Submitted by “Remember the 1980s worries about how the ‘forking’ of Unix could hurt that operating system’s chances for adoption? That was nothing compared to the mess we’ve got today with Linux, where upwards of 300 distributions vie for the attention of computer users seeking an alternative to Windows.”
Of course, it doesn’t provide one argument for its assertion that too many distros make for a mess.
Seriously, read it, not one.
It only acts as if this were selfevident.
It also goes without saying that it simply takes the number from distrowatch and acts as if it were in any way significant. What this of courses misses, on purpose I’m sure, is that most of the distros listed there are extremely small, some are mere private for fun projects, some are very specialized, some are simply made for a very special audience (e.g.: provide a distro in parsi).
How exactly it can be considered a bad thing if someone who enjoys doing this kind of stuff makes a distro that is intended as a firewall to run from an USB-stick, will be forever beyond me.
Ah, and don’t forget to call linux a religion while you are at it.
What a terrible article.
Ah, and don’t forget to call linux a religion while you are at it.
I think we should take to calling the free software movement a political party as opposed to a religion. It lends itself to more credibility, and it more accurately describes the positions we represent.
We don’t think that RMS is a god. We think he introduced an important theory concerning the social, economic, and legal aspects of software. He’s the founder of our political party, not the prophet of our religion.
I think we should take to calling the free software movement a political party as opposed to a religion. It lends itself to more credibility, and it more accurately describes the positions we represent.
We don’t think that RMS is a god.
That wouldn’t solve our problem. It’s our opponents that call our stance a “religion”, in an effort to discredit us. What they fail to mention (of course) is that in many cases they hold just as tightly (if not more) to their beliefs than we do to ours.
And you can have a religion without a god.
And you can have a religion without a god.
Exactly.
Also, if someone’s going to call the Free Software movement a religion, referring to Richard Stallman as its god is inaccurate. Some Free Software people reject RMS and disagree with him fervently. I don’t know of any followers of a religion that reject its god(s) (assuming it has any).
More so, all these too-many-distros articles fail to take into account that less popular distros disappear over time due to people gravitated towards feature-rich distros that suit their needs.
Put this argument into economics terms, how is having fewer choices a good thing?
Of course, it doesn’t provide one argument for its assertion that too many distros make for a mess.
Seriously, read it, not one.
It only acts as if this were selfevident.
Well, Ralph, I would probably say the same thing about peoples’ complaints that there are too many editions/SKUs of Windows, that it generates confusion, yadda, yadda, yadda. There are two sides to every coin.
I’m honestly beginning to wonder if OS News is trolling for ad revenue.
I think a lot of people here, including myself, installed Adblock.
No Adblock here.
…just a hosts file.
Besides, I don’t see how trolling would increase ad revenue or somehow attract visitors to the site. If anything, trolling could cause OSAlert’ readers to leave.
I think it’s unfair that people have modded down this post for voicing what is a valid concern. And why when it is at -5 did my + vote not bring it to -4? Thom?
I think anyone who says anything againts OSAlert ads will be automatically modded down to -5.
You cannot vote on your own posts
Besides that, you were off-topic.
It was my post not chekr’s
While I’m not going to cry conspiracy, the post was partially off-topic, -5 does seem a little harsh.
My comment was partially on-topic in that this is an argument that’s been made a million times about Linux but Linux has continued to grow and flourish. So why on Earth would a publication like OSAlert that must be completely aware of how tired this argument is continue to publish links to poorly written articles on the subject unless to encourage flame wars? It’s something Slashdot is being increasingly accused of too, I think quite fairly.
The text is interesting to read but full of misconceptions. Linux isn’t a forking mess. Linux distributions serve the same software, only in different versions. Nothing like a fork.
If you also are lost in the distro mess there is an easy solution: you should visit polishlinux.org to make your choice easier
* Linux distros ovieview: http://polishlinux.org/linux/
* BSD family ovieview: http://polishlinux.org/bsd/
* Distro chooser: http://polishlinux.org/choose/quiz/
* Distro comparisons: http://polishlinux.org/choose/comparison/
And you’re all set. If not, there’s always DistroWatch.
What I want to say is that ONLY THE CHOICE IS HARD. After you make the choice, the number of distributions isn’t important for you since you can use all the software other distributions use anyway. The choice is a good thing, not bad. And for those who don’t like to make choice, there are always pre-installed versions by Dell and others.
The issue arises in regards to software development companies and which distribution(s) they should create and maintain their applications for. If the market was alot bigger, it wouldn’t be an issue – but the problem is that a small market is fragmented which means even if you do focus on Novell/OpenSuSE and Red Hat/Fedora, you still risk not being able to serve the remaining 50% of so who don’t run those distributions.
In a pure opensource environment, the idea of fragmentation is a non-issue as the source is maintained by the community but the moment when proprietary vendors enter into the equation, that is where the issues regarding fragmentation comes to the forefront.
With that being said, however, I’m not missing anything major from the proprietary world – I’d like to see Sony make SonicStage available for OpenSolaris so that I can add music to my minidisc player, but its nothing major as I can do it on my flat mates computer.
Those who complain about fragmentation seem to be people who want to impose their proprietary model and the community change to meet the community; I don’t want to sound arrogant, but we were here first, its time for you (the proprietary company) to change to meet the new reality. People don’t care about your secret sauce, they just want a solution that works with a good support structure in place. If you need to keep your sauce secret it speaks volumes over the fact that the price can’t be justified unless everything is kept secret.
Companies usually run RHEL, SLES or some Debian flavor anyway and most of the software vendors (like Oracle, DB2) provide support for these distros. It’s not hard really. Usually the binaries just work for any system (if they include statically-linked third party software as most of them do).
And those proprietery vendors that deliver for desktop users (Skype, Opera, Google) somehow manage to provide custom packages for most popular distros and some genric TGZ for others.
Saying that, I think more standarization (e.g. on libraries used in stable versions of major distribuitions) should be in place — this would make the life of such vendors easier and encourage them to provide software via some standard mechanism for all distros like autopackage.
Still, I don’t think this is the major issue blocking Linux adoption. OEM is.
But if it is a C++ application you need to make sure that you have the correct C++ library installed for it to work – for example, for some time, if you downloaded Firefox off Mozilla it would not run on Fedora unless you downloaded and installed a compatibility library – for a first time user to Linux if it doesn’t load the first time they give up, assume Linux sucks and throws it the idea of using Linux out the door.
Standardisation won’t solve it either – you end up with a static distribution at the core, which may be nice, but becomes outdated quickly – you would also end up with people within the Linux community complain that it is controlled too much by either Novell, Red Hat or some other organisation whom they have a beef with.
OEM isn’t a major stoppage, over 1/2 of the computers sold in the world are from non-big name vendors – they’re small white box companies – they’re not under the same contracts as the big name vendors. I’ve yet to come across a single white box company who was unwilling to install Linux. Heck, its been almost like there is at least one Linux guru at the shop willing to peddling Linux for computers sold.
As for those examples, they aren’t immune. Opera for example will fail to run if you don’t have the correct libcompat packages installed. Skype as a mixed reputation as to whether it completely works on the machine of choice.
So statically link your application and bundle all of the required compopnents in the distribution archive or CD.
Optionally, make the installation routine smart enough to see what needs to be installed on a target box if the required libraries are not already present.
It isn’t that hard. Folks have done this on Windows and OS/2 and other platforms for YEARS.
Just because it flies in the face of UNIX tradition (where the world is dynamically linked) doesn’t mean it can’t be done.
1) “…upwards of 300 distributions vie for the attention of computer users seeking an alternative to Windows.”
Not every distribution on Linux is aiming to compete with Windows. For example, I don’t think Slackware is trying to get Windows users as much as it’s trying to get advanced Linux/BSD users.
2) An application written in GNOME or KDE compiles and works the same across all distributions, provided the distribution supports the library version you wrote the program for. That’s not the same thing as you write a program for Win32 API and you want to compile it on Linux or Mac.
3) A lot of distributions use related package management tools. For example, Ubuntu is derived from Debian. That means you can issue the same commands on a Debian box and expect the package tool to behave the same on an Ubuntu box.
In sum, different distributions are aimed at different niches and tastes. That doesn’t necessarily mean they are always in competition with each other. And programs run the same across almost all distributions.
“In sum, different distributions are aimed at different niches and tastes.”
That’s the theory, but in reality there a only a bunch of real original and useful distros.
It’s not even just things like Slackware trying for other linux users. A lot of the liveCD ‘distros’ are proof of concepts or tech demos, not really trying for any permanent users just acting as a show case, or a shortcut (e.g. Elive, Etoile’s live cd)
Not every distribution on Linux is aiming to compete with Windows.
