Time for a Linux Consolidation?

An anonymous reader writes "Are there too many Linux distributions currently available? Can there be too many? This article explores the effect of the large number of distros out right now and suggests that progress could possibly be made through a consolidation. The article is more focused on Linux on the desktop but the ideas presented would impact the entire community, especially as it is seen as a rival to Windows." From the article: "One of the less widely recognized reasons why Linux has not yet toppled Windows, despite it many advantages, is how divided the resources available to Linux are. With dozen of different distributions the Linux community is so diffuse that the power or significance of any specific entity is severally limited."

Comments

You mean like...

[ylikone]

... the success called "United Linux"??

Re:You mean like...

[Synbiosis]

They're really all the same, minus some extra shiny icons.

And the drivers included. And the (non-shell) user interface. And the installation/packaging system.

Linux could definitely benefit from *some* consolidation.

Re:You mean like...

[It'sYerMam]

But there are huge benefits to be gained in usability with a consolidated packaging system. Think what it means - projects need to be ported to ONE distribution system. Proprietary drivers need support ONE more system, make sure it works with ONE system. Users would be nearly garunteed to be able to find the program they want if there were consolidated repositories. The repos could be divided a la debian into stable, free, nonfree, testing, unstable - although preferably in my view another that's in between testing and unstable. Thataway every distro could be as stable as debian, as bleeding edge as fedora and whatehaveyou. Also preferably organised by flags as opposed to directories, thatway you can have "nonfree, stable."
You've gotta admit that the massive web of differing packagin systems is often high up on the list of criticisms against usability, and if there were consolidated repos, then you needn't worry about getting software from dodgy sites as most stuff would be right there in repo.

Re:You mean like...

[TCM]

Every distro out there is an experiment in what works and what doesn't.

Exactly that's why I can choose the network protocol to reach Slashdot, right? Oh, has to be TCP/IP you say? Where is my choice?!

Next you tell me I have to use HTTP to read comments and can't pick something I want..

It's about choice, right?

Right?

*crickets*

Re:You mean like...

[cortana]

It *would* be much better if there was an NNTP interface to reading Slashdot stories and comments.

I got it!

[captnitro]

We should form a committee to consolidate and leverage synergy for our information technology solutions in the marketplace!

THIS IS A PERFECT OPPORTUNITY FOR BUREAUCRACY!!1

Look, the fact is

[Anonymous Coward]

Linux is never going to reduce the number of available distributions, even if it's what's good for Linux, because the people making these distributions aren't doing it for the benefit of "Linux". They're doing it for the benefit of themselves.

It's all very well and good to be some kind of columnist, standing outside of Linux and going "well, Linux would be better if Slackware and Gentoo would combine". That's easy to say. But this doesn't help you much if you're a Slackware user; it might be better for Linux if that happened but it wouldn't be better for Slackware and to the Slackware developer, what's better for Slackware is what matters because Slackware is what they want to use. If it wasn't, they'd be using Gentoo instead in the first place.

Linux development, as an open source process, is fueled by self-interest. This is its greatest strength. That it indirectly produces weaknesses is unavoidable.

Maybe Not

[the_weasel]

I am a windows apologist - look at my history and you will see I entirely willing to point out the failings of Linux to the Zealots as the next guy.

But even I can see that the diversity of Linux is one of its strengths, as well as its weakness right now. Thanks to the sheer variety of work done in exploring slightly different approaches to the same task, we get to experiment with a multitude of approaches and ideas.

While that may not be a truly better product now, it can only lead to an excellent one in the future.

I am in no hurry for Linux to take over - I am not even sure that the operating system that does take over will be called Linux. Windows will have to sink a lot lower before its abandoned by the masses.

I am entirely certain that the work done in Linux over the past 10 years will shape the next generation operating system that finally does defeat windows though.

Re:Maybe Not

[The_Mystic_For_Real]

One of the reasons that the number of Linux distros is impairing its growth is that it hasn't presented any one competitor to Windows. If people want Windows, they get Windows. If they want to try Linux, should they use Gentoo, Slackware, Knoppix? Which one supports their hardware better, which one is more user friendly, which one is more secure, which one has a helpful userbase? It's not really easy to answer any of these questions, even for those experienced with Linux. Furthermore, there is no face to Linux (no offense Tux), it becomes something that occasionally goes from one to two percent of the market for much of the other 98%.

