Debian Is the Most Important Linux

inkscapee writes "Without Debian we are nothing. Debian is the most influential and important Linux, and is unique for being the largest, oldest, 100% non-commercial community-driven distro. '...just under 63% of all distributions now being developed come ultimately from Debian. By comparison, 50 (15%) are based on Fedora or Red Hat, 28 (9%) on Slackware, and 12 (4%) on Gentoo.'"

Comments

Do we need this?

[Again]

Yet another dick measuring contest? Seriously?

Re:Do we need this?

[airfoobar]

Yes.

These numbers mean one of two things. Either devs should:
1) Allocate more resources into developing Debian because it's the most important distro, or
2) Allocate more resources into the rest because Linux may be losing its diversity.

It helps to know where you're going...

Re:Do we need this?

[larry bagina]

If you use linux math, Dell, Compaq, HP, etc all load up Windows with different crapware, so they count as a new distribution.

Re:Do we need this?

[Sulphur]

Yet another dick measuring contest? Seriously?

unique for being the largest, oldest, 100% non-commercial community-driven metric.

Re:Do we need this?

[achyuta]

In a way, it's alright.

It is needed to help spread awareness that there are few distributions like motherships out there - whichever, Debian or Fedora - which we should be respectful of and grateful to for all the fun we have on a derivative product like Ubuntu - which has been in the news for all the wrong reasons of late.

Re:Do we need this?

[jamesh]

Exactly. I tried running Dell Windows and it was crap. I use mostly the kernel from Bill's tree with a few HP and IBM patches applied to support the newer hardware as required, but have stuck with the HP window manager (the Microsoft released one is rubbish).

Re:Do we need this?

[armanox]

Slackware has Debian beat on age.

Re:Do we need this?

[moronoxyd]

The statement is that Debian is the oldest community-driven distribution.

Personally, as a Debian user, I always think of Slackware as the oldest distribution still in use.

Re:Do we need this?

[JWSmythe]

If you narrow the focus enough, you could pick anyone as the winner.

    The most important due to the number of users?

    The most important due to the number of commercial users?

    The most important due to the number of contributors?

    The most important due to the number of lines of code submitted back up?

    The most important due to longevity?

    The most important due to the number of commercially released versions?

    The most important due to the number of commercially released disks?

    The most important due to the number of ISO downloads?

    You get the idea.

    I agree with your last statement though. I'm a Slackware user. I've been a Slackware user since ... well, long enough where I had to install from floppy disks, because most people didn't have a CD drive, much less a burner, and there were no ISO releases. But I'm not a Slackware zealot. I've used all the major distributions, and many of the minor ones, as well as just about every available OS since the late 80's.

Re:Do we need this?

[julesh]

Most Gentoo users would argue their package management system is the best. It can certainly manage things none of the others can, although that comes with obvious downsides too...

Re:Do we need this?

[jouassou]

Yet another disk measuring contest? Seriously?

Re:Do we need this?

[Fallingcow]

By far my favorite. I finally went back to it after a few years on Ubuntu (god has it gone from being sensible and complete yet sleek to a bloated mess in the last few releases) though I'm going the Sabayon route this time because I frankly don't care about compiling every single package, but I want access to Portage and Gentoo's config tools if I need them. The way init scripts are handled in Gentoo and Gentoo-derived systems is especially great--if I had to pick one part of Gentoo for every other distro to copy, it'd be that.

Re:Do we need this?

[Lennie]

Most of my disks are 3.5" :-)

Re:Do we need this?

[Lusa]

    The most important due to the number of ISO downloads?

The most important due to the sheer size of the ISO downloads.. stable is 8 dvds or 52 cds? ;)

I have no idea what is on most of those (nor likely care) but I like that someone does care enough to be the package maintainer. I'm sure other distributions also can say the same so I'm not saying Debian is best for that. I'm not even saying it's best, different distro's for different tastes and needs. I personally dislike the sheer number of distributions though. It's almost becoming a fashion statement and what is in this season. I predict a retro comeback in 2012, console only with blingy gold border decorations :D

Re:Do we need this?

[Inner_Child]

When looking at distros named "Ubuntu", it seems that Ubuntu is the oldest in the category!

Re:Do we need this?

[tsa]

And this, ladies and gentlemen, is the reason why Linux never made it to the desktop: as long as developers don't unanymously chose for option 1, chaos will prevail in the Linux world.

Re:Do we need this?

[AchilleTalon]

Yes.

Re:Do we need this?

[drinkypoo]

Or, option 3, they mean that Debian is the least useful distro because everyone who wants to do something with it ends up making a new distro based on it.

Or, option 4, they mean that Debian is the most useful distro because everyone who wants to make a new distro based on it can easily do so.

I find it amusing that you made a whole comment bitching about how the story says nothing when you in turn say nothing.

Re:Do we need this?

[kaizokuace]

Yes. Man has had dick measuring contests throughout history. It will continue to the future. When humans finally evolve into pure energy beings we will compare our energy space dicks.

Re:Do we need this?

[next_ghost]

--depclean has been around for quite some time now. If you uninstall packages with --unmerge, don't blame the tools for giving you rope to hang yourself.

Re:Do we need this?

[supremebob]

Funny... after reading TFA (which is basically just a huge list of large and small Linux distributions that are based off of Debian), I came away feeling like I should blaming all of these Debian spin-offs for the insane fragmentation of the Linux desktop ecosystem.

I'm not saying this is a necessarily a bad thing, but it's definitely a good reason why commercial software vendors don't want to support Linux on the desktop. I can only imagine what a pain in the ass it would be to QA test your product on each one of these distributions to make sure that it installs and works properly.

Re:Do we need this?

[Nursie]

Could be, but I think it's that everyone who has the same name agrees with each other!

Re:Do we need this?

[Nursie]

Window manager != theme

Otherwise a very witty and true post.

Re:Do we need this?

[Haeleth]

No, the reason Linux does not have a large desktop market share is because it matured 10 years too late, after Microsoft had already established a stranglehold on the desktop market. The barrier to entry is massive. Fragmentation is a minor issue in comparison to the difficulty of challenging an established monopoly.

The only way Linux will ever succeed on the consumer desktop is if it (a) runs all Windows applications and games perfectly, and (b) never presents users with any uncertainty or minor difficulties. Because the truth is this: when a user has have a problem with Linux, they blame Linux. When they have the same problem with Windows, they blame themselves or their computer. That is the real reason why Linux has only made major inroads in markets such as smartphones, where there was no existing monopoly.

Re:Do we need this?

[h00manist]

All I know is open source advocates need to stop arguing among themselves and give as much support as possible to developers. Open source developers are preciously few and very often little recognized, struggling with the lack of cohesion among systems, tools, and libraries, under financed, represented, etc. The primary reason for any platform's support is developers and their supports, production capacity and enthusiasm for their platforms. I'm all for decentralized and independent initiative for everyone, but cohesion, coordination, cooperation and agreement are certainly necessary parts of independence too.

Re:Do we need this?

[i.r.id10t]

Of course, if MS were broken into 2 different companies (one for apps, one for OS) the apps side could develop Office for Linux, etc. and possibly the rest will follow (much like Mac)

Re:Do we need this?

[http]

Plausible, but wrong. That used to be the canonical definition of "Troll"

Even Windows doesn't run all Windows applications and games perfectly. See the recent youtube video, installing Windows 1.0, and upgrading through every windows version (save ME) up to W7.

I've spent a large amount of time helping friends and customers through computer problems on a multitude of operating systems, including transitions to and from Linux, and when they have a problem, they blame their ignorance.

Re:Do we need this?

[spitzak]

Nonsense. It is very simple:

Linux could be complete crap but if when a person bought a "computer" and turned it on, it ran Linux, Linux would be the #1 operating system.

The reason OSX is becoming popular is because now a significant fraction of "computers" (the ones from Apple that the cool people buy) happen to run OSX by default. This is despite the fact that 90% of the boxed software you can get at best buy will not work on it, and that it is perfectly possible to wipe the machine and install Windows to get this software to work. Thus showing that the "joe user can't run their software" argument is bullshit.

The number of people running something other than what was on the machine when they bought it is miniscule and it has nothing to do with quality or amount of software (except that amount of software is CAUSED by amount of machines, not vice-versa).

Re:Do we need this?

[ieatcookies]

If you have to spend more time making the tool work than actually using the tool than that tool is not worthwhile. I give exception to those that enjoy fine tuning their tools, I used to be this way myself. Now I just something that works without hassle, linux is STILL not there yet.

Re:Do we need this?

[egranlund]

I agree with this - Linux is not there. I just installed Fedora yesterday after going from a *nix hiatus for the past two years (windows is required at work). Before then I had been a daily linux user for the past two years.
First thing I did after installing was I went to go play an MP3. Can't do that. The music player asks if I want to search for the plugin, I say yup. Then it eventually brings me to a page on the Fedora web page saying "we'd love to tell you how to get MP3s working, but we just can't". Really!? 10 minutes later, I finally found a website that gives me a random command to add a bunch of repos to my package manager, and then another command to install the actual LAME plugin. After that I rebooted back to Windows and haven't gone back yet.

Re:Do we need this?

[tramusen]

Yet another dick measuring contest? Seriously?

unique for being the largest, oldest, 100% non-commercial community-driven metric.

Also the most frequently falsified for the measurer's benefit.

Re:Do we need this?

[zeroeth]

Most of my disks are 3.5" :-)

Mine is 8" when floppy =)

Re:Do we need this?

[SETIGuy]

Or those numbers might mean that number of child distributions is a meaningless benchmark. Perhaps Debian is easiest to clone without getting into copyright problems. Number of systems on which the child distributions are installed would be a better metric.

Re:Do we need this?

[Risen888]

That's the beauty of the whole system. The fact that we have a base like Debian to build on promotes diversity. So I'd say the answer is "both of the above."

Re:Do we need this?

[Risen888]

I sell consumer desktops with Linux. I can assure you it is succeeding in paying my bills. Maybe no one else is trying, but when people come to me with either a broken Windows machine or wanting a new machine that doesn't suffer the same ailments as their old broken Windows machine, Linux pretty much sells itself.

Re:Do we need this?

[Albert Sandberg]

And then they wonder why they can't play their newly bought game? Or the game that someone gave them on x-mas?

I'm an avid linux user, but I'm yet to push it onto my family or friends because they have other needs than me. As parent said, windows compability is key for those people.

Tho, it should be said that with the new webgl to html5 there will be a LOT of change I think, regarding to games. If you can play a game safe in your browser, why use seperate executables? I think this is a whole new era, so devs: get your webgl skills up to date ASAP!

Re:Do we need this?

[synthespian]

Or, option 4, they mean that Debian is the most useful distro because everyone who wants to make a new distro based on it can easily do so.

And I find it amusing that you use circular arguments ("Debian is useful because many people use it") and then go about bitching about another guy's comment, as if you had presented one. Try. Logic.

Re:Do we need this?

[synthespian]

The reason OSX is becoming popular is because now a significant fraction of "computers" (the ones from Apple that the cool people buy) happen to run OSX by default. This is despite the fact that 90% of the boxed software you can get at best buy will not work on it

Those thin Mac books are damn cool. Period.
90% of the 90% software that Windoze users feel they need to install is just some cheap shit they want because Windows ships with crippleware, whereas Mac users are ready to have fun from the get go.

Also, 50% of the 90% of those 90% is just a bunch of people writing low quality software that does exactly the same as the next guy's software. That they can charge for it and make a buck is a beautiful thing.

Re:Do we need this?

