Slashdot Mirror
Debian And The Rise of Linux
There's an article in this month's LinMagAu that asks a question about how the rise of Linux will impact Debian and what that could mean. Good article, especially interesting if you have been a fan of Debian.
← Back to Stories (view on slashdot.org)
I think for Linux to grow more, it needs a reference implementation so that developers and users know that something will work for sure.
I think Debian GNU/Linux should be this system for several reasons.
It's non-commercial, meaning SuSe can't complain that the reference system is partial to RedHat or anyone else.
It's conservative, which is very important for reference systems. If you write for Debian 3.0, you know it will be around for awhile. This doesn't mean that RedHat can't extend their distribution to add more recent libraries or programs. It just mean that something written for Debian 3.0 will work in the RedHat system that says it follows 3.0.
Je ne parle pas francais.
This is a self-fullfilling prophecy, and to change this will take quite a major change from the existing Debian (fairly elitist) culture.
Where Debian will shine is not nessicarily as a mainstream distro itself, but as the basis of systems that are more widely used, such as Xandros and Knoppix. Is this a bad thing?
It does run the risk that Debian-as-distro/brand become marginalised, but all that needs to happen for the Debian project to stay healthy is that Debian-as-underlying-system is widespread.
This said, my Ideal World(tm) is every man and his dog running Deb... ;)
"To any truly impartial person, it would be obvious that I am right."
... to much people around comming from Windows.
Debian was always about doing "The Right thing", about not only making things work, but make them work like they should work.
But you cannot build a good distribution on software getting worse and worse. Think about more and more software unable to do basic things, because people did not thought about them as they are not feasable with one human before one computer. Because people grew up with windows and do not even know how it could work.
On good example is konqueror and its identification of file type through filename's suffix. Do you have time to tell 300 users of your computers to rename "download.htm" to "bild.gif" to be able to click on it. (Oh, sorry I forgot, you are using your computer alone...)
Even Debian, which was formerly known to be usable by admins, is now working on abolishing its old working menu system to one build up on KDE's
menus. (Instead that someone would finaly get a menu-method for KDE and the old one.)
It's a shame, the old system capable of creating a menu looking the same under all window-managers (except KDE, because the KDE people do not want to integrate) making life for an admin really easy, is dropped for a thing not nearly capable of it.
(No possibility to specify a menu-hirachy. And the proposed format for icons is png. absurd.)
WARNING : THIS IS NOT A FLAMEBAIT. I LOVE DEBIAN, BUT READ ON ...
I started using Linux with SlackWare when it was the only distro available out there. I used to love them tarballs, but then at the time systems still had manageable sizes, so one really could compile everything in a reasonable time.
Then I had the (mis?)fortune of being hired by a certain Caldera spinoff and was forced to use OpenLinux 1.2. That was my first contact with RPM, and that was a painful contact. Part of my work also involved writing and maintaining specfiles for various cross-platform packages. That's when I learned that (1) RPM was better than tarballs because it had dependencies, (2) RPM dependencies are not powerful enough and (3) RPM isn't backward-compatible. In short, RPM is not good but it's better than nothing.
At that company, I also had the misfortune of meeting a Debian fanatic. Note that I say he's a fanatic of Debian, not that Debian made him a fanatic. Having tried Debian long ago myself, when it wasn't ready for prime-time, and having found it complicated and messy at the time, I was conforted in this idea by the truly detestable way this guy was patronizing everybody who didn't use Debian, and was turned off Debian for another 2 years.
Then, several months ago, it was a sunday afternoon, my local computer shop was closed, and I couldn't find my RH CD to reinstall my box. I though : what the hell, I'm no more stupid than the average Debian user and I have nothing to do, let's try the Debian network-install. Well, I went through a little pain (it's not quite totally polished yet), but I've never looked back. dpkg and apt-get are just a godsend, and I too am now a convert today.
Moral of the story : I avoided using Debian for several years entirely due to the advocacy of one (well, several actually) Debian bigot. You can always say that I should have been more intelligent and I should have made my own opinion, but I never had time and the experience you get from other users do count for me.
In conclusion : what's the biggest good that could happen to Debian ? that other distros' package management got better so Debian bigots wouldn't have such an powerful incentive to behave like asses and disgust other people of Debian before they even try it. Or better still, that the Debian bigots start realizing that they won't win anybody to Debian by being patronizing.
"A door is what a dog is perpetually on the wrong side of" - Ogden Nash
Why I like Debian:
(1) Serious philosophical principles. The only people to say GNU/Linux with a straight face. People concerned with my liberty above all else.
(2) No Prepackaged Experience. I run Fluxbox, Gnome-Terminal, Mozilla, and Konqueror, and have a proper GTK/KDE library environment. It all works the way I want it.
(3) The system state is transactional. Glitz is antithetical to transactionality. Glitz hides transactions. I like transactions.
(4) No waiting forever to compile stuff pointlessly.
#1 is the crucial element. Liberty is paramount.
I don't think Debian is going to collapse soon, But I do agree the installer could use some rethinking. Things I would like to see included: System recovery (using distributed backups over the lan), hardware autodetection, an installation blog - or something like that where you can put your installation remarks/choices, etc. Also, I'm looking for a command that would backup all config files that have been changed, or all files not managed by apt.
Also, if there were a central repository for those installation blogs, developers could easily see where most of the problems arise.. Just some random thoughts..