Re:Maybe Not

[jbolden]

Which should they use to do what? To:

1) Create a development workstation
2) Run a small home server
3) Run a large multiprocessor corporate server
4) Use for children
5) Use for a keyboardless sales demo on a plasma screen
6) Use for a cash register
7) Use for an embedded system with no user interface at all

etc... Why would you expect one product to be all of those things. The car industry works fine with having a range of products from small compact cars to large trucks. People in grocery stores can shop for steak or cereal fine. Its Microsoft that has pushed a "one size fits all" approach.

Re:Maybe Not

[bhiestand]

4) Use for children

Good job. You managed to pick ONE thing the average home user would actually use a computer for. I guess you just don't realize that the average home user doesn't know the difference between spyware and a virus, has no interest in networking their coffemaker and their fridge (with encryption, of course), and certainly doesn't care for a server.

Here's what they want:

1) so easy to use, even they can do it

2) check their spam daily

3) all of the programs they already know how to use already work on it.

4) easy transition (minimize lost settings, files, etc.) This includes things such as iTunes playlists.

5) the same hardware support they have in windows. That is, everything worked with windows when they bought that computer from Dell/Compaq/HP/Fujitsu, whatever. They didn't have to go hunt down drivers, troubleshoot anything, or settle irq conflicts. In fact, they're not quite sure what a driver is, but are pretty certain about putters.

6) multimedia support out of the box. They don't want to install their Line-Ux CD just to find out that the media player can't even play their porn or their mp3s. They need codecs? "What are those? Where can they buy them? RTFA? Where can I buy an RTFA? Oh, you mean check the interweb? Google said I should download winamp!"

I really have no idea why I haven't run into a single linux distro that supports my porn and mp3s without extra codec downloads. I'm not sure why most STILL don't include read-only support for my NTFS partitions. To be honest, I'm amazed they don't have it fully figured out such that I can WRITE to my NTFS partitions as well. Especially since I've got some nifty 300GB external hard drives for my windows laptop. Why is it a windows laptop? See the above issues.

I realize most of the problems are really caused by crappy licensing agreements by the "owners" of whatever licenses or code, and a lot of the other problems are caused by hardware manufacturers refusing to release drivers or help us to write them, but the end user doesn't really care. The average 50 year old jet mechanic couldn't give a damn if it's because the software author's daughter died yesterday. He wants his shit to work to do the stuff HE does. And most people do the EXACT same thing and little else.

As for me, I'll switch entirely, and get the rest of my family running it, as soon as all the porn that windows media player plays is playable under linux, red alert 2 works perfectly, and gaim has caught up with trillian. The browser is about the same, evolution and thunderbird are great for email, the hardware support appears half-assed at best, .... I'm done with that.

To bring back to your point: "The car industry works fine with having a range of products from small compact cars to large trucks" This is a valid point, but also completely off. The true analogy would be going into a car dealership and telling them:

Buyer: "I just want a car. It has to have A/C, working lights and brakes, and get me to and from work."

Dealer: "Will you be hauling sand or rocks. What kind of towing capacity do you need?"

Buyer repeats: "just a car. For transportation. To and from work. And maybe the movies occasionally. I'd like to be able to get it on in the back seat too".

Dealer: "Oooooh. Ok, yeah, we have that. Here's what we can do for you. We've got this great frame out back, it's really happening. Then you can go across the street and buy the interior. We'll even tow it down the street to the corner for you to get the engine put in. You can repaint it whatever color you want (fully customizable!), but nobody has lights for it, so you'll have to find those before you drive it at night. Oh, and it needs 100 octane, so you'll have to go to one of these special petrol stations."

Linux's problem is your idea of "why would you expect one product to be all of those things." The truth is, the majo

Yes.

[mrchaotica]

I think there are too many distros, but moreover, I think there are too many competing technologies: QT vs. GTK, dpkg vs. rpm vs. ebuild vs. tgz, etc. If we could work out some good standards -- that everybody followed -- we could have all the distros anyone wanted and it wouldn't be a problem.

Re:Yes.

[slavemowgli]

"Pulling together is the aim of despotism and tyranny. Free men pull in all kinds of directions." - Terry Pratchett, "The Truth".

It just doesn't work that way. Most FOSS software is not developed because somebody has a "vision" of "taking over the market" or something similar; it's done because people want to scratch their own particular itches. This is a weakness of FOSS, if you view it from a marketing or management point of view, but it's also the key strength.