[synthespian]

Some distros try a different business model for Linux, one that is not for freeloaders. Linux fanboys love freeloading, hence no Linux apps. Anyways, I digress. Try Mandriva PowerPack (paid version). It ships with proprietary codecs. This makes some fanboys go insane, but then again we don't see those fanboys helping out with any codec, because signal processing is tough shit, and not the kind of stuff bashers and repackagers can wrap their head around. They better stick to repackaging other people's software in .debs.

Re:Do we need this?

[synthespian]

It doesn't promote any diversity. If it's the same distro, then the binaries are in the same place, the same configuration files are there. It's just a low-hanging fruit for the black hats. Much more so Debian, that had their servers cracked open not once, but twice.

Re:Do we need this?

[bill_mcgonigle]

The way init scripts are handled in Gentoo and Gentoo-derived systems is especially great--if I had to pick one part of Gentoo for every other distro to copy, it'd be that.

Well, don't leave us hanging - what's so great about it? (not everybody is going to go install a Gentoo machine to see what you're talking about).

Re:Do we need this?

[bill_mcgonigle]

What kind of GUI do you set up for your users?

Re:Do we need this?

[bill_mcgonigle]

So you installed a distro which has as a core of its mission not to ship non-free software, and then get gobsmacked when it takes 10 minutes to install non-free software on it?

Your beef is with the US Patent System, not Fedora (of course they would ship mp3 if not for the patents, it would then be free).

But you didn't really do any of that anyway.

Re:Do we need this?

[bill_mcgonigle]

The only way Linux will ever succeed on the consumer desktop

You say this like the consumer desktop is going to be around for a long time. New cell phones are running dual-core 1GHz ARM's. That's one generation and a WiDi chip away from being the thing you plop down next to your KVM when you want to use a real keyboard and bigger display.

Re:Do we need this?

[egranlund]

I don't have a problem with not shipping the plugin but at least make it easy to get to or make some page just giving a rundown of another repository you can point to that is outside of the distro (like ubuntu/debian). Making users go outside on the internet and hunt down the documentation to do something so simple is ridiculous.

Re:Do we need this?

[Risen888]

And then they wonder why they can't play their newly bought game? Or the game that someone gave them on x-mas?

Actually they don't. Everyone knows you can't put a Playstation game in a Wii. It's not a terribly hard concept to get your head around.

I'm an avid linux user, but I'm yet to push it onto my family or friends because they have other needs than me. As parent said, windows compability is key for those people.

Apparently it's not. I'm paying the bills. Document compatibility's still pretty important, but even that's becoming less of an issue than it was even five years ago when I went into business. Binary compatibility, I assure you, is a non-issue the overwhelming majority of the time.

Re:Do we need this?

[Risen888]

When I first went into business five years ago, I shipped a not-quite-stock Ubuntu desktop stack. In the last year and some, I (more to the point, several of my customers) started to get fed up with Ubuntu constantly dicking with their interface, and at the same time, KDE really has started to shine on its 4.x series. I am in the process of transitioning some of my more adventurous clients to KDE, and if that goes well, that's going to be my default stack going forward.

I'm excited about this, because I personally like KDE a lot more than even straight Gnome, and a damn sight more than Ubuntu's bastardized version. I think the application suite blows Gnome away, which means it's simpler for me to ship out of the box without stuffing a bunch of other shit in there, and as a vendor, the ease of desktop customization is a huge win for me. I have a custom color scheme, font scheme, and widget theme that I can ship on KDE with zero effort, and that's awesome. I'm really encouraged so far by the feedback I've gotten from the early adopters (who are obviously non-tech people, BTW, or they sure as hell wouldn't be coming to me), and expect to be shipping nothing but KDE going forward if this experiment keeps going well.

Re:Do we need this?

[Albert Sandberg]

Interesting. Next time my mother wines about the computer perhaps I should take a ubuntu live cd over to her place and show her around... I've tried this many times but I think I have to simplify ubuntu further to get her to not nod it down as too complicated, like for instance removing the multiple desktops which are ace when you know how, but just "where did my window go?" for beginners...

I made sure she ran firefox and her mail is web based plus she mostly uses the computer to manage photos.. perhaps you have a good solution for that? I'm currently running digikam but that's almost to complex.. perhaps I could use dropbox for galleries..

Re:Do we need this?

[drinkypoo]

I find it amusing that you missed the whole point of my comment and thought you were funny anyway.

Re:Do we need this?

[jbolden]

Come on, you think people are buying Apple for the hardware despite OSX? I'd buy one of those Fuji tablets in a second if I could OSX on it.

Re:Do we need this?

[GuyWhoSteals]

It is, in fact, very easy to install an MP3 codec from RPM fusion. You should have googled. (And yeah, I don't care about those patents)

Re:Do we need this?

[Risen888]

Interesting. Next time my mother wines about the computer perhaps I should take a ubuntu live cd over to her place and show her around... I've tried this many times but I think I have to simplify ubuntu further to get her to not nod it down as too complicated, like for instance removing the multiple desktops which are ace when you know how, but just "where did my window go?" for beginners...

Part of the service I provide is a brief tutorial on "how the desktop works." It usually only takes me an hour or so to walk folks through it, but I consider the most important hour I spend with a new customer. As an aside, most people really like multiple desktops once I show them how to use them. One tweak I make is to disable "scroll wheel switches desktops," because that does fuck people up.

I'm currently running digikam but that's almost to complex.

I do ship Digikam. You may also want to look at Kphotoalbum.

Re:Do we need this?

[bill_mcgonigle]

Thanks for the feedback and your work to promote free software. I've pretty much settled on the same, though this whole Nokia/Microsoft thing has me a bit unsettled as to its future.

I'm pleasantly surprised to hear that non-geeks are doing well on KDE - I would have thought that, say, digikam, would have been too complex for somebody more used to an iPhoto sort of arrangement. (I need to migrate my wife at some point!).

Re:Do we need this?

[CAIMLAS]

So does my mom, but as far as I know, there's really only been one guy working on her.

Re:Do we need this?

[Fallingcow]

Well, you could look at the documentation :)

Seriously, this part of the handbook explains how it works pretty well. The "working with rc-update" and "changing the runlevel behavior" sections are probably the most important for most users--or rather, for most of the subset of users that would care about this to begin with.

Re:Do we need this?

[laurelraven]

Even Windows doesn't run all Windows applications and games perfectly.

No, it doesn't...however, when it doesn't, either the software or the system get blamed far more often than Windows does. Even if Windows does get blamed, the user sighs and trudges along because they know there is no other alternative.

When someone tries Linux, however, every little thing that does not work the way they expect are automatically Linux's fault, because these things "Just Work" in Windows (even when they really don't), so Linux must, therefore, be inferior (this is especially true of people who are otherwise smart, and/or tech savvy..."if I'm this smart, and can't figure it out, it must be badly designed"). After all, if I have to put up with annoyances and breakages in the system, I may as well stick with something that I already know and am used to being disappointed by, rather than something that I have to go to the trouble to learn well enough to even use halfway as well as I do what I just came from. They see it as replacing one problem with another problem.

While it may be true that Linux may get desktop acceptance without doing what Haeleth said above, I think that being able to do everything Windows is expected to do, not just better but perfectly, and exactly the way that users coming from Windows would expect it to be done, is the only way for acceptance to come quickly (sorry for the run-on there...).

Anymore, I don't hate either platform, and see both as having strengths and weaknesses...one of Linux's big weakness right now is that of perception, that it is harder or that people have to completely re-think what they know about computers to use it (which is almost half right...they do have to re-think part of what they know). People simply need computers to do the task they need done. They don't care what operating system is on it, or the morality or ethics of the software that they are using. Most of the time, they don't even care what it costs, as long as it does what they need done, with minimal hassle, so they can get back to doing other things. If we can make Linux work like that and set it up for them, they won't even care or notice; as long as it just works, they will do their task and get away from their computer as quickly as possible, just like always.

Re:Do we need this?

[speculatrix]

mod parent up!!

Re:Do we need this?

[Risen888]

Thanks for the kind words. I consider myself damn fortunate to get to do what I do every day. I don't know too many people who love their job this much.

I'm also...well, not "surprised" so much, but really pleased that KDE is looking viable. I remember when 4.0 came out, the unholy mess that that was, and even then I thought "Damn, one of these days this is gonna blow everybody's socks off." I think that day has come. I'm not too concerned about Nokia, to tell the truth. KDE was around before Nokia came on the scene, and it'll be here long after Nokia comes and goes, because KDE's exciting in a way that Gnome is not, and people are always gonna want to work on the exciting projects. (No offense to Gnome devs and the work they do.)

Finally, to your last point, I do ship Digikam, but you may also want to have a look at Kphotoalbum. It's a little more spartan, very accessible.

Re:Do we need this?

[Zancarius]

This is despite the fact that 90% of the boxed software you can get at best buy will not work on it, and that it is perfectly possible to wipe the machine and install Windows to get this software to work. Thus showing that the "joe user can't run their software" argument is bullshit.

So what you're arguing is that if someone bought a computer with Linux preinstalled, they'd just wipe it and install Windows anyway the instant they discover certain software won't work (to say nothing about Wine).

I'm not sure, but it seems to me that this negates the rest of your point. I'm sure to be missing something.

The OP's argument was that Linux will do better than it is if it can run Windows software perfectly. Your argument is that it'll do better than it is if it comes preinstalled (fair point, I agree). Then you use OS X as an example and make a statement that the solution to being unable to run certain software is to install Windows anyway, which seems to defeat the purpose!

Either way, I don't think it matters. Linux on the desktop isn't going to happen any time soon for many reasons; yes, it's easy to condense it to a handful, but I don't think such simplification is truly appreciative of the hurdles that still exist. Never mind market share.

Re:Do we need this?

[spitzak]

What I meant was that if someone bought a computer with Linux preinstalled, they would NOT wipe it and install Windows. For the same reason people who buy an Apple machine won't wipe it and install Windows.

OSX does not run WIndows software perfectly, in fact it is far worse than Linux. This does not seem to be stopping it. There is a belief that this is due to quality, but I suspect the majority of the reason is that only extreme geeks actually reinstall operating systems. OSX could be the worst thing in the world (I don't think it is btw) and if people bought the machines to get the cool apple logo and shiny case, they still would not reinstall.

Re:Do we need this?

[marty23571113]

Slackware has Debian beat on age.

Yes - by about 6 months. Debian has many more packages available using synapatic, software center (in Debian a6.0) or mainly apt-get. Why does the package manager in Slackware called slapt-get? Slackware requires even more work than Debian and will never be adopted by anyone as a primary OS except by masochists.

Re:Do we need this?

[Pf0tzenpfritz]

No, the reason Linux does not have a large desktop market share is because you are too fscking stupid to use it. Look at today's internet - user are even too bloody stupid for FTP or real email. Right now systems without even keyboards are the big dead - because the main target group are lazy idiots too fat and silly to even write.

Re:Do we need this?

[Stuarticus]

Vaios are shiny, Alienware are shiny. People want a computer that works. That's why they buy macs.

Re:Do we need this?

[armanox]

Package manager in Slackware is pkgtool. Just like the package management tool in Debian is dpkg. Apt is a convenient tool that works well with dpkg and rpm, and many people think that slapt-get works quite well. Being familiar with both I'm going to strongly disagree with your closing statement - Slackware is great for finer control of the system.

short discussion

[Randle_Revar]

>Without Debian we are nothing. Debian is the most influential and important Linux,

100% true, and all that needs to be said. Story over, thread over.

Re:short discussion

[Anonymous Coward]

There was a time when we used to say that about Slackware.

Re:short discussion

[Anonymous Coward]

I think that if you would talk install base, the figure would be much closer to 90%, at least for home computers. The commercial world, I don't know... perhaps Red Hat would still win that one.

Re:short discussion

[clemdoc]

and is still working fine.

Re:short discussion

[Raenex]

Story over, thread over.

