org/~gerard/log-install http://linuxfromscratch.org/~gerard/pkgdel and http://lfs.osuosl.org/hints/downloads/files/rpm.tx t reinstalling LFS (for toolchain updates): http://lfs.osuosl.org/alfs

Since it wasn't linked in the main thingamjig, you might want to check this out: http://www.linuxfromscratch.org/ Notice, when you look at that, LFS is now at version 6.0. Solid proof that this isn't actually "new" at all. =) I can remember building LFS 3.0 actually on an old pentium 150... just made some scripts for it and left the PC on overnight.

http://www.linuxfromscratch.org/

If you are on a small budget, you can just go with simple scripting. Pick a Debian based distro or an RPM based one (SuSE or RedHat only) and you can script all you need. Enable SSH for every system you deploy, desktop and server. Then you just write a few simple scripts _once_ and you can push down any update you need.

Red Hat has their own update stuff and you can pay them extra and run your own update server on your local network. However, where I work we have found Red Hat to be _way_, _way_ overpriced (I work for a multi-billion fortune 500). We are starting to look toward Novell SuSE for our Linux needs. Novell SuSE is _way_ better priced. If you look at a Red Hat Linux solution and an MS Windows Solution, MS will usually be less expensive! I personally don't know what Red Hat is thinking. However, if you go with Novell SuSE, you will see that Novell SuSE is far less expensive than MS. Also, Novell SuSE has some very nice tech that they got from Ximian. As you pointed out, Ximian, now Novell, Red Carpet, is a very nice corporate update client. That is the whole design of the product. You have one local update server and put the client on all your deployed systems and Novell Redcarpet handles the rest.

With Linux you have tons of options. If you have a really bare-bones budget, I would personally recommend a nice Debian solution. I have been using Ubuntu on my desktops at work and at home and have been very pleased with how easy it is to upgrade with out dependency problems. I originally used Fedora Core, however I would run into repository conflicts often because every Fedora repository out there tried to be "The" repository for Fedora. So you would have 3 or 4 versions of every package and they would all conflict. You won't run into that with a Debian based distro.

If you have a bigger budget, look into Novell SuSE (which is still very cheep) and their Red Carpet client/server to handle updates. If your budget is even bigger, you can look into BigFix. However, I think BigFix is priced more as a bigger corporate product, though for our budget, BigFix was still priced nicely per/client.

As I said, you have _tons_ of options with a GNU/Linux deployment. Build yourself a seperate subnet and spend a few days testing to see what level of support you want. Obviously, the less support you or your staff want to do, the more you will pay for your solution. You could spend 10's of thousands if not 100's of thousands (or millions like us) for a complete MS software "assurance" package or you can go very lowlevel and build your own GNU/Linux system like Linux From Scratch (which was very fun for a personal project but _way_ too much work for a professional solution for more than 5 systems).

I persoanlly think your best bet is a hybrid system of Linux and MS Windows. As I said, get a test lab/network. Then use the right tool for the right job. Try to build a lab that is all or almost all Linux servers with mostly MS Windows XP desktops. On your MS Windows desktops try to use OSS software. For example, deploy Firefox and OOo.org. Maybe for some more tech users you could even get some Linux desktops in that mix. For your development needs, use OSS tech such as Tomcat or PHP.

Honestly, I would personally love to be in your position. It sounds like you have the ability to use the "right tool for the right job" without all the PHB crap or extreme OS bias. Where I work we have 140,000 employees and changing technology is like the changing of the North pole ; )

But *true* elitists use linux from scratch, anyway.

It is possible, as root, to load code at runtime into the kernel that is totally invisible to anyone poking at the computer.

Under linux, with kernel modules, this is possible. Even without kernel modules, it's possible to do it upon the next reboot (in Linux, under any distro I know of that doesn't run off a CD or other non-writeable filesystem).

I guess my lack of attention to this detail comes from the fact that I learned unix on FreeBSD, and in FreeBSD you can set the immutable flag on the kernel so that no one, including root, has permission to modify it. In order to remove the immutable flag, you have to be running in single user mode. It was actually just today that I came to the full realization that most Linux distros don't do this, and as far as I can tell it isn't even possible to do it on Linux.

My SOP for dealing with weirdness is to log into a console and run top as root. If I always ran as root, the spykit could be completely invisible. If I typically ran as user, then I'll see something weird.

OK, but you know what you're doing. The average Joe isn't going to. And if you've already noticed the weirdness, I would assume you're going to continue exploring further and further until you eventually find the culprit, even if that means booting from a CD and reinstalling everything.