As for Linux taking over Windows and toppling M$... maybe it'll happen, and sure, if it does, that'll be a day to celebrate. But it's not the principal *aim* - Linus himself has repeatedly said that his goal is not to write an OS that is better than competitor X, Y or Z, but simply to write the best OS he can.

Of course, on top of the actual developer/user community, there is another layer - namely, the companies that produce distros, like Novell (SuSE), RedHat, Linspire (or Lindows or whatever they're called this week) and so on. And yes, if you view it from their perspective, then it very much *is* about beating Windows (as far as possible, anyway), and also about taking as big a slice of the Linux cake for your *own* particular company - this is one reason why distros are different and why UnitedLinux never took off.

Of course, the same principle applies everywhere in the IT world. And of course, like everywhere else, standards still are important, because even though distros can try to set themselves apart by being being better than their competitors, they will ultimately fail if they're too different from everyone else - in that case, the majority simply has more weight, even if the minority distro is better in technical terms.

This is a kind of conundrum that you'll encounter everywhere in the computer industry: standards and interoperability are good, not just for customers and users but also for companies that develop software, but if you have *too* much, then products will start to be completely interchangable, and that's something that companies will always try to prevent.

But to get back to my original point, that's just something that happens in the business layer on top of the actual FOSS communit(y|ies), and it doesn't matter to Linux as such - because, ultimately, FOSS is not developed so companies can build new business models around it, but rather so that users will be able to scratch their own itches in the best possible way.

That's what it's about: the software. To stay with one of your examples, some people like Qt, some prefer GTK+ (for whatever reasons), but there is no reason why there shouldn't be both. It may be bad for a business trying to maximize its profits, yes. But why should the community members care?

It doesn't make a difference to us whether Joe Sixpack uses Linux or Windows. Why should we give up our choice so it's easier for a company to sell a Linux distro to him?

Re:Yes.

[Brandybuck]

"Pulling together is the aim of despotism and tyranny. Free men pull in all kinds of directions." - Terry Pratchett, "The Truth".

Awesome quote! The problem with the Free Software community is that too few people understand what "free" means. Everytime someone says we need consolidation, or that KDE and GNOME need to merge, or that we have too many text editors and package managers, all they are doing is admitting that they don't know the first thing about freedom.

Freedom is messy. It's uncomfortable. It doesn't hold your hand. It doesn't make any promises. Sometimes we need to grow up and stop depending on mommy and daddy or the big nice government or the dictatorial standards committee to wipe our butts and make all of our decisions for us. It's called "adulthood". People who can't handle should go back to a childish OS like Windows.

Missed the point

[Bloater]

> With dozen of different distributions the Linux community is so diffuse that the power or significance of any specific entity is severally limited.

The author clearly missed the point of Open-Source. *The power or significance of any specific entity is severally[sic] limited* so the users have control. That is *why* people want to use Open-Source. Indeed there are few reasons apart from that one.

Drastically unconvinced

[oneandoneis2]

Most of the people that *I* talk to who are considering switching to Linux are concerned with nothing but GUIs: So long as there's a decent desktop and the package management has a nice frontend, they couldn't care less about the inner workings of the distro.

DEs are freely interchangeable between distros, and even package manager GUIs are fairly universal - There may be hundreds of distros, but how many are there that don't use RPMs, apt-get or source code?

The amount of community time spent on distro-specific stuff is miniscule compared to the time spent on projects that can be used on a wide variety of distros. The number of distros is therefore largely irrelevant, rather than some community-draining problem like TFA says.

After all, that's the whole point of Open Source, isn't it. . ? Sharing code amongst projects. . ?

Mod Article -1 Troll

[dokebi]

First, the hyperlinks in the articles are actually advertisement links. Second, you cannot consolidate distributions when I can start my own distribution tomorrow.

Dear editors, can we please mod articles? Recently there have been numerous articles that are just thinly disguised advertisements and click-through magnets. Slashdot as a community deserves better.

Re:Mod Article -1 Troll

[slavemowgli]

You're right, of course, but the editors won't listen to you.

The reason for that is simple: unlike you and me, the editors don't care for Slashdot as a community site; or at least, it's not their top priority. What they *do* care about is their bottom line - i.e., advertising revenue. In order to maximize revenue, they need to have a certain amount of stories each day, and the stories need to be sufficiently interesting. The best story, from an editor's perspective, is one that generates lots of attention and makes people come back the next day to read more.