Fuck off.

Re:short discussion

[Randle_Revar]

>capable of interfacing with the different types of hardware optimally

Absurd. They both use the same kernel sources, in that their hardware support is identical.
Of course Debian as a whole, and as configured, supports far more archs than RH.

Re:short discussion

[petermgreen]

aptitude has got a UI for browsing and a "cleverer" (though that cleverness can often backfire...) depsolver but really it's a minor part of the debian package management system.

Debians package management is indeed among the best though I think that is less to do with any particular tool (they are good but afaict they aren't fundamentally different from those in other distros) and more to do with the care and attention that goes into the whole process (both the tools and the packages themselves).

Re:short discussion

[synthespian]

No it isn't. If you think about it for a minute, no other distros picked up on their packaging technology. The one thousand claimed distros that did are just...little Debians, not really forks or anything creative (but that I mean - it's just some scaled down version of Debian).

Debian's packaging system almost brought them down to a halt. I remember a few years ago we discussed this on /. The more packages Debian got, the more they delayed their releases.

OTOH, if you look at the rpm tech, it got picked up by different distros (amongst others, the "big distros" SuSE and Mandriva), but they changed and tweaked it. So the technology that got picked up by different distros is rpm, not apt.

Re:Android second?

[X0563511]

Google no longer maintains the code they previously contributed to the Linux kernel as part of their Android effort, creating a separate version or fork of Linux. Android's mobile operating system is based upon a modified version of the Linux kernel. It is not linux.

Specifics: http://en.wikipedia.org/wiki/Android_(operating_system)#Linux_compatibility

Trolltopic - trolldot.org

[Anonymous Coward]

I think my hosts file is messed up. I typed slashdot.org but somehow ended up on 4chan .... but where are all the hot femanons?

Re:Trolltopic - trolldot.org

[carnalforge]

Lol!
I'm sorry i don't have mod points.

Android?

[fatp]

Is Android considered linux?

Re:Android?

[larry bagina]

Red Hat, Debian, and pretty much everyone except Linus Torvalds himself use a modified version of the Linux kernel.

Re:Android?

[migla]

>Is Android considered linux?

In everyday usage the word Linux refers to the whole OS. And by that we mean the kernel, GNU stuff, (sometimes also X11 and whatnot). In light of that, Android is not Linux, even if it technically is.

Kinda funny.

Re:Android?

[mirix]

Not in the "GNU/Linux" sense, which most people mean when they say linux, no.

In that sense, BSD is more linux than android is, and it isn't. If that makes sense.

Re:Android?

[Fnkmaster]

I'd go with yes. But it is not a general purpose desktop nor server distribution, and doesn't come with the full user-space stack that you'd expect from those. Rather, it's a mobile Linux distribution with a Java-based user interface and application environment.

Re:Android?

[clemdoc]

Slackware, as far as I know, uses a vanilla kernel as well.

Re:Android?

[Jonner]

Popularity is not the same thing as influence and importance. I don't think it's much of a stretch to imagine that if it weren't for Debian, GNU/Linux wouldn't have had a high enough profile for Google to choose it for a mobile platform.

Re:Android?

[Risen888]

Arch's kernels are pretty much stock.

Thanks for the Update!

[Anonymous Coward]

Thanks for the reassurance pal! With all this talk of Ubuntu gaining traction and Arch being more widely used than ever, I was really getting scared for the fate of my beloved Debian. Debian, you will never get old and I will always love you.

This is probably the most useless post on Slashdot I've ever read.

Re:Thanks for the Update!

[Stupendoussteve]

Does Arch have secure package management yet?

Debian

[Anonymous Coward]

...is the natural "next step" from ubuntu for those looking for something less experimental.

But one distro being more important than another? Ludicrous. All distros are essentially the same, except for minor variations in desktop environment, package installer, and selection of usermode programs loaded onto the install CD. If one distro were chosen at random and all others ceased to exist, the linux world would continue as usual.

Re:Debian

[maxwell demon]

Is this so? If Debian ceased to exist, would Ubuntu remain?
Of course, if Ubuntu ceased to exist, Debian would remain.

Re:Debian

[Junta]

Maybe nowadays, but debian did do apt before yum or anything else like it, and now that sort of capability is absolutely ubiquitous. In terms of things a "distribution" does inherently, that aspect of package management is very large.

Though, I'm inclined to agree...

[The Altruist]

This smells suspiciously like flame-bait. And if you look carefully, you'll see an army of trolls off in the horizon.

Re:Though, I'm inclined to agree...

[Pieroxy]

Who is going to call the dwarves and the elves? We need help over here !

Re:Though, I'm inclined to agree...

[spongman]

I think the smell might change once the trolls get closer, though.

Descendent distributions != Importance

[mhotchin]

Isn't 'Number of descendent distributions' a crappy metric for 'Importance'? Wouldn't something like 'Installed base' be humongously better?

Re:Descendent distributions != Importance

[Hognoxious]

Isn't 'Number of descendent distributions' a crappy metric for 'Importance'?

No. It's perfectly adequate for starting a flamewar among ignorant zealots and obsessive fanboys in order to generate page hits and advertising revenue.

P.S. Ubuntu sucks.

Re:Descendent distributions != Importance

[Gaygirlie]

And how would you obtain any reliable numbers for counting installed base? Most people do not register themselves as a user of any distro or such and since most Linux distros don't call home you really have no reliable way of counting them. Atleast number of descendent distributions is known, with installed base all you have is guesses.

Re:Descendent distributions != Importance

[mhotchin]

But 'easy to measure' is not the same as 'useful to measure'. We could measure the size of each distribution, but it doesn't really tell us anything.

I threw in 'installed base' just because it was the first thing that came to mind - my real point is that descendent distributions just doesn't imply much, no matter how easy it is to determine.

Re:Descendent distributions != Importance

[ldobehardcore]

I'll have to agree. PuppyLinux has at least 3 dozen variants, and I'm probably horribly underestimating.

Does anyone see people using puppy linux as an important and critical operating system?

It's a toy, and everyone wants their own decals for it.

Re:Descendent distributions != Importance

[jgrahn]

Isn't 'Number of descendent distributions' a crappy metric for 'Importance'? Wouldn't something like 'Installed base' be humongously better?

Yes. But TFA started out by saying 3/4 of the most downloaded distros were Debian-based. The "number of descendent distributions" was just something that looked nifty and quotable.

The killer features of Debian though are that it's driven by *people*, not shareholder profits, and that these guys work in the Unix tradition. Kind of like a *BSD, I imagine.

Re:Descendent distributions != Importance

[GPLHost-Thomas]

Wouldn't something like 'Installed base' be humongously better?

In that case, Ubuntu is the most used distro, and 75% of it's packages are coming from Debian. There's a real synergy between both distro too. So in both cases, the result is the same.

Re:Descendent distributions != Importance

[Max Romantschuk]

P.S. Ubuntu sucks.

Not really biting, but on the subject of Ubuntu:

What I really love about Ubuntu is having a Debian stack underneath, with a nice desktop setup on top. I can do normal desktopy things without extra fuzz, but when I happen to need backend-component-x for our webdev stack I can just apt-get install it with a 95% probability.

I really like the blend of user friendlyness and a solid foundation that Ubuntu provides, and a lot of it is really thanks to Debian. For me, it's a win-win.

Re:Descendent distributions != Importance

[backganon]

Amen brotha.

Re:Descendent distributions != Importance

[icebraining]

Except for Mint Debian.

Re:Descendent distributions != Importance

[Saint Stephen]

Debian doesn't like 95% probability (stable should be stable). Plus Ubuntu ships with non-free parts that make it unacceptable to Debian.

I think the only reason I like Debian as a desktop is I use FluxBox as my windowing manager, no dock, and just use the whiz-bang GTK-based *apps*. Best of both worlds and fast.

Re:Descendent distributions != Importance

[ladoga]

Not really biting, but on the subject of Ubuntu: What I really love about Ubuntu is having a Debian stack underneath, with a nice desktop setup on top. I can do normal desktopy things without extra fuzz, but when I happen to need backend-component-x for our webdev stack I can just apt-get install it with a 95% probability. I really like the blend of user friendlyness and a solid foundation that Ubuntu provides, and a lot of it is really thanks to Debian. For me, it's a win-win.

Ubuntu's OS X mimicry, especially in its recent versions and fast release cycle - which tends to break things - are few things I dislike in it. Though LTS releases make the second issue less important. I'd like to see such highly visible distro having it's own style, trying to lead the way - or choosing its own way - rather than following others. In other points you mention I agree though. Ubuntu is still the distro I recommend to most people that are new to Linux.

If user has some clue what specific software he likes to use, then my vote goes to Debian. 'Stable' for servers and 'testing' for desktops. It doesn't have any of annoyances of Ubuntu, while all the good stuff (apt, Debian configuration tools, etc.) is there.

It also has the nice advantage that if certain software package in 'testing' happens to have a bug you can always fetch another version from 'unstable'. I've now had my desktop 'testing/unstable' install for some 5 years, incrementally upgrading to newer software as it comes available and haven't run into bigger problems than occasional bug that hasn't been too hard to work around. Usually the way I mentioned above, just by installing another version of the said package.

Re:Descendent distributions != Importance

[Jonner]

Isn't 'Number of descendent distributions' a crappy metric for 'Importance'? Wouldn't something like 'Installed base' be humongously better?

"Installed base" is a great metric if you include the descendent distributions.

Re:Descendent distributions != Importance

[froggymana]

Citation needed.

Red hat definitely is important though in the fact that their developers contribute huge amounts of code back into the community.

Oldest?

[SilverJets]

I think Slackware is just slightly older than Debian and this graph seems to indicate that as well.

Re:Oldest?

[Rakshasa Taisab]

In my definition of The Bost Important Linux Distribution, the one and only qualification is that it is running on one of my machines. Thus Slackware takes that price with ease.

Re:Oldest?

[FragHARD]

My thoughts exactly... I know I was using slack way before I even tried .deb ;-)

Re:Android second?

[Anonymous Coward]

Do programs compiled for Linux which are not dependent on outside libraries or utils (this includes all GNU software) run on Android? Yes. That means Android is effectively Linux.

Re:% of distributions vs % of user installations

[ljhiller]

Mod Parent Up. Anybody and their dog can start a Linux Distro. Maybe they start with Debian because it's guaranteed to be free and open, or like like Debian package management. And maybe 500 people install it. I'm sure Red Hat and Debian are #1 and #2 in ACTUAL deployments, although for all I know Red Flag tops them all.

Re:No, it's not

[Cochonou]

They have had stable releases every two years lately.

how about the BSDs

[kayumi]

Where would the BSD distros fit into this? Or how about OpenSolarinux?

Re:how about the BSDs

[clang_jangle]

Where would the BSD distros fit into this? Or how about OpenSolarinux (sic)?

Not being linux, they don't.

Re:how about the BSDs

[Lennie]

Does it matter ? Because Debian is now a BSD-distro now. ;-)

http://www.debian.org/ports/kfreebsd-gnu/

Seriously.

I think OpenBSD might have the most influence, because they created/maintain OpenSSH.

Which is used in many, many if not all Linux, BSD and other Unix based systems, routers and managed switches.

I think FreeBSD is where a lot of drivers are being created for all the BSDs and I think for the Linux kernel as well.

FreeBSD is also used by Juniper as the basis for Junos for their routers, which runs a large percentage of the internet.

OpenSolaris is dead, but OpenIndiana/Illumos will keep it going for that community. Which means there is free code which can do ZFS and Dtrace (which itself is also incorporated in FreeBSD).