In sum, different distributions are aimed at different niches and tastes. That doesn’t necessarily mean they are always in competition with each other. And programs run the same across almost all distributions.
That isn’t the complaint. It’s that there are simply too many choices, that the choices are widely divergent, and it isn’t always clear what distro a user should choose.
People tend to fork something in Linux because they aren’t “mature” enough to stand an argument, well most of the time. So this behaviour is weakening Linux development and of course quality. Have a look at the distro djs out there. Choice is always a good thing, too much choice is a mess – it’s that easy. Maybe LSB will tame the Linux beast in future, but I doubt it.
People tend to fork something in Linux because they aren’t “mature” enough to stand an argument, well most of the time.
1. Distros are not forks.
2. I’d really like to see you back this statement up.
What about the compiz/beryl fork? I know at the very least some thought that fork was a tad childish on the part of the beryl guys.
As for distros. I think they are doing just fine. Best I can tell the biggest difference is you have the Red Hat way and the Debian way.
I figure the market will pick the leaders. You may have upwards of four hundred different distros, but only a few will garner the majority of attention at any one given time.
What about the compiz/beryl fork? I know at the very least some thought that fork was a tad childish on the part of the beryl guys.
I disagree. Compiz was clearly meant to be a stable system, and the type of wild experiments going on in Beryl really didn’t belong there. Beryl was quite useful though as a playground where we could see what effects were useful and which should be forgotten about.
Ultimately, I think Compiz should die out as well and just leave the old window managers to add 3D support, so if anything I would call the entire Compiz/Beryl combined project the wasted fork. But it clearly had a purpose originally, so I won’t complain too much.
“Market” doesn’t play much with OSS. For businesses sure, but I for example am an Ubuntu user. But I’m not noted for it anywhere. I didn’t pay for it, I didn’t register it, and if I didn’t tell anyone, no one would know about it.
You can’t know the number of Linux distribution users, and saying that “market” choses the successful distributions is wrong. It has an impact, but for example Debian and Ubuntu were first community successful and only then “market” successful.
The compiz beryl fork was also a good thing(tm). If it didn’t happen, we’d loose all the beryl devels and their work.
There are only three or so distros that matters, at least if we are looking at it from a business perspective.
The difference, between the current state of Linux and the old fragmented Unix is that most Linux distros are interoperable and contain almost the same things. The differnce is mostly in packaging e.g. different support deals.
The reason we have so few distros that count is not that the other 2997 distros are bad from a technical point of view. It’s about their credability in being able to stay in business and provide business with support over an extended period of time.
Another difference, is that good changes from the 2997 distros that doesn’t matter for financial resons very quicly will propagate into distros that do matter financially. Unlike in the old unix days Linux is open source and why should for profit Linux companies avoid adding improvemnets to their products.
I really don’t see any problem having many distros, on the contrary they strengthen Linux, by providing special solutions for people not confortable by mainstream distros.
There are only three or so distros that matters, at least if we are looking at it from a business perspective.
Yeah, but a great many people adopting Linux aren’t doing it from a business perspective but, rather, as an alternative desktop to Windows.
How do these guys come up with such interesting new insights about free software? You’d think they’d have to start recycling the same old tired arguments at some point. But they keep it fresh. I bet that nobody had ever thought about the proliferation of Linux distributions until now. Certainly not in the level of depth pursued in this article.
I’ll take a break from the sarcasm to make a couple serious points. UNIX fragmentation was bad because everybody kept fighting over the copyright and there was no pervasive culture of open source. Nobody could decide whether the derivative UNIX implementations were legal, and it was often impossible to port applications across distributions.
In the Linux world, we have an incredibly high awareness of copyright. We’ve learned from the UNIX wars that we can’t be casual about copyright. Patents are slightly more problematic since there’s no established best-practices for ensuring patent cleanliness in software development. It’s a minefield we have to cross. All we can do is be careful and hope we don’t get torpedoed down the line. We’re all in this together, regardless of distro proliferation.
Furthermore, open source licensing effectively mitigates the application compatibility issues associated with distro proliferation. Most distros provide packaging tools that allow teams of employees and volunteers to efficiently package open source software. This is often the vehicle through which end-users make the transition to become contributors, adding strength and diversity to the free software community.
Proprietary software is more of a challenge, and that is why there will never be more than three major Linux distributions widely supported by proprietary ISVs. No amount of distro proliferation can change this, and so for those who depend on proprietary applications, there is no reason to be concerned about proliferation.
You make some excellent points. And the author didn’t even get the concept of free software right. He’s referring to “free as in beer”, not the “free as in freedom” that free software refers to.
And the author didn’t even get the concept of free software right. He’s referring to “free as in beer”, not the “free as in freedom” that free software refers to.
I actually think he does, sort of. He gets the freedom part to an extent because he knows that forks can be freely made and used, but he has the common misconception that freedom has to always equal (or lead to) being “free as in beer”.
Nonetheless, there are other, more valid reasons not to use Linux if InformationWeek wants to spread FUD. This whole forking argument has seen more forking than Linux itself. Personally, I don’t see the “mess” or the big deal about forking. Linux itself (the kernel) is not made “messy” by forking, nor are the top two or three distributions. The “mess” is what leads a lot of people to Linux, in fact, and I’ve not known of anybody turning down Linux because of its forking “mess”.
The “mess” doesn’t hamper productivity, either. If a developer was going to contribute to a small, esoteric Linux distro, that doesn’t mean (s)he’d contribute to a bloated, mainstream distro–(s)he would have contributed to it to start with if that was what was going to happen.
Also: Which begs the question: Isn’t one of the tenets of the Linux “religion” the belief that open source advocated are wiser than Windows users? (The other core concept is that idea that free software is somehow nobler than stuff you have to pay for.) They’d never let Linux evolve into a “giant hairball,” which was the colorful way Sun Microsystems’ chairman characterized Windows.
The stupidity and ignorance of that boggles my mind beyond words. It even made me chuckle. Seriously. Preventing a “giant hairball” is exactly why the whole concept of distributions exists, instead of having one or two uber-bloated Linux flavors that have and do everything, blindly assuming that one size can fit all.
Since when did ‘choice’ become a downside to an operating system? Really… there are only a handfull of major desktop distributions (openSUSE, Fedora, Ubuntu and possibly Gentoo)… that is not a definate list, but it is a good roundup of some of the more popular distro’s. Like someone else pointed out, a lot of the other distros out there are special purpose only, and some are just for fun… Heck even I had a distro a while back.. this is just another article that wishes to spread FUD about Linux being divided… when really it is going stronger than it ever has.
It isn’t so hard. Really.
Having dozens of automobile companies and models hasn’t made using cars any more difficult. Why should it make using software any more difficult?
Most American’s choose to drive an automatic, sort of like windows (I guess they are just lazy… the reason given here for anyone who actually prefers Windows).
All those cars… those European imports with manual gearboxes… and what do Americans want. Automatics. Huge, fat, BLOATED automatics. With lots of eye-candy – a mobile living room – games, tv, music…
Automatics mean that anyone can drive them without having a great knowledege of the car or a good feel for how the car performs. People who care about things like this, that aren’t lazy, are interested in learning how to really drive, and want to CHOOSE when they change gear drive manuals. In Europe only old or disabled people drive automatics.
All those cars to choose from and people still want easy options… but people can still make mistakes. Buy the wrong car. Read a review of a car by a fan of that type of car and you’d think it was made of gold. There are plenty of reviews of Linux distros, all claiming that their subject is perfect for those switching from Windows. If a company does as you suggest and just chooses one (based on what they find online) there’ll probably just end up using a piece of crap that make them go back to Windows. That’s very easy to do because there is only one Windows.
All those cars… those European imports with manual gearboxes… and what do Americans want. Automatics. Huge, fat, BLOATED automatics. With lots of eye-candy – a mobile living room – games, tv, music…
I’m an American. I use Linux and OS/2 as well as Windows, and I’ve driven a stick since 1988. I much prefer the feel of a 5-speed to a PRNDL.
You might be correct about most Americans, but not all.
If a company does as you suggest and just chooses one (based on what they find online) there’ll probably just end up using a piece of crap that make them go back to Windows. That’s very easy to do because there is only one Windows.
The companies I’ve worked for tend to make well-informed decisions about software and run the stuff in a test lab before actually deploying it. Something to do with the attitude that airlines need to be safe and all that.
One Windows? Which one? Win2k? XP? Vista? And which flavor of Vista? You gotta be kidding me…
“That’s very easy to do because there is only one Windows.”
Uh, that’s not true. It’s been said that Windows greatest competitor is Windows. There is 3 different codebases; win9x, winnt and wince and 7 flavours of win9x (Win95 a/b/c, Win98, Win98SE and WinME) I still see a lot of Win98 and WinMe and 3 important versions of WinNT (Win2k, 4 versions of Windows XP, Vista). Each is a little different, even though MS has dome a pretty good job of compatibility up to Vista.