They can't change root's LD_LIBRARY_PATH. That's the whole point of having multiple users.

It just seems to me that as you increase the number of things that you need root access to do, you also increase the ease with which an attacker can get a user to run a trojan. After all, if one needs root to install most programs, then a user isn't going to think twice when running an installer for a trojan which tells them to use sudo or su or whatever. And then, when it comes down to you and me, well we always check the MD5sum of the program before we compile and install it, whether we're root or not, right?

In my opinion, in order to properly protect an end-user from himself you need a whole paradigm shift. Firefox shouldn't run as root, but it shouldn't run as "anthony" either. Firefox should run as "firefox". And it should be installed as "firefox". Now there's something which can actually protect your data when you accidently click on a link to a website which exploits a hole in firefox. On my server system I've set up these "package users", although perhaps ironically, my only login to the system is root (it makes sense to me, because I only log in to the system to perform maintenance on it).

Yes, I did it to a system built off of Linux From Scratch and was almost done with the build process about ready to reboot and load the new system for the first time... 15 hours of work down the drain.

I wrote "Unix-style". Built my own Linux system from source ages ago a la Linux From Scratch, waiting for next releases of NetBSD and FreeBSD to take these for a spin, and writing these comments from Gentoo Linux. Does that count?

As for your partitioning problems, I didn't have that problem with Debian (I used it on a PC as well), so there may be something about your drive geometry setting in your BIOS that Debian doesn't like (is it "LBA"?); maybe you need to specify the drive geometry on the Kernel command line (I forget the syntax). Been a long time since I had to deal with that stuff.

"As for freetype and fontconfig -- are you fucking cracked? Those are part of X.org."

They're also dependencies for Pango:

http://www.linuxfromscratch.org/blfs/view/stable/x /pango.html

Though I agree that you can get better support for something well-tested like a commercial distro, there is something to be said for building a distro from scratch, if your goal is to learn how the system works.

Personally, I think everybody should build LFS at least once (at least everybody who wants to learn how linux works anyway). Gentoo makes it too easy, you don't learn nothin' ;)

I believe what you mean to say was: here

http://www.linuxfromscratch.org/

Personally I would like them to start with Linux From Scratch. After all, aren't they here to learn about computers? What could be more educational than "building" your own OS? In fact, a short course in basic electronics should be included also. It would help to make the whole binary thing a bit more understandable.

You should try using Linux from scratch. You can't believe how comfortable 128M of RAM can feel. :)

After the initial setup, the admin of this system wouldn't compile a line of code. The package building box would do all of that for him or her. The beauty of this system is that it's 90 - 99% automated, but not dependent on RH/Novel/Debian fuckups or incompatibilities that might slip down the pipe. Any fuckups are the responsibility of the admin. They have no scape goat. They have to be on the ball and perform 100% or they will be fired.

Debian, Novell/SUSE, and Red Hat don't have incompatibilities within their own configured setups, and APT (for Debian and Red Hat) and YUM handle the dependencies.

What's more, all of these are well documented, used by multiple distributions, and are supported by teams known for their conservative operations with stringent QA.

Gentoo is known for being bleeding edge and not getting the dependencies right - and is not that well known even in Linux circles. What are the chances the next Linux admin will know Gentoo? Close to zero.

First, the additional work would be in designing and implementing the system. Otherwise he'll probably have the same workload as any other admin. Also, he won't be using the standard Gentoo package set. He'll be using packages custom built on a compilation machine for the specific needs of the site, then deployed based on machine group need (ie, render farm, workstation, mail/DNS/samba server). What would he get? He would have an intimiate working knowledge of the entire system that he designed and implemented himself, allowing for quicker response, shorter trouble shooting times, and increased job security. Generally be a more efficient worker overall because he knows more than the average sysadmin about his system.

And it then becomes his system - not a well-known, plain environment with a toll-free number for support.

If you really want to get a sys admin learning about his system, make him build his personal workstation using the Core Linux Distribution or Linux From Scratch. But if an admin did that on company time, he'd probably be in trouble.

A Linux admin should know about Linux before he starts work.

Yeah, well, you gotta do something to set yourself apart from the hordes of redhat/debian/suse zombies out there with no real job skills other than ability to walk from desk to snack machine 10+ times a day, and press buttons. This is by no means a solution for your standard corporate entity, it's more a flexible, inhouse system for a company that needs it.

...hordes of redhat/debian/suse zombies... ??