Try CentOS

[Nova Express]

If you go by the importance of infrastructure run, I would guess that CentOS (binary compatible with Red Hat, without Red Hat fees) is is the most important Linux distro out there. The last three companies I worked at that use Linux in the data center used CentOS.

Re:Try CentOS

[hdparm]

CentOS is not binary compatible with RHEL, CentOS is RHEL, sans RH branding stuff.

I agree with you - since Debian inception, lot has changed in where and how Linux is used. So Fedora, being the base upon which RHEL is built seems to be the the most important one these days. Some will argue that it's not community driven though.

Re:Try CentOS

[Alain Williams]

It is somewhat naive to try to say ''best'' across all domains. Red Hat (and to a lesser extent) SUSE are very important in the commercial sphere. This is because of the nature of PAID support - something that large companies demand. It all depends on where you are viewing things from.

Re:Try CentOS

[Kjella]

Except by the logic of the article, that would make RHEL the most important, since without it there wouldn't be CentOS. But if you're first going down that line you might as well go one more step, without the individual apps there wouldn't be anything to make a distro of.

I think the article is confusing convenience with importance. As long as Debian does the job well, it's very convenient for the other distros to not do the same job and just work off that. I think many employees have discovered that as well, your job responsibilities may be important to the employer, even critical and essential but you are still replaceable.

If Debian wasn't there, most of the same people with the same interests would still be there and create a system reasonably similar to it. It's not like all the distros that "depend" on Debian would just fold up and disappear, nor would most of the packagers and such. I'd be a lot more concerned about the future of the essential packages, it's not Debian that determines how Linux / X / Gnome / KDE / OpenOffice / Firefox etc. evolves.

Re:Try CentOS

[supersloshy]

CentOS is not binary compatible with RHEL, CentOS is RHEL, sans RH branding stuff.

That's exactly what binary compatibility means... If the software is essentially the same as RHEL but without branding, any software made for RHEL should work/compile on CentOS (same with most other Linux distros but that's besides the point).

Re:Try CentOS

[Aighearach]

Sorry Junior, but by that logic you're not you; you're just compatible with yourself. Try it in a philosophy 101 class, you'll get better results there than in real life.

Re:Try CentOS

[supersloshy]

Sorry Junior, but by that logic you're not you; you're just compatible with yourself. Try it in a philosophy 101 class, you'll get better results there than in real life.

...Are we talking about totally different things here? If the software on CentOS is EXACTLY THE SAME as RHEL, it's essentially the same thing. The only things changed are artwork and references to Red Hat. Why wouldn't that be binary compatible with RHEL? Explain it to me in a good example why a program that was written exclusively for Red Hat won't work in CentOS (doesn't RHEL follow the Linux Standard Base anyways?). Mere and minor aesthetic changes, if you don't change the code itself severely, should mean that something is binary compatible... right?

Re:Try CentOS

[IgnitusBoyone]

His argument is that BC is more then just if I copy and paste this source code it will run programs already compiled. Its a given that CentOS and RHEL are BC because they are the same code base. BC is normally used to imply a design relation where as two different source bases can be run code compiled on either. Generally the two projects will be radically different.

https://secure.wikimedia.org/wikipedia/en/wiki/Binary_code_compatibility

I also use it to describe updates to a code base that does not require projects dependent on it to recompile. In this case the implementation may be completely different but the interface is the same so the dynamic libs will still work. I guess this whole speech is that if you say CentOS and RHEL are BC your oversimplifying BC and you should default to there the same damn thing.

Re:Try CentOS

[dAzED1]

Some will argue that it's not community driven though.

And those people are misinformed. Fedora is substantially more community-driven than Ubuntu. The lack of ability to add random shit to the distro just means it's also more secure; the general fedora community is more concerned about having something they can trust, I guess.

Re:Try CentOS

[xdroop]

That's because they're big, clueless dinosaurs, who don't understand that Debian is the more complete, better maintained solution [blah blah blah blah]

No, they run it because the tools they run on linux demand a platform which can reasonably be depended on to be universally the same everywhere, and available with support contracts so that if something is wrong you stand a chance of getting it fixed beyond the usual dude, you have the source, fix it yourself that free software so enjoys.

Seriously -- these tools can cost so much that even a "real" Red Hat support license for the platform is noise. Just pay the man and be done with it.

Some of us have work to do.

Re:Try CentOS

[hdparm]

Yep they are (misinformed). I really didn't want to start flamebait thread within the flamebait article, so decided to refrain from explanations.

Re:Try CentOS

[dAzED1]

apologies - was already annoyed at something else. Never go to forums when you're already annoyed, I guess ;)

Yeah, we need Debian

[Compaqt]

See, with
1. RedHat doing their weird patches thing, and their restrictions when you use RedHat Network (Red Hat Stops Shipping Kernel Changes as Patches), and the huge lag times between RHEL updates
plus
2. Ubuntu doing stuff that some people don't like, plus the whole Unity/Wayland thing,

the importance of a good, free, working and fresh distro is highlighted.

OK, so you're going to say "Debian, fresh?" But I think this might be a good time for both Ubuntu users to test the Debian waters, and for Debian to get its act together.

Re:Yeah, we need Debian

[Provos]

RHEL's lag times are fine by me. While I don't work with a giant-multinational corporation, enterprise stability is still of utmost priority to my staff and I. We like that while RHEL stays on top of security issues, they do not make frequent jumps to newer application releases without significant warning.

As far as RHN, if you don't like paying for support, use CentOS. We do for our development and testing environment, and can be assure that it will stay in sync, once the base platform port occurs.

For desktops, our devs use whatever makes them happy. I'm running Fedora 14 right now, and several others are using various debian-sourced releases.

I think the real detriment, as mentioned earlier, was the growing lack of diversity in the distribution sphere.

Re:Yeah, we need Debian

[Haeleth]

See, with
1. RedHat doing their weird patches thing, and their restrictions when you use RedHat Network (Red Hat Stops Shipping Kernel Changes as Patches [slashdot.org]), and the huge lag times between RHEL updates
plus
2. Ubuntu doing stuff [slashdot.org] that some people don't like, plus the whole Unity/Wayland thing,

the importance of a good, free, working and fresh distro is highlighted.

I absolutely agree 100%.

That's why I'm so glad Fedora exists. It's as usable as Ubuntu, avoids non-free software just like Debian, and is fresher than either.

Re:Yeah, we need Debian

[OFnow]

RH drove me off. First they dropped end user support.

Then Fedora was a gigantic pain: Upgrades did not work (had to reinstall) and every single time, I struggled for weeks getting sound and skype to work after getting latest Fedora, Even getting simple modern monitors to work took effort.

Ubuntu fixed all the problems Fedora had...

Re:Yeah, we need Debian

[loupgarou21]

You do realize that Ubuntu _is_ Debian based, right? It's part of the 63% of Debian based distros.

Re:Yeah, we need Debian

[Compaqt]

Yeah, I do. And, I don't mind you pointing it out because some people might not.

My point being that Debian might be considered to not have the latest weirdness that Ubuntu is moving toward.

Re:Yeah, we need Debian

[zaivala]

Are you forgetting that the large percentage listed for Debian "influence" is mostly made up of Ubuntu users? Your statement seems to imply that Ubuntu is not Debian...

And yeah, Ubuntu left me behind when they went to Unity -- my laptop is too old to have a driver that allows it to run Unity. But I can still run older versions of UNR/UNE, or slightly older versions of GNOME, or just load Puppy (MacPup 5.11 preferred), and I'm still using Ubuntu AND Debian.

Re:Yeah, we need Debian

[GuyWhoSteals]

How hard is installing Skype from the Skype repo? A yum install worked for me.

Re:Yeah, we need Debian

[CAIMLAS]

As a long-time Debian user (13 odd years now, I think) and someone who has been using Ubuntu since early 2008 on the desktop, I've proposed a different approach for a distro that is somewhere between Debian and Ubuntu.

For those who remember: something more akin to Stormix Linux or Progeny Debian in principle and action, but still somewhat different (in the following fashion).

Instead of having a 'distribution' you would have distribution tiers. Debian excels and appeals to many for a number of reasons which Ubuntu almost completely lacks; namely:
* The ability to maintain consistency across releases
* Keeping things at the "system" level as similar between releases as possible, despite the significant kernel/library/infrastructure changes taking place
* Stability and completeness
* dpkg/apt/aptitude

However, it falls short in a number of places where Ubuntu succeeds (and has made it wildly popular):
* End-user experience/polish
* Near features where people/end users want them
* Accessibility/ease of installation
* Inclusion of all the necessary/desired drivers/packages to get pretty much any system up and running easily (eg. nvidia binary drivers)

Of course, Ubuntu has largely broken a lot of the things which make Debian appealing (eg. the inclusion of upstart/replacement of the old linear init), which burns people like me.

Debian is much more the 'traditional' distro: we're just packaging things, not making decisions for you.

As mentioned earlier, the distros like Progeny Debian and Stormix Linux back in the dotcom era illustrated basically what made Debian great: you could take a stable, well designed base distribution and drop your "integration", installer, and the like on top of it to provide a very polished product.

Finally: here is what I would like to see. Instead of a massive "debian" distribution, I would like to see Debian broken down into further flavors, if you will. Debian would, in essence, be a base: everything you get with a base install of Debian would be "Debian" (or, alternatively, a subset thereof) - akin to how NanoBSD is a base for a number of other projects. This would be a stable base which would be updated independently (or at least, as a prerequisite to) everything built on top of it. Ubuntu 12 could use Debian 7 as a base; you'd get your consistency. Arch (or whatever) could do the same. You update your packages independently from the "OS", the packages of which do not need regular updating anyway.

Speaking personally, if it weren't for the headache of doing so, I'd still be doing the "install Debian, X, and a couple other staples and then build my own awesome/KDE/thunderbird/firefox/chromium/whatever" approach. I want a core set of 'stable' things I can run the latest software on (sort of how Windows has done things all these years). I don't want to have to upgrade the whole damn system just to get a newer version of a package (particularly when there's nothing 'wrong' with the rest of the software): that just means I'm going to have to deal with a whole new, different set of small bugs that were introduced. (Case in point: Ubuntu 9.10 had a bug in something that caused clusterssh to not work for months; that wasn't something I could easily regress from, because it was a fresh install; my options were to go back to 9.04, 8.10, or try something else. Not cool.) I should not be 'stuck' like this, short of building it and its requisites from source pulled from a CVS repository! I used to build + patch my own kernels, and do a myriad of things like this "just because". It was fun and I learned a lot, but for goodness sake, I shouldn't have to do those things to get a semblance of stability in my basic tools.

Like anything, it's something that requires time to complete (else I'd have done it myself, of course).

Re:Yeah, we need Debian

[OFnow]

Sorry I was unclear. The Fedora Skype problem was that skype would not generate sound output, making the otherwise-working skype useless. Or if sound output worked, microphone capture would stop working. Sometimes a Fedora mid-release small sound update would break general-sound or skype-sound too. Ugh.

The first task with microphone was always to look thru multiple sound-preferences screens and the alsa-sound-configure apps looking for muted microphone. Because my settings got botched by a minor update. Rather ridiculous, and sometimes that was not enough and I would have to write a configure file of some sort, which configure file I would later have to delete after the next sound update (to get sound working yet again). Good grief.

Re:Yeah, we need Debian

[laurelraven]

The Fedora Skype problem was that skype would not generate sound output, making the otherwise-working skype useless. Or if sound output worked, microphone capture would stop working. Sometimes a Fedora mid-release small sound update would break general-sound or skype-sound too.

This.

For a long time, I hated Red Hat and everything that they did because of the horrible experiences I've had with Fedora (had to use it for school, they insisted). I've since tried out CentOS (forced myself to use it for a project to get more used to the RH tools) and was very impressed how everything just worked so well, how easy the command line tools were to use without taking away power...it felt like they put a lot of work into making CLI administration of a server as straightforward as possible. There were things I didn't like, such as the default startup daemons (do I really need to run the bluetooth daemon by default on a webserver?), but they are easy enough to change.