Having dozens of automobile companies and models hasn’t made using cars any more difficult. Why should it make using software any more difficult?
But it makes customising the car we bought that much harder. We can’t just walk into a car parts store, grab PartX and attach it to our car. We have to find PartX-Y-Z for CarA, and make sure we have the supporting MountB to make it work. And if something goes wrong, we have to track down Mechanic1 and AutoRepairShop2 and hope they’ve seen our model of car, otherwise we’ll need to look around online and try and figure it out ourselves (although not all cars have manuals readily available).
Just think, if there were only one or two standard models of cars, with standardised interfaces for all the parts, it would be so much simpler to get 3rd party parts for our cars.
Edited 2007-07-20 17:38
Err, i just bring the car to the garage where i bought and pick it up when it’s repaired. the parts are the garages problem.
In your analogy, if the different car companies used the same set of components: same engine / chassis / wheels / transmission, while just painted them differently and had different logos, then it would indeed be a mess in the name of choice.
The point here is that it diverts creative energies towards less significant pursuits.
I use to think something along the lines of the article, but then I realized a few things. One, it seems that most distros fall either under the red hat way or the debian way. Two, the market will determine the winner. Three, applications probably have more impact on the operating systems design than people give credit for.
My only problems with the large number of distros is that it might affect the decision of some people who are contemplating a switch to Linux. However, I think in this day and time it would only play a very small role if any at all in that area. I also think that if all the developers of the different Linux distros came together to support only a handful we might see a even greater operating system, but that’s just speculation.
At any rate I think Linux is progressing along fine even with about 400 different distros. Needless to say I’m not to worried about the large number of distros.
Disclaimer: I know a some of you don’t care about converting WIndows users to Linux – this comment is for those of you who do.
Some of you say that too many distros is a tired, old argument. Well, guess what .. I’m a Windows user, and I think there are too many distros. If you want me to switch, does my opinion not count for anything?
If there were only 2 or 3, I might be more inclined to try them. But as it is, you’ve got Ubuntu, Fedora, Debian, Knoppix, Mandriva, Linspire, Xandros, Suse, Gentoo, Mepis, Slackware, etc. I mean, you’ve gotta be kidding me!
And the distro chooser that somebody linked to asks such questions as “How important stability and maturity is for you?” and my favorite, “How important speed and responsiveness is for you?” What kind of f**king question is that????? Yeah, I’d like a distro that’s slower than snot on a doorknob, kthx.
Just give me a desktop distro that’s fast, stable, customizable out the ass, but easy enough for relative newbies like me to get started with, but flexible enough so that when I’m comfortable, I can start tweaking without needing a degree in rocket science. Is there not a distro like this? If there isn’t, then I would say that makes all of the current ones (that aren’t server distros) rather pointless.
Edited 2007-07-20 00:21
It really doesn’t matter how many distros there are though. You only use one. What if they are all the same? It doesn’t matter which you choose. What if they aren’t all the same? Read some reviews and pick the best one.
Easy!
Actually, some of us more than one.
I use one distro as a dedicated firewall, another as a permanent desktop, and a third and fourth as LiveCDs for support purposes.
It *still* isn’t that hard to keep things organized.
With you there WorknMan. [OS order – XP/OSX/Zeta..BeOS] I’ve tried maybe 10 different flavours of Linux, and ultimately, what makes me wipe them is pure & simple typographical/UI ugliness. [Relax, this is clearly not a technical appraisal] Linuxpeeps, you’re not handsome, not even in a certain light.
You should come to my house, where we have three Windows PCs and one Linux box. The fonts on the Linux box and one of the Windows PC’s are lovely, and the fonts on the other two Windows’ boxen suck. And it was as easy to set them up right in Linux as in Windows. What are you using, twm?
Plus, the Linux box has been reinstalled from scratch once in the last two years, and it was a hell of a lot easier, that one time, than any of the multiple Windows reloads I’ve had to do in the last quarter. Even without a lovely cuddly graphical screen. Which, by the way, Windows doesn’t give you until you’ve done the partitioning.
what makes me wipe them is pure & simple typographical/UI ugliness.
On my Ubuntu Feisty with Font Preference; Smoothing Grayscale, Hinting None. The fonts look much nicer than the fonts on my Win XP system at work. The Linux fonts look pretty much the same as the fonts on a Mac but the Windows fonts look all emaciated and ugly. I would wipe the Win box and replace it with Ubuntu if I could get away with it.
Certainly the properly set up Linux fonts look much better on the same system compared to XP, when I use the Ubuntu live disk. Not like the old days five years ago when the Linux fonts looked definitely ugly.
Edited 2007-07-20 09:29
Ubuntu 6.10. Want another?
What gave you the idea that anyone *wanted* you to “switch?” Do what you will.
How about the people working on Linux? Why do you think they do it? So that nobody uses their software?
What gave you the idea that anyone *wanted* you to “switch?” Do what you will.
Oh, I dunno .. maybe this? http://www.badvista.org
Also, just about every time a Microsoft/Windows-related article shows up here or somewhere else where there’s a comments section, there’s at least one or two people who recommend that I ditch Windows and install Linux, so it seems there are those who are interested (too interested if you ask me) in what OS I’m running.
My point is that linux is not aimed at you, obviously.
1) You have demands which you don’t feel are being met – fine, linux isn’t for everyone. It’s absurd to insist that linux bend to your will, though.
2) Your refusing to use it isn’t hurting anyone, or, frankly, making a difference at all. If you hate it and are happier using other OSes, use them instead. You’ll be happier, and most likely so will the people who would have to read your posts in linux support forums.
3) Free and Freedom require some elbow grease and some shopping around. If that’s more than you’re willing to do to find a linux distro that suits your needs, see my first point.
Seriously – I think Windows is dogshit dipped in hogshit as far as OSes go, but I use it at work because it suits me to. I do not use it at home, or in any situation where I’m expecting to enjoy myself. I do not, however, make posts about what Microsoft had better well do if they want me to stoop so low as to buy Vista. I doubt microsoft gives a damn what I use on my machine since I am statistically insignificant. So are you to the linux world.
However, evangelists and zealots are, as you point out, everywhere, and can be a major PITA, and are embarrassing at times – that point is taken. While I agree, in general, with their stance that linux is better than windows, I tend to keep my opinion to myself in most cases. At the very least my response to “I goted a viruses again d00dzors- what should I dooo???” is never “Run Linux.”
Too many? Let me put it this way: Fedora, Ubuntu, Mandriva and openSUSE. All of them will be more than capable of suiting your needs (from what you mentioned). Still too many? Write four little papers labeled with a distro name each one, put them on a bag, and pick one paper. Do not look inside the bag, eh?
There, problem solved. I could give you many more distro names, but that would take a lot of paper and ink.
Someday, people will learn that distros are just a bunch of software packed together. While is true that not every distro is exactly like the rest, if you focus your search in a given target purpose (Desktop use, development, server task) things will be far easier.
Saying “too many distros to even give them a try” is just a silly excuse. You’re better off saying “I won’t use GNU/Linux because I actually like Windows” reather than “You’re not worthy enough of my attention, solve my problems for me“.
Don’t get me wrong, there’s a lot of people willing to help newbies, heck, that’s why we call ourselves “a community”, but solving your pre judgements is something way out of our reach.
No offense here, but if you’re not willing to do something, to learn a bit o to ask other people, I foresee a sad experience with F/OSS. That’s just the way it works.
By the way, if you really want to try stuff and learn, I’ll give you as many hands as I can, just drop me a PM.
Again, there’s a lot of people willing to help, you just need manners and some curiosity. Oh, and a spare gigabytes on your HD too
What a rant really!
If you find it that difficult to choose a linux distribution that suits your needs, then you probably find it equally disturbing choosing a pasta sauce from the supermarket. How hard is it to do a little research and choose a name brand. I mean really, it costs you nothing but a little time. Do you not see how ridiculous you’re being?
Oh and here’s a little tip, Fedora or Ubuntu will do just what you asked for!
More like, you picked a name brand, but they have 350 derivatives of that brand. Different flavors. A few are good, chunky, creamy, etc. But 300? Yikes.
Of course, in reality, things aren’t *that* bad. It’s more of a choice between slightly less than 10 or so distros for new users.
It’s still an extra layer of complexity for them. Most don’t even care what OS they are using. They see the computer as one unit. Having to select an OS is confusing enough for most of them. Then having to select a distribution if they choose linux is another layer.