Sadly enough, troll stories do fit this bill quite perfectly - they're simply exploiting human nature and its hunger for sensations. Ever wonder why there are so many tabloid newspapers and why they're read by so many people? It's the same thing.

Of course, Slashdot has a reputation as being a news source for people who're more intelligent and more interested in technical issues than the average Joe Sixpack from the street - but it's still the same fundamental mechanism.

You can't really change anything about it, either - you could stop reading Slashdot, of course, but chances are that due to the sheer number of users, it wouldn't be noticed. There are alternatives, of course, that you could turn to, but they, too, suffer from the same problem. Kuro5hin, for example, caters to a specific audience just as much as Slashdot does, and uses the same tricks - somebody reading Kuro5hin might condemn Slashdot for what they do, but will probably fail to realize that the same thing is happening on K5, too. Ultimately, all news sources find their target audience and cater to that - if you like it, good, if you don't, not good, but it's not gonna change.

Incidentally, this is why we get dupes so often, too - contrary to popular belief, editors *are* aware that they're posting dupes, but they need good stories, and if something garners a lot of attention the first time it's posted, then it'll likely be posted again. And of course, if an editor still feels uneasy about it, they can always rationalize it away by pointing out the fact that people from different time zones might have missed the original story and so on.

But yeah, that's Slashdot. Love it or hate it, but you're not gonna fundamentally change it; and personally, I can live with daily dupes and troll articles as long as *most* articles are good, at least.

Re:Mod Article -1 Troll

[William-Ely]

I always thought articles were simply prompts for people to make comments about something losely related to its description.

Yeah right...

[One Louder]

I see.

So, a consumer walks into a computer store to buy a computer, and they're overwhelmed by too many choices of Linux.

Sorry, I don't buy it.

The problem is that, with few exceptions, you can't buy a machine at retail with *any* Linux on it. The only way Linux ends up on machines now is when a consumer decides to get rid of the OS they got for "free" on the machine.

Consolidating Linux distros doesn't do anything about getting it into the hands of users and onto machines - an effective sales and marketing organization does that.

just for some things

[diegocgteleline.es]

Well, no, we don't really want this. What we need, more compatibility between different distros, not a single distro

DEB vs RPM vs ebuilds vs ..., for one. And then, even if we'd use the same packaging format, we'd have lots of troubles anyway: The "package namespace" is different in each distro, xorg can be split up in 50 different packages called "lib$FOO" in debian xand in fedora 25 called "xorg-lib-$FOO", or some shit.

That's the biggest problem for inter-distro compatibility IMO . And the one way that it can be fixed is by moving the "packagin work" to developers (ie: let the developers write the spec files / debianize them, don't redo all the packaging work yourself as ALL distros currently do). But then, the one packaging format that encourages developers that is autopackage, which nobody is going to use because it's not .deb or .rpm. Sight....

Re:Wrong questions

[Rosco P. Coltrane]

Perhaps, but as long as they exist we have to waste our time resolving them over and over and over again.

The question is more like: how much time do you spend working on installing/configuring/repairing the Linux (or BSD or whatever *nix) box of your choice, versus how much time you spend installing/configuring/repairing Windows?

For me, it's easy:

1 - Installing Linux: maybe twice as much time as Windows, due mainly to the lack of prepackaged drivers for this or that

2 - Configuring Unix: 10 times as much time as Windows, because I want to have everything neat and well installed, and KDE can be non-obvious at times.

3 - Maintaining Unix: 0. Maintaining Windows: it's an endless pain in the butt (patching, running Norton, de-spyware-ing, de-virusing, renewing licenses, etc etc...)

So, in terms of time, I spend a lot more time installing and configuring Unix, but then after that I'm done for good.

So even with the minor differences in distros as they are, I'm winning over using Windows anyway. And I'm not even talking about the hard-dollar price of Windows and Windows software, so that's why I'm saying that, for moderately technically-savvy people, Linux is already a better choice than Windows, even with its flaws.

68 millionth verse, same as the first

[petrus4]

More FUD of the "Eine Reich, Eine Volk, Eine Fuhrer!" variety.

I'd like to think I'm only going to have to explain this once (yeah, right) so here goes my attempt to explain a few things to the Windows using world.