I've since given Fedora another chance...and the only positive thing I can say about it is that it was very pretty. It's just too unstable. Ubuntu, Mandriva, PCLOS, all make far better desktop OSes than Fedora, because things just plain work better, and release grade software actually feels release grade, not "it was time to release the next version, so where everything was at that moment is what we put in it" quality.

show us the stats

[bguiz]

When laying claim to a statment that "X is the most important of Y", one would expect that to be backed up my statisitics proving that point.

The only half-serious attempt that the author has made at this is in the 3rd paragraph. And even then, he is merely quoting select figures from distrowatch, without further derivation or detail, let alone an attempt to paint a balanced picture. The rest of the article is basically a listing of the various distros based off debian.

That is precisely what the title of this article should have been: "List of distros based on debian"

Instead, the author has chosen to go for the dramatic, attention grabbing headline - and has in some respects succeeded, in that as he has gotten his article slashdotted.

Nothing interesting here, don't waste your time RTFA, move on.

Re:show us the stats

[GPLHost-Thomas]

When laying claim to a statment that "X is the most important of Y", one would expect that to be backed up my statisitics proving that point.

Maybe a quick look at distrowatch may help you?

Re:show us the stats

[DataDiddler]

They also forgot to mention that 80-90% of these distros which are derivatives of Debian are just Ubuntu remixes with a different default wallpaper image when you install it. If HannaMontana Linux is a valuable diagnostic in determining which Linuxes are more "important" then...?

Re:show us the stats

[twocows]

I'd just like to point out that, as someone with a little statistics background, statistics never prove anything. Statistics establish associations and correlations, significance and probability; never proof.

Re:show us the stats

[IrquiM]

When laying claim to a statment that "X is the most important of Y", one would expect that to be backed up my statisitics proving that point.

90% of the statistics on the Internet is up anyway - who cares?

Re:There is no important Linux

[maxwell demon]

Even if that were the case, the subject would be meaningful. All ants are small. However, you certainly can ask which is the largest ant. Indeed, it makes even sense to speak about large ants, because while being small at human scale, they are large compared to the average ant.

Of course, Linux is important even in an absolute sense. You probably even have it at home, in your home router.

questionable statistics

[postmortem]

So if there are 10 distros, 1 of them having 99% market share and other 9 have 1% market share, this statistics say that other 9 are "most important". Well they are not. Give us number of actual number of computers using RH-based vs. debian-based distros.

Android is a Linux distro by definition

[Morgaine]

Linux is officially just a kernel, and a "Linux distro" is any suite of user-side, open source software that provides a complete operating system based on that Linux kernel.

That makes Android a totally kosher Linux distro, even if it is an unusual one with a special Java-based UI by default. It can't be suggested that lack of X11 means that it's not a Linux distro, since there are lots of other Linux distros without X11 too.

Re:Android is a Linux distro by definition

[Lennie]

I think officially you should call it: a Java-like language ?

Re:Android is a Linux distro by definition

[Carewolf]

To be a distribution would so of imply that it is being distributed. Android is trying to be an operating system, not a distribution of an operating system.

Re:Android is a Linux distro by definition

[ToasterMonkey]

That makes Android a totally kosher Linux distro, even if it is an unusual one with a special Java-based UI by default. It can't be suggested that lack of X11 means that it's not a Linux distro, since there are lots of other Linux distros without X11 too.

Sorry, Android is a "kosher Linux distro" as much as Phelps & WBC are "good christians".

Re:Android is a Linux distro by definition

[SwedishPenguin]

It's Linux, but it's not GNU/Linux, which is what the term Linux distribution generally refers to, it has nothing to with X but with the GNU toolchain. More like Dalvik/Linux...

Re:Android is a Linux distro by definition

[morcego]

You know, I was thinking along those lines. Considering the sheer numbler of nonpc devices running Linux, ranging from Android phones to CC machines, I have to wonder if embedded linux didn't surpass desktop/server computer in the number of instalations. In my house it is true, with 3 Linux computer and 5 (maybe more) linux based devices.

Re:Android is a Linux distro by definition

[hey!]

I think officially you should call it: a Java-like language ?

I don't think so. Android programs are compiled with the standard JDK compiler then the resultant bytecode is cross compiled for Dalvik rather than the standard Java Virtual Machine, so clearly they are written in the Java language. I think you might mean that programs are written in a "Java-like execution environment", although that's a bit of a stretch; you could just as well say that dotNet's CLR is a "Java-like environment". In some ways that's *more* true than calling Dalvik "Java-like".

Re:No, it's not

[gdshaw]

You're missing the point. Debian may not be the most popular distribution in and of itself, and it is not necessarily the best choice for many users, but it does provide the foundations for (amongst others) Ubuntu and Linux Mint. That makes it technically very important, even if the users who benefit from it don't know they are using it.

As long as you spell my name correctly

[rcpitt]

I don't care what you say about Linux - just spell it correctly.

Debian, Ubuntu, Red Hat, Android, Gentoo, Mandriva, Knoppix, SUSE, Slackware, Puppy, Slax, Freespire...

Who cares?

As long as you don't spell it Microsoft

p.s. The Kernel is Linux - the rest of the stuff is Open Source. Even Apple's OS/X gets it 95% right - they just use MACH instead of Linux and then apply a different GUI.

No matter what - Redmond loses.

Re:As long as you spell my name correctly

[Stormwatch]

I don't care what you say about Linux - just spell it correctly.

Apple's OS/X

Uh...

Re:As long as you spell my name correctly

[Billly Gates]

"As long as you spell my name correctly"

WRONG!!! It is pronounced GNU/Linux. GNU/Debian, GNU/Ubuntu, etc.

Get with the program ... wipes foam from mouth.

Re:As long as you spell my name correctly

[froggymana]

Time to sick the gnu on him!

You've all been trolled...

[Saroful]

by CmdrTaco!

Almost 2/3 of the distros are based on Debian

[G3ckoG33k]

"323 currently active distributions listed on Distrowatch, 128 are based on Debian, and another 74 on Ubuntu. In other words, just under 63% of all distributions now being developed come ultimately from Debian. By comparison, 50 (15%) are based on Fedora or Red Hat, 28 (9%) on Slackware, and 12 (4%) on Gentoo."

Almost 2/3 of the distros are based on Debian, as it includes Mint and Ubuntu. Given the assumed popularity of Ubuntu, that is a lot.

However, should you turn your head to commercial server space, I guess RedHat based systems like Fedora, Oracle, and CentOS would give another picture. The Suse distros would probably climb in such a list too.

Still, the metric is valid in its own setting.

Re:Almost 2/3 of the distros are based on Debian

[Rich0]

Yeah, but isn't the whole draw of RHEL the fact that it isn't one of 500 splintered forks? There is only one RHEL, and that is why companies all over the place use it.

This is an interesting metric, but it points more to what random groups of devs use when they decide to create their own novelty distro rather than the "importance" of the distro itself.

Now, don't get me wrong - if there were some objective way of determining which was the "most" important linux distro out there Debian would be right up there in my list of contenders. I'd also include Android, RHEL, Fedora, Suse, Ubuntu, Slackware, Arch, Gentoo, and maybe even ChromeOS (not necessarily in ranked order). The reason that all of these have followings is that they all have different pros/cons - if you're doing something really crazy or starting your own distro I'd probably consider every one on that list as a possible starting point, and the right choice depends on what you're doing.

In any case, I consider diversity a strength of linux, so all this my-distro-is-better-than-yours nonsense is needlessly divisive. I happen to like one distro more than others, but I use other distros in VMs or whatever when I need to solve particular problems. I also own several flat-head screw-drivers even though I detest them, go figure.

Re:Almost 2/3 of the distros are based on Debian

[Synn]

>Yeah, but isn't the whole draw of RHEL the fact that it isn't one of 500 splintered forks? There is only one RHEL, and that is why companies all over the place use it.

No, companies use it because it's vendor supported. Specifically you can buy support from Red Hat for it, but beyond that because Red Hat does a lot of work with hardware vendors to ensure that the hardware is "certified" for RHEL. Both of those are important for the pointy haired bosses, so you see it used in the corporate environment. And where people don't want to pay for RHEL, they just drop CentOS in it's place in the office.

It's really nothing to do with fragmentation. Debian itself isn't fragmented. It's been running strong since 1993 and has an extremely stable community. It's stability and solid foundation are why it's often used as a base for new distributions.

Re:Almost 2/3 of the distros are based on Debian

[Aighearach]

You're both right, companies use it because it isn't splintered, things are done in a standard, predictable, pragmatic, traditional way, where there is one. And where there isn't, such as audio, they standardize on whatever looks like it can be complete and useful going forwards.

And other companies use it because of commercial support.

The one part that was wrong was where you're telling him "no."

Re:No, it's not

[dbIII]

It didn't think it was that relevant either - but the distro on my phone, eeePC and even knoppix comes from Debian. That is what it is about and not distro install fests from a decade ago.

But Linus says that Debian is pointless..

[Anonymous Coward]

http://www.omgubuntu.co.uk/2011/01/our-exclusive-interview-with-linus-torvalds-lca2011/

“I’ve tried it a couple of times over the years, mainly because the thing Ubuntu did so well was make Debian usable. I always felt that Debian was a pointless exercise because to me, the point of a distribution is to make everything easy. Easy to install, to be pretty and to be friendly and Ubuntu did that to Debian.”

That must hurt.

Re:But Linus says that Debian is pointless..

[Lennie]

He also says:

“I’ve always had a few problems [with Ubuntu.] It’s not very friendly to kernel developers, and I just end up giving up. That’s kind of okay, because clearly I am not the target audience.”

I guess the same goes for Debian. For example Debian does really well on servers.

But Linus obviously is not a 'user' anyway, as he had no clue which distros are the first ones that popularised live-cd usage.

Re:But Linus says that Debian is pointless..

[Synn]

> That must hurt.

Not really. Linus can be pretty stupid. He likes things to be easy and often makes dumb choices in regards to that(hello BitKeeper). In this case saying Debian is "pointless" because Ubuntu make it easier is stupid. Ubuntu wouldn't exist without Debian and the "point" of Debian is to be a stable 100% free open source distribution that will always be there for you.

And it's funny considering the shit that Ubuntu is pulling these days and how people now are looking to move away from that distribution. Guess all that "easy and pretty" has a price, eh?

Re:But Linus says that Debian is pointless..

[elashish14]

And I respectfully disagree (with Linus). Debian makes it easy to build and create other distros, for whatever purpose they serve. Ubuntu puts the finishing touches on the user experience for desktops, while Knoppix does the same for LiveCD users. Continue on for the millions of other distros. And let's not forget, Debian is a great choice for servers, plus they have a great security team from whom a lot of patches flow downstream. Think of it as a meta-distro if you will, one that serves primarily to enable others.

And truth be told, just because Linus says something doesn't make it true. Sure, he's the original architect of Linux. But if he were to say something ridiculous like 'Everybody should use $DISTRO,' would that make it true? Of course not. Do we really want to be like this?

CentOS anyone?

[LS]

Every company I've worked at and virtually every ISP I've utilized have used CentOS as their main Linux distro. Maybe Debian is tops for hobbyist use, but CentOS / Redhat Enterprise is king of the corporate world.

In other words, there is no one distro to rule them all - depends on the context.

LS

Re:CentOS anyone?

[Lennie]

That depends on where you look. Many, many places run Debian servers.