“If there were only 2 or 3, I might be more inclined to try them. But as it is, you’ve got Ubuntu, Fedora, Debian, Knoppix, Mandriva, Linspire, Xandros, Suse, Gentoo, Mepis, Slackware, etc. I mean, you’ve gotta be kidding me!”
you don’t only have to use one why not use all of those you just mentioned. I use more then one linux distro becasue i also have a hard time picking just one because they are all to good not to use.
Yeah, I’d like a distro that’s slower than snot on a doorknob, kthx.
Of course you would, you’re using XP (and as for Vista)!
Seriously though, do you know how many brands GM sells with the /same body parts/? Just in the US? And some of them get sold as Vauxhalls/Opels (in the UK and Europe) and Holdens (in Australia/NZ) too.
And yet when was the last time /anybody/ complained about the number of identical car brands GM produces?
it’d be really nice if whoever is responsible stopped linking to this kind of bullshit articles that do not contain any meat and just serve the purpose of causing a flamebait.
what does “Error: This comment is already at the upper limit.” mean
“That was nothing compared to the mess we’ve got today with Linux, where upwards of 300 distributions vie for the attention of computer users seeking an alternative to Windows.”
Clearly a misunderstanding of open source.
1.Its about choice and freedom to change as we see fit.
2.Those 300 distros are not generic.Basically they are grouped for target audiences each having a different need with the option to make a choice as the person(s) see fit.
3.Every Linux distro is not seeking to compete with Windows.
4.Linux is not in a mess.If anything Linux is at its most cohesive point to date making some serious inroads onto the desktop , and no I am not saying 2007 is the year of the Linux desktop.Just you cannot overlook Ubuntu being offered on Dell desktops and notebooks , especially later into Europe, a linux stronghold.
I think the writer needs to be more insightful if he has an axe to grind with the success of the open source initiative.
Edited 2007-07-20 00:30
Why bother even having the discussion about whether the existence of many distributions is good or bad, when there doesn’t exist any sort of authority in the universe capable of regulating the number of distributions? This is like arguing over whether there are too many different books or songs in the world: ultimately pointless, because you can’t stop people from writing new ones. The simple fact is that most distributions exist because someone thought it would be fun to create their own distribution. What more justification do you think is needed?
All of the tedious articles about what the open source software community “needs to do” miss a vital point: There is no such thing as the open source software community. There are lots of people who write software, and some of them release the source code to that software. That’s the only thing they all share in common. The “community” isn’t a community any more than atheism is a religion, and it has no universally shared goals, aspirations, responsibilities, or duties.
Edited 2007-07-20 00:32
So what’s the ideal number of distros? One? Two? Seven? I know, forty-two. Right?
The too-many distros argument is becoming as tired as the “BSD is dead” drivel. Instead of over-dramatizing the proliferation of distros, how about talking about the amazing accomplishments over the last few years to bridge the underlying compatibility issues to ensure interoperability *and* choice?
You never see a mention of what Xorg has accomplished since (gosh) forking from XFree. You never see them gush about the successful work freedesktop.org has done and continues to do. No reference to the substantially improved interoperability among different DE’s. No credit to the work that went into dbus or hal et al. And they certainly never mention the work being done through LSB, since that would undermine the whole “too many distros” argument.
I love that there are 359 distros. I wish there were more. Every one of those means that someone at some time had an itch to scratch and did it the best way they felt they could. Sure, most of it is probably white noise. I’d probably consider many of them useless or pointless, in fact, I’ll probably never get around to using 356 of them, but I still like that they’re there. I like that a developer with an inspiration can still throw their hat in the ring with a new idea. People seem to forget that although linux has evolved into a multi-billion dollar industry, it’s still nice that the grass roots spirit remains. Besides, I’d rather have to comb through 1000 useless ideas to find one truly brilliant, rather than see that brilliant idea discouraged from ever being realized by the need to conform.
Sure, one can argue about the need to focus resources etc. but the people making these arguments are inevitably in the linux MUST BEAT MICROSOFT!!! camp. I question whether they truly want to linux to “succeed” or just for Microsoft et al. to fail. Novell, RH, Mandriva and a few of the others pay their devs to work, but there is a lot of blood, sweat and tears contributed by people driven solely by the desire to see their idea realized, no matter how grand or how insignificant. I think that is a good thing. I find it a little ridiculous to complain that too many people are working for free to produce free software, instead of allowing others to arbitrarily decide where they should put their efforts. I understand the logic with that argument, but I don’t agree with it and find it a little bit hypocritical.
Ok, I am a little concerned when I see an NIH mentality, I will agree that can be wasteful if purely done for that reason, but even so, that’s simply a side-effect of the overall model, you can’t have one without the other. Human nature being what it is.
Still, as always, vive le choice.
As a French speaker, I am gladly helping correcting that last sentence. It is “Vive le choix”.
D’accord. Comment vous dites des ” tongue-in-cheek ” ?
I’m Canadian, so mangling French is my constitutional right, although frankly, my phrase was intentional. I use it frequently, much to the dismay of my French wife.
What a small world! I am also Canadian. Lets get on topic before someone will complain.
I wish there were a way to PM
P.S: “tongue-in-cheek” roughly means “deuxieme degre”.
Well, those 359 might not quite be white noise, but consider that various groups even put out more than one set of distros that often go by different names, that may vary by nothing more than what’s included in the default install. I consider Ubuntu, Kubuntu, Xubuntu, Flubuntu, Edubuntu and Ubuntu Christian Edition essentially the same distro.
Personally, I don’t think you can call them all that different because in this day and age with small exceptions everyone’s using the same software. They differ by package managers, versions, size and type of install target… but really, aside from minor differences KDE is KDE is KDE and the difference is who has it set up better or put their logo on it.
“Linux is a forking mess…”
Well, the way I see it, Windows IS a MESS and it is not even forking! In fact innovation seems to be reaching a stand still, probably a dead end, while complexity (for development) is at an all time high.
Oh come on, he was trying to be vulgar and witty.
You see, Windows doesn’t matter in all of this because it CAN’T fork. So whatever you have, is what you have, and you just gotta deal.
i agree with most comments in that choice of linux distros is not a bad thing considering the kernel and software are generally the same or can be made the same.
but let me play devil’s advocate, i’m not saying that i agree with what he’s saying…
but let’s say windows kernel was available and had 300+ distributions and several different ways of installing software (add/remove programs, win-apt-get, automatic update, win-emerge, win-install-pkg, etc) and each distribution had its own set of software packages. that would be kind of confusing and you’d have all these different guys creating packages for all the different install types. i kind of hate the fact that you have rpms, debs, etc. imagine all the wasted time collectively by everyone trying to get packages for each installer built, sure it’s scripted, but add all that time up for every project. i know i had a tough time finding slackware packages because it didn’t seem as popular. anyway, i solved my problem by just always getting source and compiling.
i guess windows had different install types too, install anywhere, zero installer, wise, etc, but microsoft killed off those interfaces with the msi interface *i think*.
i dunno, i think it’s not comparable, windows to linux. you can’t, after all, take windows kernel source and rebuild throwing out everything you don’t want (not that anyone in their right mind would want to).
i’m a slackware user and i don’t have a problem that everyone else is using other distros, so long as the software that i use (linux kernel + linux software applications) continue to get improved, it doesn’t matter if slackware goes away some day due to insufficient user base, i’ll still be able to run linux.
Isn’t that number an order of magnitude off?
I just had this argument in another thread. The notion that there are too many distros is ridiculous. Choice doesn’t seem to be a problem for people buying computers from different companies with different processors, different graphics cards and different hard drives even though most people have no idea what the difference means. If choice was a bad thing we would have an appliance for a computer, with no options for OS, applications, or even system performance. I think everyone who reads OSAlert can agree that that is a bad thing.
Choice is good in the sense that we need more operating systems (like Windows, OS X, Linux, Syllable, Sky OS, Haiku) not more types of Linux.
Choice is good in the sense that we need more operating systems (like Windows, OS X, Linux, Syllable, Sky OS, Haiku) not more types of Linux.
Says who?
From a biological perspective, heterogenity is a positive trait, and it also applies to IT to a certain degree for similar reasons.
The presence of multiple executable formats makes it more difficult for a single rogue program (virus, trojan, etc.) to take out a given infrastructure.
From a biological perspective, heterogenity is a positive trait, and it also applies to IT to a certain degree for similar reasons.
The presence of multiple executable formats makes it more difficult for a single rogue program (virus, trojan, etc.) to take out a given infrastructure.
Considering Linux doesn’t have a virus problem, I don’t seem to find your evidence very compelling. Besides, Linux distributions are different enough from each other to make it very difficult for a binary virus to propogate.
Considering Linux doesn’t have a virus problem, I don’t seem to find your evidence very compelling. Besides, Linux distributions are different enough from each other to make it very difficult for a binary virus to propogate.