"It's too bad that stupidity isn't painful." - Anton LaVey
The company i work for uses a good deal of redhat for workstations and high end servers. Even our NT admin converted to redhat and uses vmware for when he needs to run NT. I personally now run debian at work, and am trying to get them to change to deb.
After using redhat for many months here, then changing to debian, ill never go back to RH. It can be a pain to get installed, but once there, its solid. where as on redhat I had lots of dep issues because I was always installing cutting edge crap. I have done the same on debian, but with alot less issues. With in a few weeks ill have the chance to change over our DNS server to debian. And onward from there...
We have seen that living things are too improbable and too beautifully "designed" to have come into existence by chance.
If Linux gets a lot bigger, but Debian doesn't get bigger with it-- so what?
The Debian developers seem to be happy to work on Debian for their own use and for the use of the people who use it now. As long as that audience doesn't shrink too much-- and I doubt it will, for though many slashdot posters love to scoff at this, there are some people who use Debian for philosophical and other reasons-- then the same number of people will continue to use Debian.
Yeah, I agree that Debian needs to move forward and needs to make sure it stays as close to the "cutting edge" as possible. But I don't understand why other Linux distributions exploding into extreme popularity among people not currently using Linux at all must detract from Debian. That sort of "must be the market leader to survive" mentality may work for commerical entities (be they open or closed source companies), but Debian isn't one such beast.
Indeed, I suspect what will happen is that the "mainstream" distros will become more attached to proprietary offerings. Red Hat's made amazing contributions to the open source community, but if their users are demanding crossover office sorts of things bundled with Microsoft Office, and M$ agrees to licence that, I'd be surprised if Red Hat didn't go for it. There will be those who will stick with Debian for philosophical reasons-- and so long as there are enough of them to provide a core of Debian maintainers, why not? It doesn't hurt anybody else.
That's the great thing about free software. Anybody who wants to do their own thing can do their own thing, without being beholden to what somebody else is doing, and without requiring anybody else to be beholden to them.
-Rob
More to the point: Debian is already marginalized to a certain extent. In the semiconductor industry, if a simulation or regression tool runs on linux, it runs on RedHat linux. A specific version of RedHat linux.
It is one of the first questions that technical support will ask: what version of linux is the tool running on? And if you answer incorrectly, you get a free trip to the sorry but that is not a supported configuration hang up. I am responsible for about a hundred linux boxes and none of them are Debian, for precisely this reason.
The real question is: so what? If the Debian developers are really as keen as everyone says they are, then it really doesn't matter -- they will keep coming up with technical innovations which will get tried, proven, and then absorbed into "more popular" distributions. Let Debian users be on the cutting edge, while those of us with real work to do can use the distilled and canned solutions to get on with our lives.
you should read everything on the internet as if it had "but I'm probably talking out of my ass" appended to it.
The article claims that since Debian's technical advantages can (and to some extent have been) be "borrowed" by other distributions, and since Joe User doesn't care about the policy advantages of Debian, then Debian is doomed to be marginalized as the Linux market grows with unprecidented numbers of Joe Users. I strongly disagree.
Debian has always had a strong following with Systems Administrators who want a strong, stable, supportable platform for their GNU/Linux based services that can be centrally administered without waisting a lot of time. The same forces will make Debian significant as a corporate desktop. This is a huge market, and while Joe User might be on some of those computers, he's not the one making the decision.
Red Hat wins its share of this market through marketing, Debian wins its share through precisely the same policy superiority that the author discounts. Sure, Joe User doesn't understand the policy advantages, but Joe User doesn't play in this field. Sure, Red Hat and other corporate marketted distros will mean Debian will probably never even get a majority share of this field, as long as there are systems people who are allowed to make systems decisions, Debian will be a player here.
The other two markets are Small/Home Businesses, and Home Users. These are the fields Joe User plays. And no, he's not necessarily likely to gravitate towards Debian (actually, from my experience he is, but all my evidence is anecdotal, and it's irrelevant for my point). What the author misses is a key differentiation distros that borrow from Debian.
Some distros, like the example of Red Hat borrowing apt-rpm/apt-cacher, are alien distros borrowing a tool that was developed by Debian. While they probably will contribute to development of the tool, these don't do much for Debian as a whole.
Other distros are derivative of Debian. They put their own installation and look and feel, do their own marketing and often usability testing. They might not even mention their relation to Debian, but, at their core, they're Debian, and developers developing for these Distros are directly helping Debian development. Some significant distros in this category are: LindowsOS, Progeny and Libranet. They're not Red Hat, but they're growing, and growing strong.
I feel Debian's chances of being marginalized are slim.
----
Open mind, insert foot.
Why care so much about Joe User?
Let RedHat, ALT Linux and other commercial firms
care about them. They would get their revenues
and give their contribution to OpenSource world,
including Debian.
Users switch to Debian not from Windows (or complete
computer illiteracy), but rather from other Linux
distro's.
Personally I switched to Debian from RH (four or five years
ago) when I found out, that when I need some piece
of software which is not included in my distro,
I routinely go to ftp.debian.org and grab orig.tar.gz from there.
There should be at least one distro in the world,
which cares about clever people, not stupid ones.
Debian perfectly fill that niche. It is created
by clever people and targetted to clever people.
With apt-get dist-upgrade who need installer
at all, once he learned how dump/restore work?
And for first time in the life you better
to call some more experiencd friend.