And if it wasn't for CentOS I don't think people would be using as many RPM-based distributions on servers. I don't think many people run Fedora or OpenSUSE on production servers. Because all the other big RPM-based distros are for-pay. And with for-pay distros you don't get such as much community support.

Re:CentOS anyone?

[Junta]

Well, to be fair, CentOS isn't 'driven' by a communtity (other than the mostly simplistic mission of 'rebadge RHEL'. Indirectly, RedHat does the 'driving' for CentOS. That is explicitly why it is so popular, the users piggy back on the hardware vendor and ISV support RedHat has.

Debian is largest among distributions that are not ultimately beholden to a commercial entity and not following the lead of another distribution.

Re:CentOS anyone?

[carnalforge]

That is not my experience. At all the places i've worked as a consultant in IT the only linux distro i've seen is RH. That is 10 years working on unix servers. Another thing that i've noticed is that non commercial distros are automatically banned, there should be some commercial entity with whom to share the blame when something goes wrong.
Maybe it's just Italy? I'm really curious on this.

Re:CentOS anyone?

[petermgreen]

In particular redhat actually delivers (and centos repackages) long term support. Sure ubuntu has releases that they say are long term support but one year of overlap on the desktop and three on the server seems rather low to me for "set and forget" boxes that you don't want to mess with but do want to keep security updates on.

Re:Android second?

[X0563511]

Does it use the Linux kernel? You're answer is also the answer to the question of "is it Linux."

Ubuntu is essentially Debian Unstable (Sid)

[alizard]

with bug fixes by Ubuntu devs (which get backported to Debian), some cooler packaging, and a better driver selection. I thought everyone in the Linux scene who's been around for any significant length of time knew this.. For more details, read Ubuntu is based on Debian unstable.

I would say that Debian has a certain importance to the Ubuntu community, given that if Debian disappeared, the Ubuntu community would have to whip up all these cool new packages from scratch rather than fixing what's broken from unstable and adding non-free drivers and cool packaging. If Debian disappeared from the face of the earth today, Natty might come out on time... but with luck, its successor might come out in 2 or 3 years. Or maybe Canonical goes bankrupt and it never comes out at all.

I switched from Debian Lenny because of the better driver (as in non-free) selection in Ubuntu, I had to get my new motherboard running immediately.

I think Ubuntu (I use Kubuntu Meerkat) is a better desktop distro, it was intended for ease of use, which is important for most people who make a living with computers in an office environment and pretty much succeeds. If I were going to run a server, I'd put Debian Stable or Testing on it.

freedom

[be1993]

Debian is the definition of open source software, and the proof that we don't need actually companies to impose their interests. That Man is intelligent enough to stand on his own and he doesn't need any mediator to tell him what is the best for him....

Re:Android second?

[Urkki]

So where does that place Android? Maybe third because there are some embedded Linux distros that are in everything from TVs to coffee pots.

This was about Linux distributions. There aren't any Android Linux distributions. Therefore Android doesn't qualify to take part in this comparison.

Or if there is, please give me link to an ISO so I can try it out! A Live CD would be even better!

Re:No, it's not

[GPLHost-Thomas]

Exactly what do you call "infrequent"? We release Debian 3 times a day (that's the number of Dak run per day in SID), and stable every 2 years (that is at least truth for both Lenny and Squeeze, and everyone is trying to keep the pace). More and more people that once moved to Ubuntu are returning to Debian. Exactly WHAT is making you say that "development seems halted"? Is that the 10 000 new packages that came with the new release (Squeeze), which brings the number of packages to nearly 30 000? Stay in the past if you want, don't read the information, but please, we don't need such a silly comment.

First?

[dna_(c)(tm)(r)]

You didn't RTFA did you? Debian came in first. Not you.

Re:maybe it's just not so good

[Isaac-1]

Because a base Debian install can still be small, so it is a great starting point for so many speciality distros. You try installing Red Hat or Fedora and even with lots of trimming it is hard to make it fit on anything smaller than a 9 gig drive.

Re:Android second?

[ChunderDownunder]

Android 2.2 for i386

Debian pure blends

[GPLHost-Thomas]

Instead of having silly derivatives, we'd like to see more "pure blend" flavor of Debian, in order to avoid to spread the efforts, and also avoid useless forks. TFA talks about Knopixx, but totally misses the efforts of one of our very active DD: Daniel Neumann, who wrote major parts of Debian Live.

Low-hanging editorial fruit

[Platinum Dragon]

Hey, while we're all getting drawn into well-burnt flamewars, can we get a vi vs. emacs story for our weekend debating pleasure? Maybe some BSD vs. GPL flavour for good measure?

Re:Low-hanging editorial fruit

[Aighearach]

For programming you really can't beat emacs with zenburn and mumamo.

That said I also use vi/vim every day; for system administration. Mode based editing is really great when touching mission-critical files that aren't hooked into a version control system. (no, backups don't count)

As for BSD vs GPL the modern trend is BSD _and_ GPL. Yay! "You can be anything you want this time around!"

Re:Low-hanging editorial fruit

[Risen888]

We did that already. The answers are Emacs and "it depends," respectively. Next question.

Re:No, it's not

[jimicus]

Seriously, are you trolling?

The latest version of Debian Stable was released about a month ago. It's not great for the desktop, mainly because other distributions (most notably Ubuntu) have overtaken Debian - mainly because they've opted for a pragmatic "hey, it works, who cares about the license as long as it's legal to distribute?" approach.

But put Debian on a server and it absolutely shines.

Stable, reliable, timely security updates which don't tend to break anything, upgrading in place usually works pretty damn well. And while it might sometimes take a while to get the latest versions of a given package, by the time they reach Debian Stable you know they're usually pretty damn solid. Which is what you want on a server.

Most important

[soundguy]

The most important distro line to me is the one that's been sitting in my data center generating my paycheck for about half a decade. That would be Centos. Before that it was RHE. Before that it was RedHat 9.

Desktop distro pissing matches are irrelevant. I have bills to pay. Any desktop/OS combo that provides SSH and a modern graphical browser is adequate for my business needs.

Re:Most important

[Lennie]

Coincidentally many, many run Debian servers.

I personally think Debian is a server operating system first and a good base for desktop distros second.

Re:Most important

[Aighearach]

And for many many geeks around here, half a decade before that. And they were the new kids then!

Re:Android second?

[dreamchaser]

Yes, it is Linux. Just because you say it's not doesn't make it so. Just because it's become a fork, for now at least, doesn't make anything you've said valid. It runs a Linux kernel, modified yes but still Linux.

Re:No, it's not

[chip_s_ahoy]

I think that most of what you write is lies.

Root Prompt

[codepunk]

I could care less what it is, debian, centos, solaris, aix etc etc etc, just give me a root prompt and I am at home. Oh and just for the record no, I don't give a damn about you sudo troopers.

Re:Root Prompt

[bmo]

Protip:

You can change it so you can login as root. It's not a good idea from a security standpoint, but you can. It takes about 5 seconds if you're a slow typist.

Also sudo -i will get you a root shell so you don't have to keep refreshing root privs with a password after they expire.

--
BMO

Re:No, it's not

[xnpu]

Not sure what you use your servers for, but for our purposes Debian is more than adequate. Very few applications really require yesterdays cutting edge release of whatever software. The slower release cycle of Debian provides us with a much appreciated stability. Security fixes are much more manageable than entire release updates. Meanwhile, the long "testing" period Debian goes through allows us to test properly test and prepare a platform-wide (read thousands of servers) upgrade and ensure it doesn't cause any issues, which it rarely does.

So then, should i just set my desktop with debian

[unity100]

then ? its currently ubuntu. dual boot with xp.

Totally off base

[vga_init]

Debian isn't even in the ballpark. The most important linux is FreeBSD, but nobody wants to admit it.

Re:Totally off base

[Lennie]

Let's see, they create new drivers which are used in the other BSD and the Linux kernel, they are used as a base for Junos for the Juniper routers. I guess FreeBSD like the Linux kernel is used for TCP/IP-research.

But OpenBSD maintains OpenSSH which is used in pretty much all Unix-like environment and embedded system like routers and managed switches.

I still think it makes Debian interresting because they adopted FreeBSD as their second kernel, having the same environment on all systems is very useful.

Re:Totally off base

[Junta]

We'll see Netcraft confirms...

Re:Totally off base

[micheas]

I suspect that debian is going to make FreeBSD a lot better for FreeBSD users.

Debian does a lot of work making sure that all of software works on all the architectures that it supports

Mozilla claims that Firefox runs on Linux, but debian had to jump through hoops to get Firefox to compile, much less run on the MIPS platform.

By making Kfreebsd a first class platform, a lot of fixes for FreeBSD should make it upstream, which should improve the quality of the software in the ports tree.

The big contribution debian makes is debian policy and the QA on all the architectures that it supports.

Some of the billion respins are probably interesting, but the copyright fights, and the code improvements to support cross compiling are things that leak into other distributions. Debian was one of the reasons that AMD64 support is as good as it is under Linux. The Redhat, gentoo, and slackware users that use the 64bit versions are benefiting from Debian getting their distribution to run on 64 bit platforms for years.

Personally, I think the title of the article is true, but that the article provided no evidence about all the contributions that Debian brings to Linux users and just argued "it's the parent" Which if that is the case, BSD386 is the most important OS, as it is in many ways the ancestor of Solaris, OSX, FreeBSD, NetBSD, OpenBSD, The GNU project and all the GNU systems. In other words, everything other than Windows. And Windows has some FTP and telnet code from BSD386 in it, and at one time the windows TCP/IP stack was based on the BSD386 network stack.

Gentoo, Redhat, SuSE, Slackware, Canonical and others contribute in ways that help build the Linux ecosystem, but it is hard to overstate Debian's importance to the ecosystem by being an large, neutral, cross platform distribution, that is very transparent. Unfortunately it is possible to completely miss this and ramble on and on about the number of respins that exist.

Re:Totally off base

[carnalforge]

Must of what you said is true and fair, but i call bullshit on this:

Debian was one of the reasons that AMD64 support is as good as it is under Linux. The Redhat, gentoo, and slackware users that use the 64bit versions are benefiting from Debian getting their distribution to run on 64 bit platforms for years.

FWIK it was Andi Kleen from SuSE to port linux's kernel on AMD64. Quick googling:

http://www.x86-64.org/pipermail/announce/2001-June/000020.html

Re:Totally off base

[TBBle]

The grandparent post is not talking the kernel port, it's talking about the applications and libraries. And not just AMD64, but the earlier 64-bit ports as well (Sparc, MIPS, IA-64 and PowerPC64 all predate AMD64, although I haven't checked whether they were supported by Debian before the AMD64 port. I suspect Sparc 64-bit was supported much earlier...)

Which is the same benefit it discusses bringing to FreeBSD: Getting all those user-space applications and libraries working properly under the FreeBSD kernel, where many of them will have been developing under a Linux-only environment, and merely _hoping_ they worked under other kernels.

Re:Totally off base

[micheas]

I said "one of the reasons"

The kernel is a small, but vitial, part of what you need to have a usable desktop.

Andi Kleen got the kernel ported. The paper you link to states that the toolchain GCC and the kernel work, and you can use them to port things to linux on AMD64. That is about a dozen of the 30,000 packages that are in Debians AMD64 distribution.

A lot of user land apps were written with the assumption that they were running on a 32 bit platform, especially graphics and multimedia programs.

Imagine how much less 64bit mode usage there would be if xorg, xfree86, php, perl, python, and java did not run in 64bit mode on AMD64 machines.

Re:So...

[Dracos]

If "most popular" == "best", than Debian is "best" in the same way that NT4 is the "best" Windows, I guess: NT4 and its derivatives are the most widely used.

Popularity does not necessarily equate to quality, though.