That’s the point though! Since there is a large variation, it’s hard for one virus or some other malware to have a chance to harm the OS. If a linux system goes down or is not available anymore for some reason, it is just as easy to switch to another one. Also linux is very configurable and open source. I can change and configure my desktop environment or the underlying kernel to fit my needs if I need too. If I need a server, or a stand alone music player or a system for process control, I can go to a linux project (and contribute a little money too) and it is available for free or like I said, *contribute* too the project.
Linux as a whole is one of the best thing that has ever happened to the software industry.
That’s the point though! Since there is a large variation, it’s hard for one virus or some other malware to have a chance to harm the OS. If a linux system goes down or is not available anymore for some reason, it is just as easy to switch to another one. Also linux is very configurable and open source. I can change and configure my desktop environment or the underlying kernel to fit my needs if I need too. If I need a server, or a stand alone music player or a system for process control, I can go to a linux project (and contribute a little money too) and it is available for free or like I said, *contribute* too the project.
Linux as a whole is one of the best thing that has ever happened to the software industry.
Who are you arguing with? You’re agreeing with me.
Who said I was arguing.
My fault,…. I’ll try to be a little clearer about how I direct my comments:)
But that is precisely the point!
If Linux used the exact same executable format and APIs as Windows, it *would* have a virus and trojan problem.
Sure, the virus could only infect files hat a given user process had permission to modify (unless you were running as root or using a distro like Puppy which only runs as superuser), but so what?
That would STILL allow the process to infect other files it had access to (how many advanced users have their own /home/username/bin directory?), it would require the user to remove and reinstall software to remove the problem, it would require the use of virus or trojan scanners to prevent in the first place, and it would generally cause a mess if the program in question were distributed to any extent before the infection or destructive behavior occurred.
It is the relatively heterogeneous nature of the x86 operating system ecosystem that helps to prevent Linux users from experiencing some of the hassles that Windows users have to deal with. The fact that Linux does not run Windows software is an advantage in this instance.
It also illustrates a potential risk with Wine. Not to the extent a real Windows system might experience, of course, but potentially irritating in some cases.
Besides, Linux distros aren’t that dissimilar, and most are using the same basic set of C libraries and the same basic kernel. I suspect it wouldn’t be that hard to get a destructive userland virus or trojan out there if the incentive existed.
Edited 2007-07-22 03:15
If Linux used the exact same executable format and APIs as Windows, it *would* have a virus and trojan problem.
They don’t so strike one.
Sure, the virus could only infect files hat a given user process had permission to modify (unless you were running as root or using a distro like Puppy which only runs as superuser), but so what?
Not if home is mounted noexec, then good luck getting your malware to run at all. Even if not mounted noexec there is no way to cause the malware to start on boot globally like there is in windows. That alone eliminates a huge number of issues. Also most people used packaged programs from their distro, unlike windows where people download crap from anywhere. That’s a huge advantage for package managers in my opinion.
Besides, Linux distros aren’t that dissimilar, and most are using the same basic set of C libraries and the same basic kernel. I suspect it wouldn’t be that hard to get a destructive userland virus or trojan out there if the incentive existed.
They are actually much more different than you think. Binaries are often installed in different directories, library versions are usually different, package format is different, desktop environments are different, so on and so forth. All of these things, and others I have not mentioned, make it much more difficult for virus writers to construct a universal Linux virus or other malware. Also don’t forget that executable have to have executable permissions which is not default.
… with that 369 distinct distribution figure.
It ignores that many of those distributions are dedicated purpose products, like firewalls and recovery CDs.
It ignores that many of those distributions are actually a single distribution (i.e. Ubuntu, Edubuntu, Kubuntu are all repackaged versions of the same distribution).
It ignores that many of those distributions are just plain experimental toys, and not something that a new (or even experienced) person would ever use.
Are there too many distributions in the Linux world. Yeah, I think so because many distributions differ nominally in their goals. But I certainly wouldn’t dive into hyperbole and suggest that there are 369 distributions. Simply put, there aren’t.
@ MacTO But I certainly wouldn’t dive into hyperbole and suggest that there are 369 distributions. Simply put, there aren’t.
Whether there are 369 or just two, very few people actually know that there are any alternatives. By very few, I refer of course to ordinary people. You know; the paying public.
It ignores that many of those distributions are …
Keeping with your theme, it also ignores that Distrowatch now includes at least 6 BSDs, which aren’t Linux, and at least 3 Solaris distributions, which aren’t Linux.
Let me go a step further…
I think the 369 number really fails to account for the fact that there are really a small number (<10 ?) of what I would call “parent” distributions. These include Debian, Fedora, OpenSUSE, Slackware, Mandriva and Gentoo. These distros form the basis for other distributions (e.g., Debian->Ubuntu->[Freespire,Mepis,Mint]) which then form the basis of other distros (Freespire->Linspire) or are repacked into more focused distros (appliance distros).
It would be interesting to see a tree of the 396 with links based on the upstream distro which provides the basis of their packages. I suspect it would look like the tree which claims we are all decended from one of six women.
And some distros are based on the parent of a parent..
It would be interesting to see a tree of the 396 with links based on the upstream distro…
This is the diagram you are looking for:
http://www.linuxlinks.com/portal/news/staticpages/index.php?page=20…
And by opera here I mean the musical type of dramatic production, not the web browser.
The answer to this isn’t to be found in the art, if you like, it’s to be found in the audience. Imagine if opera lovers everywhere started asking, “why is there German opera, and Italian opera, and French, Russian, and (God forbid) English opera? We demand one opera!” Much of opera would die. So, you hae many languages and conventions, but the rules of opera remain the same, otherwise it would not be recognizable as a form of art across nations.
Is opera easy? For some (who already have linguistic skills), maybe. Is opera necessary? Well, what sort of question is that?! Is ‘The Magic Flute’ necessary? I dunno, go and look see! Is opera accessible – well, there’s always The Mikado…
The point being that opera does make demands of its audience, more demands in a way than does popular music. There’s nothing wrong with either: you may immediately ‘get’ Little Richard and Wagner: but the audiences by and large are going to be very different.
If Linux wants to be popular, it may have to swallow the idea of making itself more Gilbert and Sullivan than Das Rheingold (OS X is maybe like Philip Glass), but it will still be opera by any other name. But I suppose Linux doesn’t need to be ‘popular’.
This is being written by a terminal Linux noob, who would love to be able to go to a performance and completely understand all its nuances and sophistication. But I think now I would no more demand things of Linux, that it be this or that, than I would expect to see anyone at all interested in opera ‘demand’ that Der Ring des Nibelungen be in English and cut to 3.5 minutes.
If Linux wants to be popular, it may have to swallow the idea of making itself more Gilbert and Sullivan than Das Rheingold (OS X is maybe like Philip Glass),
Then I guess you never heard of HMS Ubuntufore or The Mandrake* Wives of Windsor!
*Yes I know it’s now Mandriva. I’m referring to older versions, when it didn’t have a reputation for sucking.
Though personally, I suspect most people these days would rather listen to Das Rheingold than Gilbert and Sullivan, if they would listen to opera at all. I certainly would.
I think the author has a point, though it is poorly worded. There is a problem of duplication of work within the linux community. A trip to Distro Watch would lead one to believe that proof of the problem lies in the number of distributions. The real duplication is much more subtle. When Red Hat/Fedora decided that they needed an official graphical means to handle package installation, they reinvented the wheel. They could have adopted Smart, APT4RPM, Yumex, Yast, URPMI (or whatever it is that Mandriva uses). All of these are quality programs that do their job. Even if they don’t have all the features Red Hat/Fedora may be seeking, they certainly could dedicate a team of programmers to add said features. Instead, they chose to create Pirut/Pup. Needless to say, this seems like a waste of time and resources.
More recently, Adam Williamson of Mandriva posted on his blog the fact that Mandriva is interested in implementing a piece of software that provides feedback as to the hardware profiles of Mandriva users. Somebody on Planet Fedora sent an open invitation (though aimed at Adam) to all linux distributions to use Smolt. I read Adam’s blog from time to time, and I have yet to see an entry talking about Smolt. Should they (Mandriva) go their own way in creating another piece of software to duplicate what Smolt currently does, this would be another waste of time.
For all this talk of choice is good, it should be remembered that duplication is bad. Defending choice which promotes duplication is foolish. I would like to see more distributions build upon the already spectacular work of other distributions. There is no shame in this. Software like Revisor allows for choice *and* cuts down on the need for individuals to duplicate the work of others (think of Blag with regards to Fedora). The author of this article is right to call out the linux community for the duplication of work, but the number of linux distributions is not proof of this.
In response to the article summary directly, commercial Unix is closed source. Linux is open source. There’s nothing beyond technical hurdles and personal preferences to keep projects or distros from sharing. The comparison is apples and oranges.