1. Windows sucks. NT is probably the single most technologically inferior operating system to have ever seen the light of day. In terms of usability, on the surface it might seem great, but go even a few microns below the surface and it is revealed as an absolute dog. (Keep this point in mind, kids, cos it's a very important one)

2. Microsoft have taught the computer using world to think in a number of perverted, unnatural, and generally harmful ways. One of these ways is the insistence that one size has to fit all, i.e., the concept of a monoculture. There can't be more than one operating system in existence at any one time, goes the old saw. Unfortunately what Microsoft doesn't understand (aside from virtually everything else, that is) is that diversity is actually good for computer security, rather than bad for it. If different people run different operating systems, or even different versions of a similar or same operating system, it means that the anarchic 14 year olds wanting to break into said computers will have to work harder...because they will need to write versions of a given virus for a greater number of operating systems than just one. We could even hope that faced with that much effort, they won't bother.

3. Another one of these bad ways of thinking is the insistence that every GUI on the planet be identical to Windows'. You'll normally never hear me praising Apple (takes deep breath, wonders if he can really do this) but they also came up with some great ideas for user interface design, as well. The people currently designing KDE for Linux have even managed to come up with a few.

4. Yet another of Microsoft's evil ideas is the concept that programs should be designed monolithically. This actually follows on from the "Eine Reich, Eine Volk, Eine Fuhrer!" groupthink mentioned earlier. The Linux way of doing things on the other hand tends towards making various pieces which snap together, so that whoever works on a piece only has to worry about the bugs in said piece, rather than the entire program. Because the pieces are often fairly small, they're also usually a lot easier to understand than the sort of software Microsoft writes, and it's therefore easier to figure out how they work.

5. So from these few examples, we can see how Microsoft's ideology is bad. Therefore, I humbly beg you to kindly cease and desist authoring screeds about how Linux should supposedly be more like Microsoft's monstrosities...because it really shouldn't. Microsoft should be taking pages from Linux's book, not the other way around...for many different reasons.

Re:68 millionth verse, same as the first

[macklin01]

More FUD of the "Eine Reich, Eine Volk, Eine Fuhrer!" variety.

Wow, I never knew Hitler was so feminine! Perhaps it's like Adolph Elizabeth Hitler in The Producers." You need a different article: "Ein Reich, ein Volk, ein Führer." ;-) -- Paul

Re:Ford?

[cmarkn]

And the reason that Ford does not sell enough cars is that it has too many models?

It's not that there are too many models, it's that they are too different. Some models make you drive on the left side of the road, some you drive on the right side of the road, some you drive right down the middle, and some require you to build your own road.

Meanwhile, busses carry 95% of the traffic, even though they are uncomfortable and smell bad and break down a lot. To switch to driving yourself requires developing too many new skills that you don't need to ride the bus, and those do not carry over to the different models.

So it's not necessarily that the number of choices is, in itself, a bad thing; it is that each choice locks you in to that model, so that even though there are 100 choices, once you pick one, you are stuck with it. That makes the cost of picking wrong so high that it just isn't worth getting off the bus.

I think they mean like Microsoft.

[einhverfr]

No. I don;t think it was a reference to UnitedLinux. From the article:

"With dozen of different distributions the Linux community is so diffuse that the power or significance of any specific entity is severally limited."

Evidently, they are suggesting that centralized control over the operating system is what is required for Linux to spread. This is not like United Linux because no single entity controlled United Linux. This is instead the Microsoft model, which is why they are wrong.

Clearly there are a lot of areas that need improvement on unifying the base platform for Linux-based systems. And clearly the fragmentation has caused problems (was that Runlevle 2 or 3? Was that runlevel 3 or 5 in this distro?) but these areas are being worked on. The answer comes in many forms, from FreeDesktop.org to the LSB project.

The average Linux user should not have to worry about the holy wars regarding KDE v. GNOME. They should be able to get KDE apps and GTK apps running on the same system and integrating seamlessly without any problems. This is happening.

The average Linux admin should not have to worry about which utilities are on a system, which runlevel is which, what the device name for the serial port is, and half a dozen other annoyances that have at one time or another plagued the Linux world. These things should be standardized. And it is happening.

In short, consolidation is not the answer. Standardization is the answer. Interoperability is the answer. On top of that, each vendor should be encouraged to extend the standard as a way of trying out new things. Eventually new ideas will make their way in, just like other standards (POSIX, SQL, etc).

And just to mention this, there may be cases where the standard does not apply. I for one don't think that TiVo's need to be LSB compliant...