Maemo

[Weezul]

It also suggests that Nokia and Intel were idiots for switching their mobile Linux distributions over to RedHat based systems.

Re:Maemo

[Pharmboy]

I don't get it. RH is a very accepted standard, whether or not it is more or less "influential" than Debian. RH isn't fly by night, has been around as a commercial product for many, many years, has the backing of the parent corporation, which is actually profitable. The CentOS project was born from it, which makes the RH system completely free for use by people such as myself.

And being used by a number of other distros has nothing to do with market penetration or quality. RH (and CentOS) haven't been forked to death because they already do exactly what many businesses want: provide a steady and reliable server platform. This isn't nearly as sexy or cutting edge as desktops for a reason. I don't want my Linux web server to be "cutting edge", I want it up 99.99% of the time.

RH isn't the only choice, and isn't always the best choice, but it is certainly a viable choice for many applications.

Re:Maemo

[synthespian]

And being used by a number of other distros has nothing to do with market penetration or quality

Exactly. No real innovation has come out of Debian. In fact, their situation was so messed up, they needed a millionaire do-gooder to sort their mess up.

From the technical standpoint, Red Hat is the distro that advanced Linux the most. That's a fact.

Re:% of distributions vs % of user installations

[Lennie]

Based on the number of visitors to Wikipedia, over 50% of all Linux desktop users use Ubuntu:

http://stats.wikimedia.org/archive/squid_reports/2011-01/SquidReportOperatingSystems.htm

But more visitors use Android than Ubuntu.

And more people use Ubuntu than MacOS on PowerPC.

Many people run the Ubuntu LTS-version, but more run the latest stable.

Re:No, it's not

[Lennie]

It is not so much dpkg that is important, it is the quality of well it is packaged. This is partly because of the long release cycle, all possible conflicts and problems will be found because many people use testing and update frequently. Even though Debian has (I think ?) the largest number of packages (maybe I should source packages).

Re:Android second?

[Urkki]

Ok, I stand corrected, and thanks, have to try that out!

Re:Really?

[Lennie]

I don't think most of the Debian derived distros are forks. Just look at Ubuntu, they take Debian as their new base every few years and incorporate most of the 'community supported' packages directly from Debian.

Simple example clamav or something on Ubuntu 10.10 and look at the changelog.Debian.gz file which tells you who did what changes in Ubuntu or Debian:

Wed, 01 Dec 2010 16:46:37 +0100 -> change by Debian maintainer
Thu, 02 Dec 2010 21:04:10 -0500 -> change by Ubuntu maintainer: merge from Debian
Fri, 17 Dec 2010 10:49:56 -0500 -> change by Ubuntu maintainer: Cherry pick [git-id] from Debian
Wed, 23 Feb 2011 14:27:51 -0500 -> change by Ubuntu mainter: an Ubuntu change
Wed, 23 Feb 2011 14:27:51 -0500 -> change by Ubuntu mainter: security fix

That is pretty clear. But you say that is not very desktop orientated ?

Let's take the xserver-xorg-core-package:

Ubuntu took the package from Debian experimental on 17 July 2010 and did their changes on it.

And they will take the original Debian package again every few years, mostly 'just' after the LTS release.

Re:Android second?

[WaywardGeek]

Comparing Debian to Android is actually very interesting. Debian has something like 32,000 packages that can be installed, but it's taken something like 15 years to get there. Android blasted to over 100,000 in something like 2 to 3 years. Debian is all about community contribution, while Android is all about selling closed-source apps, with no sharing of code between. In theory, it should be easier to publish an app in Debian than Android, but this is not the case at all. In Debian, you have to find a sponsor, do a complicated job of packaging, pray your package gets uploaded to Unstable, and then wait a few years while it migrates to Stable before other programmers will generally have access to your work. I call this the Debian Red Tape. It's suffocating innovation in the open-source community, and it's the reason Android is kicking Debian butt.

I believe there is a solution, but it requires a completely new packaging system. Let's compare Android and Debian packaging:

- Android ships every dependent binary in the .apk app file. This eliminates the nightmare of having your app crash because some library you use gets updated.
- Debian is all about resusing .so files across applications. This made sense in the '90s when disk space was scarce, but now days, it's just dumb. The reason it takes years to get a packaged library into Debian Stable is that it takes years before we believe you library wont cause other apps to crash.

A new packaging system could share identical binaries between apps to save both disk and memory space, but it should not ever change a binary used by an app. Also, publishing new packages should be as easy as creating a repo on github.net. You simply declare that it's available, and everyone can use it. Whether a developer decides to depend on your code should be a matter of trust, which could be scored based on developer reputation, code stability and what other packages use it.

Without a major upgrade to our packaging system, Debian will continue to fall further and further behind. Why do so many people feel they have to build a custom Debian based distro? Because Debian incapable of addressing the needs of modern users. Frankly, even with the total lack of libraries available for Android, and with Google having their heads up there arse with respect to accepting contributions from the community, I am able to contribute more to Android than I can to Debian. Check out my library that I've made available to both at dev.vinux-project.org/sonic. I'm basically done for Android, while I'm still waiting for a Debian sponsor over in Debian land.

Wrong conclusion

[degeneratemonkey]

An alternative interpretation of these statistics is that a lot of geeks like parts of Debian and hate other parts of it, ultimately leading to a higher volume of forks. DRTFA. In fact I rarely RTFA these days, because 90% of /. stories are just bloggers trolling for hits. Not even good ones.

GNU/Linux

[tepples]

In everyday usage the word Linux refers to the whole OS. And by that we mean the kernel, GNU stuff, (sometimes also X11 and whatnot). In light of that, Android is not Linux, even if it technically is.

Which might even validate the point of the people who insist that Linux distributions similar to desktop Linux be called GNU/Linux. This would at least serve to distinguish "GNU/Linux" on N900 from "embedded Linux" on Android phones.

Re:GNU/Linux

[spitzak]

This would also place almost all embedded linux (ie routers and storage devices) in the same catagory as the Android phones. They don't have the GNU utilities and instead run busybox. This would be useful to remove a large class of devices that are not normally considered when asking how many users run Linux verses Windows (there is also embedded Windows but it seems that is mostly limited to ATM machines).

Some embedded boxes may be running a little GNU software, in particular glibc.

Also it is quite possible to remove all GNU software from Linux (using the BSD utilities instead) and make it run X, and the average person would consider that a "Linux machine".

So I think a better distinction is X/Linux, in that it uses the X window system.

An annoyance is that X may be replaced by Wayland, yet this should still count. X should just mean "window system".

Re:GNU/Linux

[tepples]

This would also place almost all embedded linux (ie routers and storage devices) in the same catagory as the Android phones.

Which was sort of my intent, as Bionic fills much the same role as Newlib or uClibc. People see their phones as appliances and their home routers as appliances.

Also it is quite possible to remove all GNU software from Linux (using the BSD utilities instead) and make it run X

Is there such a distribution of, say, FreeBSD running on Linux? I tried Google linux bsd userland but didn't find any active projects.

On evolution

[sean.peters]

If a Distro has a lot of descendents is because it's evolutive succesfull, just like living beings, if that isn't a measure of it's worthiness, what else???? Mother nature seems to use the same standard to judge their offsprings .Mind you, you're a primate and not a lizard and a long time ago we were tiny, scarcely populated and by every measure the less important animal in the world.

By this measure, beetles are way, way more important than all mammals put together. I agree with the top of this thread - number of descendant distros is not that interesting of a measure.

(Of course, measuring by "installed base", bacteria beat every other species on earth... combined. All of which is to say that maybe the number of evolved species ain't such a good metaphor for linux distros.)

Fail

[jvillain]

Debian isn't a "Linux". Debian is a distribution of GNU licened and other software that happens to ship and use a Linux kernel. But it also does that for one of the kernel for one of the BSDs as well. As for the statistics. I couldn't think of a worst set of them to try and make a case of greatness with.

Slackware is Older

[Lawrence_Bird]

by a month so you can remove that from the summary, not that the whole article isn't rather pointless. Distros come and go. If "Debian" were to vanish tomorrow another would take its place. And lets be honest here - repackaging other software and doing custom configurations of the same isn't exactly rocket science. A PITA maybe but groundshaking? Influential? No. Which is exactly why so many distros have come and gone since 1993. When a perceived need arises which is not yet met individuals coalese into groups which develop a new or better (forked) product. The article uses the fact that there are derivatives of Debian to claim it is influential but isn't it the case that Debian is lacking in some way(s) such that others are willing to go to the trouble to do something different?

This isn't meant to be a negative critique of Debian at all - I just find it joke to say it or any other distribution is the most "influential" having been around since the days of Yggdrasil.

Re:maybe it's just not so good

[supremebob]

A base install of CentOS 5 (no GUI or packages installed, just the kernel and a few utilities) is about 400 MB. From, there you can use yum to customize it to your liking.

Hell... I've done usable kiosk setups with XWindows and Firefox using CentOS 5 that ran off of a 4 GB CompactFlash drive.

Re:Android second?

[Stupendoussteve]

You're forgetting that out of those thousands of Android apps, most of them are built for Dalvik, not "Linux." There is a project porting Dalvik to FreeBSD, will it still be Linux after that? There are far more Debian packages _for Linux_ than there are for Android, and there are no Debian packages for Dalvik that I know of.

The reason so many people are making custom Debian distros is because Debian is an extremely stable base with a pretty nice package manager, and they want to customize it. That's it. I would not say it is "falling behind" at all, the sheer number of distributions based on it seems to rule that out.

Android is primarily used in cell phones of which there are millions. That is the only reason for its market share. It's not kicking anything's butt on the PC.

Re:Android second?

[tao]

- Android ships every dependent binary in the .apk app file. This eliminates the nightmare of having your app crash because some library you use gets updated.
- Debian is all about reusing .so files across applications. This made sense in the '90s when disk space was scarce, but now days, it's just dumb. The reason it takes years to get a packaged library into Debian Stable is that it takes years before we believe you library wont cause other apps to crash.

Shared libraries still makes just as much sense. Packages that include their own copies of libraries (or for that matter are statically built) are a serious seurity issue. It's mostly fine for applications that can be sandboxed, but mostly it's a security nightmare. A library change that causes applications to crash either indicates incorrect assumptions made in the application (possibly due to a poorly documented library), or an ABI-break in the library -- the latter should always trigger a bumped SO-name.

Part of the reason that it takes a long time for Debian to release is the size of the archives, part of it is because in a lot of cases we do quality assurance on behalf of application writers who believe that embedded library copies is the greatest thing since sliced bread.

Do these numbers really prove what they claim?

[Kenneth Stephen]

Debian would be the most important if the most number of users depended on it. If the 63% of distributions that depended on Debian had a total base of 5%, then it would be ridiculous to claim that Debian is the most important. This number is missing. Maybe the user base for Debian and its derived distributions is a lot higher and Debian is indeed the most important distro out there. This article certainly doesn't make its case for that statement.

Two comments

[snookiex]

1. Debian won two of seven categories at the Linux New Media Awards 2011
2. While the article focuses on Debian with the Linux kernel, it also supports a FreeBSD kernel and (I don't know if this actually counts) GNU/Hurd

Re:Android second?

[reub2000]

Well to the end user, it's not linux any more than your dvd player is. Native linux apps won't run.

Up Yours Debane

[steak]

I have never liked Debian, for whatever reason it always rubbed the wrong way. Slackware on the other hand rubs me the right way.

Re:Android second?

[Colin Smith]

"Debian is all about resusing .so files across applications. This made sense in the '90s when disk space was scarce, but now days, it's just dumb."

No it isn't, it's about RAM, I/O performance, security updates, bug fixes and binary drift over time.