Think about it, the more distributions we have in linux and other open source software, the better off we are.
There are a lot of brilliant people developing, coding, inventing and re-writing new code for linux everyday. Being able to meet challenges, grow, and diversify while also adapting to new situations is the hallmark of a successful living species. These are all qualities that linux and the open source community have.
Yes we do have several large distributions of linux, but like all large beings, they are made up of smaller ones or are derivatives of each other. This makes linux strong and adaptable.
Here’s some other things there is simply too much of:
* Types of pizza. Been to a pizza place lately? Wow, there’s a lot of them pizzas. How could Joe Average possible pick one? No, what we need is less choices, maybe 3 or 4 different types. Also, there are too many places serving pizza. What else but Pizza Hut do we need?
* Types of beer. Seriously, what is with all these different beers and all those companies making them? This is unacceptable, we need maybe 2 or 3 breweries making a handful of different beers. We simply cant expect consumers to make up their own mind.
* Shoes. Wow…just wow! Sure are a lot of different shoes to chose from these days. Back in the day you only had a choice of sandals or… sandals. We need to get pack to those simpler times and rid ourselves of this total mess of choice we have today.
Etc etc etc.
What if you had 350 types of pizza and no one made ingredients for them, at least not any good ones …
“What if you had 350 types of pizza and no one made ingredients for them, at least not any good ones …”
Uh, if I have 350 pizzas they already come with ingredients.
What if you only had one pizza and it sucked?
What if you only had one pizza and it sucked?
The pizza analogy is flawed, as consumers are (most likely) very familiar with the ingredients on a pizza. You know you like or don’t like ansjovis, you know you like or don’t like salami, and so on. In other words, consumers can make an informed choice as to what kind of pizza they want to buy.
Computer-illetirate computer users haven’t a clue about the ‘ingredients’ in a Linux system, and what differentiates Linux system A from Linux system B. Hence, they are overwhelmed, leading to the inability to make an informed choice – meaning, no choice.
If people needed to buy any of the 87594 different Vista editions shrink-wrapped (instead of OEM) they would most likely be overwhelmed too. Choice can easily be crippling if you have no idea as to what to base your choice on.
Edited 2007-07-20 10:45
“The pizza analogy is flawed, as consumers are (most likely) very familiar with the ingredients on a pizza.”
I’m fairly certain human beings are not born with the ability to know exactly what they like and what they don’t like and in what combinations. This is learned by actually trying different things.
“Computer-illetirate computer users haven’t a clue about the ‘ingredients’ in a Linux system, and what differentiates Linux system A from Linux system B. Hence, they are overwhelmed, leading to the inability to make an informed choice – meaning, no choice.”
And naturally if you’re a Linux distro with aims of selling to Joe Average you have to, tada, literate him and make him aware of your product.
You know, just like how if you sell any other product you have to make consumers aware of it.
If you’re a Linux distro with a small market, you don’t.
This has nothing to do with the number of distros though, it’s just market logistics.
Linux diversity is a fictious problem, the huge number of distros have no impact on Joe Average since most of them aren’t aimed at him and he don’t even know they exists. Joe Average doesn’t exactly go to Distrowatch for his Linux needs, he’s more likely to ask someone who knows Linux, look at wikipedia or Google it (2nd entry is Wikipedia).
Ask a random person on the street why he isn’t using Linux. I seriously doubt you’ll get a lot of answers like “Oh, yeah, I’d like to try it but there’s just so many different ones to chose from!”
More likely you’ll get “Uh? Say what? Linux? who is that? I think I may have heard about it…somewhere”.
Edited 2007-07-20 12:45
Same goes here.
When you really look at it, there are about 5 distros which are competing for mainstream adoption. The other 295 are really targeted for very specific needs that don’t effect adoption in the slightest. That argument is like saying that the mini OSX on the iPhone is going to hurt apple adoption, or that Windows Embedded is going to lose windows marketshare. The idea that damn small linux is holding back RHEL in any way is just as ludicrous.
The entire article shows a shocking lack of understand of how opensource works. The author seriously needs to sit down and read the CatB papers before making an embarrassment out of himself. His knowledge of the Linux world doesnt even seem to extend past distrowatch. I mean, LFS is really a learning tool, not a base for major distro creation. The three main “parent” distros are RH/Fedora, Mandrake, and Debian (and to a lesser degree, slack). I can’t think of a single distro that is based off of LFS.
I can never understand how people like this get published, even on the net.
“I can never understand how people like this get published, even on the net.”
The good thing about the internet is that anyone can use it to publish their ideas.
The bad thing about the internet is that everyone is using it to publish their ideas.
OK, so another bad article from Disinformation Week, as others have already said….anyways, he completely misses the point. Thanks to the GPL, all publically distributed changes have to be made public. Most of them make their way upstream to GNOME, KDE, etc., in fact, UNIFYING. There’s no secret embrace and extend happening there as happened with UNIX, closed source or BSD.
So it’s not at all like the UNIX forks.
Edited 2007-07-20 03:54
I’m a software developer as well as a long-time FOSS user and advocate. Please try to understand how having too many choices makes it difficult for us.
First, there’s the choice of library to use. For any given purpose, there could be several libraries with their own APIs to choose from. Some of the API designers differ in the philosophy of their design with others and it can be sometimes downright confusing. Each API-library set has their own devel approach or config tool (think pkg-config vs. xmkmf, etc.).
Go up a level and we have to choose between developing an app for Gnome or KDE. With limited resource, we can’t just develop for all the Desktop Environments out there, and we WANT to maximize the audience / consumer potential.
Then up another level to the distribution. Each distro group has their own packaging scheme, their own package naming convention, their own set of libraries to install.
Just take that into consideration and realize that we have limited resources to work with and it’s understandable why it can be a mess.
You don’t have too choose between KDE or GNOME. Most of those library’s are shared or duped. If you use a distro like Mandriva you can perfectly use a GNOME app in KDE and vice-versa.. The same is true for non-KDE or non-GNOME apps. So what’s the problem?
You also don’t have to bother about what package manager to use. Just give a compilable tar.gz file with a complete list of library’s and programmmes the app depends on…
If the distro makers think your app is valuble or there is a demand from the users, the distro makers will recompile the app and package it in the appropriate package manager (and give the modified source file back – so the developer can use it).. Again – what’s the problem?
So – I really doubt if you ARE a developper. The above arguments sould be known by you, but for some reason you fail to see them – why is that?
Forgive my bad english – it’s not my native language.
“If you find it that difficult to choose a linux distribution that suits your needs, then you probably find it equally disturbing choosing a pasta sauce from the supermarket.”
Amen.
In my book there are too few distros. Seriously, over the next 10 years I hope there are several more for the MS boys to rage about.
The free market based on competition is healthy for the US and the rest of the planet.
“There’s no other way to put it: Linux is a forking mess”
even though many distros are forks from others like ubuntu is a fork form debian,etc i don’t it’s a mess. that is what i like about linux anybody is free to use another distro code base, minus the trademarks, to create and distrubute their own linux os, something that ms and apple does not let you do. oss is a great, fun, friendly community to be part off.
“where upwards of 300 distributions vie for the attention of computer users seeking an alternative to Windows.”
i wish i could use all 300+ distros but i can’t i mainly stick to three distros: fedora because of RPM, debain becaue of apt-get and gentoo because of portage.
People miss the point. Choice is NOT always good when it in fact can limit your choices in the long run, because commercial vendors (ooh, there’s that ugly term, sorry FSF) won’t develop for linux JUST because it’s too much effort. People TALK about linux like it’s a single OS…
RPM, DEB, PORTS, source tarballs… Hell, how many different directories can KDE be installed to? I think it’s four to six) How many different directories can RPM’s be handled in just between different distro’s? (six last time I counted) libc or glibc? The whole situation ends up such a mess that if you are attempting to maintain your own packages you not only have to worry about kernel version and distro methods, but different versions of the SAME distro. Hell, you can’t even rely on binary compatability between kernel versions ON THE SAME PROCESSOR.
It’s not a surprise that many commercial software vendors, independant software vendors, and hardware vendors are going to consider the above effort for such a small market to be not worth the wasted time. I realize the FLOSS ‘open or nothing’ zealots are going to say “so who cares, commercial software is against what we are about”
To which one can reply “Yeah, because god forbid a programmer get paid for the labor of actually WRITING software when you can bleed some high school/college student, professional lecturer, career educator, hobbyist with a day job or other naive idealist for it.” – Stallmans dream, the day when there is no incentive to go into programming as a career.
But as usual people prattle on and on about users without once thinking about developers. He may have looked like an idiot, but Ballmer hit it on the head with “Developers!Developers!Developers!Developers! Developers!Developers!Developers!Developers!”