Please! I've likely used many more distributions (on more platforms) than you have (but no, I won't get into a competition...) - but the point is that all three of those distributions (Red Hat, Debian, SUSE) have a much stronger QA than anything that Gentoo has, and are much more friendly to a conservative corporate environment than Gentoo is.

What's more, all three have certification programs; Gentoo has none. All three have corporate sponsors. All three have stringent QA. All three have support programs available. All three are time-tested, and extensively documented in and out of the project core.

Gentoo is none of these - and still doesn't have a decent install program.

Gentoo is mainly a hacker toy currently; I'd recommend any of the three you mentioned (Red Hat, Debian, SUSE) over Gentoo in a corporate environment.

The parent is absolutely right.

And as a competent admin, I choose the distro that don't get in the way, that let me do the things my way. By that aspect alone, LFS would be the best, but it a bit exagerated. (I highly recommend to install it once though, if you are interested in better understanding of the system, it's parts and how they work, from boot to the password prompt and applications. I used it at my machine at home for quite some time.)

I choose Slackware. I used to install everything (after the initial instalation from the distro CD) from source, but it got tedious. Now I use swaret to upgrade the security-related packages. The software more importantly used (in my case: postfix, clamav and spamassassin and squirrelmail) are monitored from freshmeat and upgraded manually as I see fit (some from source code). Other software are not upgraded unless needed (if it works, don't mess with it).

The main source of problem, in all the distributions I tried, is the package system. Ugrading (or, in some cases, even installing new packages) can break the system. Of course when installing from source you also have the risk, but things are more under control if you know what you're doing.

The parent is absolutely right.

And as a competent admin, I choose the distro that don't get in the way, that let me do the things my way. By that aspect alone, LFS would be the best, but it a bit exagerated. (I highly recommend to install it once though, if you are interested in better understanding of the system, it's parts and how they work, from boot to the password prompt and applications. I used it at my machine at home for quite some time.)

I choose Slackware. I used to install everything (after the initial instalation from the distro CD) from source, but it got tedious. Now I use swaret to upgrade the security-related packages. The software more importantly used (in my case: postfix, clamav and spamassassin and squirrelmail) are monitored from freshmeat and upgraded manually as I see fit (some from source code). Other software are not upgraded unless needed (if it works, don't mess with it).

The main source of problem, in all the distributions I tried, is the package system. Ugrading (or, in some cases, even installing new packages) can break the system. Of course when installing from source you also have the risk, but things are more under control if you know what you're doing.

Recommendations for a n00b?

Go with FreeBSD, and do so for the following reasons. First is the amount of documentation available. It's very extensive. Second is the package management system. You have available binary packages or you can compile everything from source (what I did). It doesn't take as long as the binary package zealots claim. The tools are easy to use and powerful. Ports acts as an enormous dependency tree, and the binary packages are produced from ports commands--it's quite logical. Next, is the configuration. Many of the ``user-friendly'' Linux distributions get their supposéd ease of use from custom GUI configuration tools that edit both standard and nonstandard text files in hard-to-predict ways. Although ease of use is important for a newbie, it's just-as or more important that you learn how to edit things on a low level, too. Finally, the standard OS in the next years won't be Linux based. It won't be FreeBSD, either, but FreeBSD is closer to it than Linux is, and you'll benefit by getting a head start.

If you insist on using Linux, I recommend staying away from Mandrake. Ubuntu or Fedora would be the best suggestions for a good combination of easy and powerful. Slackware is more powerful and just as easy to use, but only once you've acquired the configuration skills Mandrake abstracts away from you. Debian would be a good suggestion, too.

Lastly, if you do decide to go with Linux, start looking at the Linux From Scratch books. The LFS and BLFS books are the documentation that should come with every open source project. Extensive doesn't touch it. You'll gain more from blindly following the LFS book than you will from using Mandrake for a year.

Xfce is a desktop environment, it's an alternative to GNOME or KDE. Both GNOME and Xfce use GTK+.

If you don't want a desktop environment and just want a window manager. Then blackbox, window maker, or twm are all possibilities. (there are hundreds of window managers for X11, take your pick).

You hate Debian? is there a rational reason for this hatred? Maybe you should just use LFS or Crux.

Have an LFS installation. Have been compiling KDE for every new release since 2.2

Yeah, it takes some time. I just start the compile process before I leave for the office. My compile script shuts down the PC when it's done. By the time I'm back, it's done. Simple. No time wasted. It could take 8 minutes or 8 hours - no matter to me coz I'm not using the machine at the time - no slowdown noticed.