Debian packages libraries as soon as there is a demand for them. No interest, nobody cares.
 

primary term

[epine]

Everyone knows deep down that "most important of" originates from the part of the brain responsible for mating behaviours and penis size comparisons. People are attracted to the dialog in the aspiration to become one of the lucky lekkers. And even if the lek has nubility factor zero, it's good practice just in case if your prospects are poor and you have nothing better to do.

At a certain age, you tire of the loud clatter of penis rulers and you just want to hand the participants a scalpel and a bassinet labelled "least important" to find out whether they really believe that every infinite series can be approximated (for the purpose of getting laid) by the first term alone.

I sometimes wonder if the donning the coat of arms of truncated approximation functions as a sexual status symbol. If the well runs deep, no need to bother with second order effects; leave those worries to the mincing greybeards whose primary term has shrivelled up.

For me the miracle of conception is how quickly the brain reprises all those forgotten terms, if there were any in there to begin with.

Congratulations to the person posting this story for telling the world that your swimmers have yet to enter the pool.

Yep.

[mrflash818]

I agree that Debian is the most important GNU/Linux distribution.

Fedora & redhat

[Billly Gates]

It's strength is in the computer room and many .com websites.

Redhat made Linux popular. In the server room, it is a different story than a bunch of hobbyists and computer science students fiddling around on their laptops with Ubuntu.

Redhat Enterprise is HUGE for business critical server apps. Debian is there too but not as big. Business users need support and well tested QA. Sure many slashdoters have bad memories of RPMs (myself included) but Yum takes care of the problems today. Infact Yum and RPMS are more robust and professional than apt-get and deb files as a bad installation on a debian based system will leave changed config files and other files scattered across a system. An RPM wont do that which is important for business users as well.

Redhat funds more software than Canonical does including Gnome, Gnu, and Java. If you love your Ubuntu experience you can thank red hat for funding and donating code for major potions of your gui.

Re:Fedora & redhat

[micheas]

It is hard to say whether Debian or Redhat has done more for users of other Linux distributions.

The reason why it is hard is because the big contributions they have made that are not not directly comparable.

While both of them spend time cleaning up software licenses, Debian has spent a lot more time on it, and essentially came up with the working definition of open source/free software.

Debian's clean separation of Free and non-Free software helps the entire community identify what software is a potential issue and may need relicensed/rewritten.

Redhat has done orders of magnitude more work getting linux non x86 architectures. Debian has done orders of magnitude more work getting optional userland programs to actually run on architectures other than x86/AMD64.

RPM is more powerful than dpkg, but that is because a lot of the features of RPM are not needed if you are installing debian policy compliant software on on a debian system.

Debian's policy has resulted in a fair amount of software becoming much easier to upgrade. Mixing data and code is not only bad for websites.

Redhat has pushed SELinux into being usable in certain use cases.

Redhat pays for more development of Open Source Software than anyone else.

Debian has more people testing software in a developer capacity than anyone else.

A panel discussion among people that can collectively identify the marjority of what RHAT and Debian have contributed to the ecosystem would be an interesting event. (at lest to dweebs like myself.)

One could make similar arguments about Oracle America(formerly SUN) and Novel/SuSE. for their contributions to LibreOffice and Gnome. However the Linux desktop seems to have not really reached the tipping point, while Linux is the dominant platform for servers.

Which editor

[Billly Gates]

I want to know which is the most important editor to use for the most important distro. Vi or emacs?

Re:Which editor

[neminem]

Nano!

Re:Android second?

[WaywardGeek]

I'm afraid I have to disagree with the replies above. While obviously not immune to security problems, Android sand-boxing seems like a good step, enough that keeping older binaries around makes sense, IMO. I can't believe anyone thinks is an RAM issue. I mean, my phone seems to handle it just fine. Also, we can still share identical binaries, so even popular libraries are only likely to be duplicated at most a few times on any system.

So, instead of having to live with the nightmare of all those developers out there who statically link, and all those other developers who change their ABI without bumping the .so name, and instead of testing for literally years before allowing a new app or library into Debian Stable, why not modernize the packaging system, and run apps sand-boxed? The strength of Debian's original packaging system is what made Debian great, in addition to it's community development. However, now days it's very hard to contribute to Debian, and the packaging system is 90's vintage, and out of date.

I contribute a lot to the Vinux distro, which is Linux for people with vision impairments, currently based on Ubuntu. One reason for branching off Ubuntu is it has more of the latest and greatest innovations, while Debian Stable is a couple years out of date. We try to just ship the packages we change, and have users download most binaries from Ubuntu. The result is that we're unable to modify core packages, like GTK+, without breaking everything, because the ABI versioning stuff just doesn't work. In reality, you have to recompile everything if GTK+ changes. Because of the fragile nature of Debian shared binaries, we avoid fixing low level things. I have GTK+ accessibility fixes that are being abandoned. It's just not possible in most cases for a single contributor to offer improvements to Debian in a meaningful way anymore. There's too much red tape.

So, if you just want a firewall or file server, go with Debian. If you want modern touch-interface tablet software, look elsewhere, because Debian will never make the transition. It's packaging system and it's friction to new software is dragging Debian into the dust-bin of history.

Re:Android second?

[perlchild]

What you call suffocating innovation, I call code quality.

Or do you expect someone to package xorg just by creating a github repo, and then gnome, xfce and kde will all "just work"
Perhaps this will clarify my point further.
apt-get dist-upgrade working from debian 2.0 to 3.0 on a live, production server(in a maintenance window), the first time, is a metric of package quality for me.
No other linux has had it, not for major version numbers, imho.
If it doesn't work this good, I call it not working(I'm looking at both Fedora and RedHat here)

You're free to disagree
I don't think the article has any meaning, myself, they're all important, for different reasons, to different people.
And how many derivatives isn't as important as what they do(for myself ubuntu and knoppix are pretty important, in terms of what they enable/do, others have different needs, like oracle-on-linux for instance, haven't checked the debian state of support for it recently)

Re:Android second?

[Risen888]

In theory, it should be easier to publish an app in Debian than Android, but this is not the case at all.

Why on earth should that be the case?

I call this the Debian Red Tape. It's suffocating innovation in the open-source community

Obviously it is not.

and it's the reason Android is kicking Debian butt.

At what?

I believe there is a solution

To what?

but it requires a completely new packaging system.

Hoo boy, here we go. Doesn't it always?

publishing new packages should be as easy as creating a repo on github.net. You simply declare that it's available, and everyone can use it. Whether a developer decides to depend on your code should be a matter of trust, which could be scored based on developer reputation, code stability and what other packages use it.

That's brilliant! So it should be just like...Debian? Waitaminute...

Without a major upgrade to our packaging system, Debian will continue to fall further and further behind.

Behind whom? At what?

Why do so many people feel they have to build a custom Debian based distro?

Oooh, I know this one! Because Debian provides the simplest, most robust base upon which to build a distro!

Because Debian incapable of addressing the needs of modern users.

Um. No. Actually, you're not even close.

Frankly, even with the total lack of libraries available for Android, and with Google having their heads up there arse with respect to accepting contributions from the community, I am able to contribute more to Android than I can to Debian. Check out my library that I've made available to both at dev.vinux-project.org/sonic [vinux-project.org]. I'm basically done for Android, while I'm still waiting for a Debian sponsor over in Debian land.

Oh, so this is all about you being pissy that your package hasn't made it into Debian? Shit, you should have said so, I wouldn't have had to waste my time on this.

Re:Android second?

[WaywardGeek]

Debian packages libraries as soon as there is a demand for them. No interest, nobody cares.

Bull. Debian allows packages to be included once a Debian sponsor is interested. It doesn't matter how many users want a package. For example, I've got a package of significant interest to blind Linux users, which is still waiting for a sponsor. It doesn't matter that there are 200-ish Vinux users who want it in there. What matters is that there aren't blind Debian sponsors. At Ubuntu, there's a guy with a vision impairment, probably similar to me, who uses Orac and some magnification. He cares about this stuff, and that's why Ubuntu is advancing faster in accessibility, even though Canonical doesn't do accessibility testing in their standard flow.

If you want to talk about red tape, let's talk some more about Debian.

Re:Android second?

[WaywardGeek]

What you call suffocating innovation, I call code quality.

I guess in that case you wont be interested in any of those many thousands of low quality apps available on Android or iOS. I agree that it is amazing that you can simply dist-upgrade in Debian. It's very impressive. But... to get there, Debian has had to introduce multiple years of delay to get a non-super-hot package from Unstable to Stable. There are "sponsors" who are the gate keepers, rather than letting users create apps for other users freely. "There's an app for that", is not going to be said about Debian anytime soon.

Re:Android second?

[WaywardGeek]

That rant doesn't really deserve a response, but I'm sitting here with a glass of wine with some time to kill...

In theory, it should be easier to publish an app in Debian than Android, but this is not the case at all.

Why on earth should that be the case?

Because we're all sharing our code freely with each other, rather than keeping it secret like most Android developers. All of that sharing should make it easier to build and share apps, but we've bunged it up and made it super-hard.

Without a major upgrade to our packaging system, Dy

Behind whom? At what?

It's falling behind Android and iOS. Look, we probably both like the tech under the hood in Debian, but random users want apps that work for them. I've got shared grocery lists, shared to-do lists, a geo-caching app that rocks, and a hundred other apps I actually use all the time. My Nexus One is the best productivity tool I've experienced since the adm3a terminal. Try getting any apps like that into Debian. The future of personal computing will be devices using speech recognition, multi-touch, and docks for tablets. Debian is awesome in the server farm, but if you just want to write and share an app, frankly Debian sucks.

Re:No, it's not

[synthespian]

"Stable" in Debian does not mean the same as, say, when FreeBSD says "stable".

"Stable" in Debian means "stalled."

Re:always hated "distro" anyway

[TBBle]

Maybe an Australian? It's a common form of abbreviation in Australian English to replace -*ion with -o. Also replacing things that aren't -*ion with -o:

e.g., Registration (for motor vehicles): Rego
Bottle shop (off-site alcohol vendor): Bottle-o

I'm sorry if our language sounds moronic to you. I'll try to avoid speaking to you in the future...

Re:Android second?

[Risen888]

Because we're all sharing our code freely with each other, rather than keeping it secret like most Android developers. All of that sharing should make it easier to build and share apps...

And it obviously does. Look at the ecosystem that exists around Debian. That's all "building and sharing," and Debian makes that possible by making it easy. You seem to think easy means "They should accept my code." That's not what it means. Easy means "They make it easy for me to base my work on it." And it does. Any idiot can set up their own Debian repository. You could too (which you don't seem to have). For God's sake, there's not even a .deb on your download page.

It's falling behind Android and iOS.

At what? What exactly is it you think Debian's trying to do?

WTF?

[hotfireball]

Some more useless article about "Who has longer dick". And the answer is: Dubai.

Debian most likely to inspire change

[jkinney3]

I read the article not as Debian is the most important distro but rather is Debian is the most likely to cause people to say "I can do better than that" and then expend the energy to do so. The RedHat derivatives are far fewer in numbers because RH does such a good job of covering all the bases people want to see in a distro. That is to say, RedHat users are more satisfied with the efforts of RedHat than Debian users are with the efforts of the Debian community.
That said, Debian is an outstanding community distro whereas RedHat is an outstanding enterprise distro. Use each appropriately for best effect.

Does not mean what you think

[O'Nazareth]

I guess the reasons for Debian to be at the origin of birth of more other distributions is because: 1. they have a good package manager (this is a good side) 2. they are very very conservative, which make people run away (the bad side). Now, the reasons for Gentoo not to spawn more distributions: 1. it is hard and to install (so not that many new users) 2. you do not need to, actually each installation is its own distribution. What was counted as distribution are the ones with a live CD.