Hell, even the people behind the different packaging methods are waking up to this… Ever heard of the “Linux Standards Base”?
http://www.linux-watch.com/news/NS4586903228.html
Specifically:
According to Ian Murdock, CTO of the FSG and chair of the LSB, what ISVs want is “to treat Linux as a single platform, which means they want to offer a single package for Linux, much as they do for Windows.”
To give the ISVs what they wanted, the Red Hat and Novell maintainers of RPM and the authors and maintainers of APT, yum, alien, and klik quickly decided that the best real-world solution was to construct “a single API (application programming interface) that could be implemented across the various package systems, because APIs make for nice evolutionary steps and can, done right, mask underlying implementation differences,” according to Murdock.
THESE people get it – and it’s fairly apparant from the rabid fanboy responses in this thread so far that a lot of people do not.
Edited 2007-07-20 06:15
So we missed the point eh? Welcome to the club then because so did you.
1) The original article does not talk about developers in ANY way.
2) you list ONE problem; the different package managers. Now, there are maybe a handful of big distros (the ones commercial vendors, ISV’s etc needs to care about) which leaves us with perhaps 3 different package managers. The obvious solution is then to contract/hire one person to handle the linux package. This has the side effect of more people getting paid to be programmers.
It’s not like there isn’t a boatload of different installation systems for Windows either; Wise, InstallShield etc etc.
“Hell, how many different directories can KDE be installed to?”
As many as you can think of. You know, just like how many Windows installers will let you chose where to install an application.
Heh, Nice try there soulbender.
>> 1) The original article does not talk about
>> developers in ANY way.
Because of course, listing an example of where it is bad in no way adds to the conversation – RIGHT. Besides, he brings the “F” word to the party, and who exactly would even have the skills to do a fork again? DEVELOPERS. Developers bring the user by bringing quality desirable software and hardware compatability… no developers, no users.
>> The obvious solution is then to contract/hire
>> one person to handle the linux package. This has
>> the side effect of more people getting paid to be
>> programmers.
Oh yeah, THAT’s a desirable option to support less than 6% of the available market. Nothing like wasting a full time employees wages for something that for windows takes one person five to ten minutes. (If even) You may want to consider some business math courses.
>>It’s not like there isn’t a boatload of different
>>installation systems for Windows either; Wise,
>>InstallShield etc etc.
That’s just … Wow. #DDD
Regardless of which package manager you use it’s still a single file deployment (in two file extensions that function identically – .exe and .msi) that works on every windows machine out there… We’re talking the difference between being able to pick any of a billion package managers and have ALL OF THEM work in 98/ME/2k/XP/Vista compared to so many flavors and incompatabilites between distros and package managers you need different packages for each and every possible combination of installer just to work native – which is why most of the time they reduce it to three – a source tarball, .deb and .rpm – screw the rest. (and some older software STILL doesn’t do .deb). Hmm, make one .exe that targets 90% of the desktops out there, or make a dependancy list, then compile at least three different .deb, five different .rpm, include a source tarball, get flooded with questions on why you don’t have a slack install…. To hell with that.
>> “Hell, how many different directories can KDE be installed to?”
>> As many as you can think of. You know, just like
>> how many Windows installers will let you chose
>> where to install an application.
Again wow, you completely missed the meaning and implication, much less the context. I make a windows program, I don’t have to worry about /windows being in four or MORE different possible locations, and if that IS going to break my install there’s a environment variable %windir% that will ALWAYS point at it. KDE applications have four different possible locations to contend with, no system variable to read it from.
All I want to do is read/manipulate the damned config file(s), and there’s no system var to point me at it, meaning brute force checking for the existance of those directories… and it’s something you have to code into your installer or end up stuck in dependancy hell. (and linux zealots complain about DLL hell)
You might want to read:
http://distrowatch.com/dwres.php?resource=article-rpm
I particularly like the quote from Shawn Gordon of “The Kompany” –
“Depending on your distribution you’ve got KDE in /usr, /opt/kde, /opt/kde2 or god knows where. For packaging everyone decided to make a new name for the directory between /usr/src and /RPMS, you’ve got ‘redhat’, ‘OpenLinux’, ‘RPM’, ‘rpm’, ‘packages’ and those are just the ones I’ve noticed myself. We are getting to the point that we are seriously considering not supporting distributions that don’t support the LSB. I’ve been very encouraged by SuSE’s work in this direction, and disappointed at how bad Red Hat and Mandrake are.”
and that page is JUST talking about the problem with JUST RPM. That’s one package manager, and even then you have to jump through hoops from hell just to support the distro’s that use it, and even the SUB VERSIONS of each distro. (see the headaches with some GTK applications between FC2 and FC3 – where all new .rpm’s had to be built)
Edited 2007-07-20 10:59
I won't refute your points one by one because, quite frankly, they haven been beaten to death by other posters and YOU know that, but one thing that strikes me on your post was this:
Again wow, you completely missed the meaning and implication, much less the context. I make a windows program, I don’t have to worry about /windows being in four or MORE different possible locations, and if that IS going to break my install there’s a environment variable %windir% that will ALWAYS point at it. KDE applications have four different possible locations to contend with, no system variable to read it from.
Not only Windows application installers (including the OS installer itself) let you pick different paths for installations and have been doing that for quite some time which makes you sound like a hypocrite but KDE also has an environment variable called $KDEDIR that does exactly the same thing as %windir%. And if that is not enough, you can query KDE itself by issuing the command kde-config –prefix on the shell. Either way will give you the result that you are looking for.
Windows knows where it is installed to, KDE knows where it is installed to.
You don’t even have to run a KDE session at the point of querying and you can get you can get directories for all resource types (executables, plugins, application specific data, configs, etc.)
See above. Additionally KDE even provides tools to read and manipulate configuration files.
I’ve been arguing the same point about fragmentation for nearly ten years now all the while being told how I’m wrong, how Linux will rally around vendor supported distributions. Linux zealots refuse to recognize that open source is a double edged sword.
Vendor X merely needs follow the GPL and then they can blame vendor Y for not implementing X’s solution to problem ABC. Vendor Y says “our solution is better” as vendors tend to do. In the end users are left with software that isn’t interoperable. Of course they are free to pay us to implement it which wins support for Linux but doesn’t solve the root problem.
Zealots insist that people want choice – they just can’t understand that people don’t want to be forced to make them. People a good default and the option to replace it if they don’t like it. The constant KDE vs Gnome, Distro X vs. Distro Y has, in part, crippled Linux adoption on the desktop.
Of course the fact that it relies on X11 only makes it worse. Which distribution? Which desktop environment? What window manager?
The proof is in the pudding.
I mean why not include five WMs, desktop environments, browsers, file systems, etc. and make the user pick during installation?
That is a hideous kind of thing to do to all but the most advanced users. The fact that Linux is contrary to its own philosophy in that almost every vendor provides a default and the options thereafter but yet the Linux market just can’t seem to offer that same kind of standard to those on the outside looking in.
Argue till you are blue in the face about it – it won’t change the facts that people only want the option of choice and that nobody likes to be forced to make any decision. It is true that nobody is holding a gun to anyone’s head however we are asking the world to adopt new technology and it wouldn’t be the first time the technical and scientific communities were in error to do so.
can be such a pain in the ass sometimes..
I didn’t really mean font rendering or anything else fancy; I meant the typographical styling within menus etc. Even after a culling, the menus & dialog boxes look ugly, disjointed and unplanned. Now [& as I said earlier] this is not really technical criticism appropriate to OSAlert, so er..I’ll stop now.
You are forgetting one thing: people are actually quite pleased with Windows (at least XP). For them to switch to Linux you have to give them more than XP offers, not the same, or about the same – MORE.
Still, I don’t think that choice will put off a newcomer in a significant way, it’s just a small hurdle.
In my opinion, the biggest problem lies with the software and hardware makers. Think about it: there are 350 distros and maybe 10-15 that really count. Every one of those distros puts out a new version about every 6 to 12 month. I believe it’s a nightmare to keep up with all that, considering that the global marketshare of each distro is somewhere under 1%.
This kind of diversity will make them think twice about releasing drivers for their new hardware or porting their latest piece of software.
Another big problem I see with the large numbers of distros is the duplication of effort. In some way or another, every distro has to reinvent the wheel. It’s worse when you think about the ones that die out in one or two years – that effort is wasted. Now think about all that work going into 1 or 2 big distros. They would blow the competition to bits.
I like how people try to pick it apart.
One person states too many flavors of Linux is a bad thing. Another person seconds that. A third has to pop in and ask who said it was a bad thing despite names obviously attached to the posts.
Once more from the ground up for those slow on the uptake; people do not like to be forced into any sort of action rather they prefer a default with the option to replace it with something else if they don’t like it.
NEARLY EVERY SINGLE LINUX VENDOR AGREES as evidenced by a preferred desktop, preferred WM, preferred filesystem, preferred applications. Ubuntu one of the most popular desktop distros doesn’t ask a user to pick from twenty-five window managers so why should a user have to pick from twenty-five different versions of Linux which suffer from interoperability issues?
Kaiwai,
I disagree that a pure open source environment makes fragmentation a non-issue. That is based on the flawed notion that a community never disagrees. History is replete with examples of supposedly interoperable technologies failing to do so. DVD burners are often known only to burn discs from certain manufacturers for example. Even when the specs are identical problems will exist and are statistically more likely to occur given multiple opportunities to do so. In other words you are more likely to break a leg if you have four of them. This has a very real impact on development. For example Apache’s security record is looking pretty grim compared to recent MS offerings. I am the last person in the world who would want to admit such a thing by the way.
In any event proprietary software has existed for a long time and it isn’t going to just go away. I hate to break it to you but the open source movement didn’t really develop until the 1980s and in truth it took an outside party (Linus) to write the kernel for the FSF/GNU operating system which in fact was merely a clone of … an existing proprietary system.
Unix was originally proprietary. Linux was not here first.
If “people don’t care” then explain why the vast majority of the world has decided to remain with proprietary operating systems even when it costs them real money?
While I can’t say you seemed terribly arrogant your post does seem ignorant of reality.
Apache’s security is an interesting issue; is it due to the fact that there are many people who are surveying the code or a biproduct of sloppy code. Given that there has only been 2 worms in the many years that Apache has been an around, it speaks volumes that as long as Apache is maintained, security patches are made available, distributions provide updates on a regular basis – and most importantly, the security issues are not blown our of proportion to their true impact on the system, any security issues shouldn’t be an issue.
The issue with Microsoft is that they have security holes, which is ok, these things happen. The issue happens when it takes months (WMF vulnerability) before they get fixed or worse, just simply ignored (look at unpatched IE vulnerabilities) hoping they go away. Add that the long time between the security notice and update, some ‘enterprising person’ will release a exploit for it.
Like I’ve said, it isn’t about the security hole, its about how it is handled, and quite frankly, even with Windows Vista released, they haven’t improved very much.
It was never ‘created’ because software was only ever seen as a means to sell hardware – hardware was where the bucks were, the software was the cherry on top of the ice cream as a tempter. There was no ‘value’ in software to the same degree we see today.
Like I’ve said about proprietary software, its not going anywhere, but the software industry is going to go through the ‘diminishing returns’ phase – just take a look at software right now – the only real reason people upgrade is so that they don’t feel ‘outdated’ or find that their ‘old version’ is left without updates and other support.
Take a look at the opensource applications right now – their survival and those companies who base their future on it, isn’t based on pushing products with more and more features but instead the customers willingness to pay for support policies which allow customers who pay to demand features to be added, bugs to be fixed etc. The need for ‘releases’ to make money will become a non-issue.
Considering that at the ‘peak’ of the UNIX wars there was no UNIX standard – even when UNIX95 came out it was very immature and worse still there was a sizeable number who were conforming. When only half the market conforms to the single UNIX specification, it’s asking for trouble.
Lack of marketing muscle for instance. If you could pick up an Office suite that was cheaper than Microsoft Office, had all the features that you needed, and ran on Windows – would you run it over running Microsoft Office?
How many end users *know* about OpenOffice.org/StarOffice? heck, I was chatting to a mate in the US who was excited that he had *just* found OpenOffice.org this year! This is a guy I would call a ‘Windows user’ who would know his way around a computer fairly well.
If this, as an intermediate level user, just found out about a project which has been in existence for over 2 years – one doesn’t hold out much hope for the lowly end user who relies on their ‘expert friends’ and the ‘friendly sales assistant’ for advice on software and hardware matters.
My mother only runs SLED by virtue of the fact that I installed it, showed it to her, and now she is a convert – she can’t believe how easy it is and loves playing the games included with GNOME. She’s happy with GNOME, Banshee, OpenOffice.org and all the other stuff bundled with it.
How so? because I don’t subscribe to the notion that the only way to make money is to keep the source hidden then extort money out of the customer on a regular basis?
Edited 2007-07-20 12:44
so why should a user have to pick from twenty-five different versions of Linux which suffer from interoperability issues?
So likewise we should only have 25 different versions of evrything. Only 25 different cars,houses,underware,food.
Lack of interoperability? Is the ext2/3 filesystem or the tcp/ip protocol stack different between the distros?
While I can’t say you seemed terribly arrogant your post does seem ignorant of reality.
Fortunately so does your post. I’m glad there’s choice. And since nobody forces you to abandon an OS. Let’s not forget the majority of linux distros are being run by volunteers in their spare time. So as of now many of them aren’t allowed to work on a fork because you say there’s to much choice?
redhat/mandriva
suse
slackware/arch/..
debian/ubuntu/..
gentoo
On those you get more or less the same and you can quite confident ignore the rest ;D
I guess one could say it’s same old + gentoo but with lots of stupid useless variants.
That every time I read an article about Linux, apparently I am trying to not use windows? How about I like all three of the big three? Here on osnews, I have been called a Linux zealot, a OSX zealot, and a Windows zealot.
How about they all have programs I use every day? I love quanta, I love the cs3 suite for osx, I use windows for games.
All have good points, and all of them have bad points, in fact, a lot of times, it seems like many people have a “if you aren’t for windows, you are against it.”
Can I be indifferent please? Just ONCE?
Honestly speaking I know what fits my need… I haven’t bothered much to check what other distros are offering after I found my match
I explored 5-6 distros before finalizing. After all freedom of choice is Open Source is all about!!!
forking isnt the issue, merging is…
Kaiwai,
First I’d like to address an earlier point you made in “Solution for too many distros” specifically regarding standardization. It is true that standards processes can slow development of new functionality however it is the only real model to achieve what needs to be achieved in the sense of Linux and offering users one truly better solution. In fact I offer that a standardized Linux distribution supported by multiple vendors would arguably be more secure and able to grow even faster as a result of reduction of duplication of effort. How many thousands of man hours are wasted every year making the same code work in the same way yet slightly differently?
To proceed with the discussion …
Apache is maintained however some issues remain, apparently unpatched, according to some industry sources. In any event despite Vista’s track record MS has arguably improved the situation with IIS. I agree SBO is not a good practice and I believe in disclosure and patches to address problems but I don’t feel this practice is in any way unique to the open source movement though it may light a fire under the asses of some more lazy developers.
Software might not have been seen in the light of MS profits however Unix licenses were not originally cheap.
I’m not convinced those are the only reasons people upgrade. I’ve seen users want genuine new features pay money for a product. Time and time again. In fact new features are about the only justification for a paid update as all maintenance updates should be available free of charge, IMO.
I believe the state of diminishing returns you speak of is already upon us in some regards so I won’t argue the point but as for your open source applications example you are forgetting that the only place Linux has really succeeded is in the server farms. Of course businesses want support – that way they can sue somebody if something goes wrong and they lose money. The same is not entirely true of consumers though, IMO.
On a “lack of marketing muscle” it is true that Microsoft and Apple have larger coffers than most Linux vendors. Why don’t you give their business model some credit here. If everybody wants support and that is what the Linux cash cow really is then how is it they don’t have a marketing advantage over proprietary systems? That said I’ve seen plenty of Linux ads – the problem is they are never about the same product even though at heart they really are. The outside world will not accept such a condition.
In regards to the statement about ignorance it isn’t so much that you don’t subscribe to a single theory but rather that you choose to trumpet your preferred choice over what has panned out in the real world and seem to infer that it is somehow superior despite the fact that with so much raw energy, passion, and effort behind it … still hasn’t really made a dent. The “we were here first” comment had a lot to do with why I said that, as well.
netpython,
ext2 and ext3 implementations on different systems have been known to have interoperability and stability issues, in my own personal experience, as well as that of others. TCP/IP is a bit of a different story but for example the old Mac OS 7 TCP/IP configuration could only support 64 connections maximum; this was before Open Transport. Likewise there have countless documented bugs and problems with various stacks from different vendors.
I’ll take a page from Kaiwai’s book and ask exactly how it is that you think I am being ignorant. What have I ignored in my statements?
I’m not the only one who feels there is “too much choice” as you put it. I see “too much choice” as an opportunity for bugs, duplication of effort, forcing decisions on the user. You are probably the first person who complains when somebody uses Windows and wonders why the world just won’t come to its senses and use your favorite platform. Then you berate the people who take the time to tell you why.
Nonsense. Of the “300 distros” only a very small handful are actually big distros that vie for attention. The rest are primarily custom distros or hobby tinkering.