–Carla Schroder, author of Linux Cookbook
Carla’s rant in response to my review of Vector Linux is well taken if misplaced. Vector Linux has been around since the late ’90s. Her point, though, is very valid. There are literally hundreds of distributions out there if not more. Ryan Lortie made the same point, albeit less clearly, in his article in response to the Free Software Foundation’s Bad Vista Campaign which Chromatic lampooned.
Do we really need a separate distribution like
Carla is right. We don’t need to keep reinventing Linux, creating distributions that put critical bits in interesting and inventive if unusual places. An application written for Linux should be relatively simple to install on any Linux distribution. It ain’t so. Do we really need hundreds of general purpose distributions, all with diffferent tools, different filesystem layouts, variations on three major software package management schemes and a host of oddball ones, and so on? Do we need yet more to crop up?
I can understand why FSF recommends Ututo Linux (a well established if not terribly well known distro) and GNewSense. They match FSF ideology which requires 100% Free Software and ideology is all important to FSF. I’m not at all sure putting that ideology at the forefront helps Linux Torvalds’ half-joking, half-serious goal of “world domination”, though.
Linux has reached somewhere between 30% and 40% of the server market (depending on whose figures you believe) because the major distributions just plain work on most server hardware from major manufacturers like IBM, Hewlett-Packard, and Dell. We need to convince enterprise customers who buy servers to demand hardware that can use Open Source drivers. Until that day arrives proprietary drivers remain a necessary evil in a Linux distribution. That’s just as true on the desktop as it is in the server room. Nobody will migrate from Windows to an OS that doesn’t work for them on the hardware they have.
The nice thing about popular distributions is that you don’t have to sacrifice any deeply held beliefs to use them. If you want an truly free as in freedom system you can choose to have it that way. The laptop I’m writing this on doesn’t need any proprietary software to work as it should. I can have a truly ideologically pristine system and I don’t need to reinvent the wheel or change distros to do it.
This is a silly argument. Who are you, or Carla Schroeder or anybody to tell people not to create their own distros? Who gets to decide which distro is "major" enough so people are allowed to work on it?
All those little guys are having a conversation - hey, let's try another way to make it all work. And once in a while ideas percolate from single developers that on take a life of their own. Knoppix, anyone? I use Ubuntu on the desktop - that didn't even exist a couple of years ago.
Sure, you end up with hundreds of dead distros, but so what? Creating and maintaining them was great schooling in software development, maintenance, systems administration, and project management for hundreds or thousands of people.
Do you complain about hundreds of people learning, say, to play an instrument? All that wasted effort! They should leave music making to the professionals and record companies! They know better!
It's about freedom. Live with it.
kvenlander: Sure, it's about freedom. The freedom to make Linux a confusing, fragmented mess that will drive users away rather than attract them. We have hundreds of live distros, most of which, as Carla points out, are half baked. A newcomer to Linux tries one and judges Linux as a whole by the results.
Who decides? Actually, the user and developer community decide. Distros like Debian, Ubuntu, Knoppix, and Vector Linux, which have attracted large volunteer developer communities, survive, as do ones with large corporate backing like Red Hat/Fedora, Novell/SuSe, and TurboLinux. Without one or the other a distro cannot thrive.
Also, I didn't see Carla or I telling anyone what to do. I saw both of us ranting about a legitimate problem in the Linux community. Can you live with that?
I don't know whether you are right or wrong when you say there are too many Linux distributions (although I'm inclined to disagree). However, I think you are telling people what to do when you say We don't need to keep reinventing Linux. The implied statement that follows is that people should stop creating their own distributions.
I'm also not sure that the argument against reinventing the wheel is a valid one. It seems to me that people reinvent the wheel in various fields all the time; otherwise, we would never have had electric or hybrid cars (let alone cars themselves), or progressed from the use of leeches to treat any disease under the sun.
I also think having any number of smaller distributions allows developers to get ideas out there that would be lost or ignored if proposed under the umbrella of a mega-distribution like Red Hat or SuSE or Mandriva. Personally, my most enjoyable experiences have come when using smaller, perhaps lesser-known distributions (am I am not a coder, scripter, or any form of geek - I just don't like being locked into Windows), such as Feather Linux, Zenwalk (before it became more popular), and Puppy.
With regard to Linux, your argument could be extended to desktop environments and window managers. Do we really need KDE, Gnome, Xfce, IceWM, Fluxbox, etc? Personally, I avoid KDE and Gnome-centric distributions like the plague. I personally would like to see more applications that are desktop environment and window manager-independent. I don't like KDE or Gnome and don't see why I should have to install KDE or Gnome libraries to be able to run certain applications. (QT and GTK are one thing, but when I have to install kdelibs or libgnome this or that, I avoid that application and hope I can find a similar one without those requirements.) Again, it's the idea of being locked in.
Finally, although I don't have the same ideological sensibilities as some others, I can see a purpose, even a need for a distribution like GnuSense. Why should someone have to figure out what to uninstall or have to recompile a kernel (did it once - wasn't comfortable doing it) in order to get the base system they want? My only personal belief is that they shouldn't.
i doubt the FSF cares if its goals are in line with tovalds' jokes. ;)
Home desktop users are so accustomed to there being only one OS, Windows - a monoculture - that suddenly seeing a great diversity of Linux flavors is very confusing and strange - that's entirely understandable. Many distros have irritations and incompleteness, but right now I'm just thinking how glorious it is to be seeing this kaleidoscope of color in choice - like feeding gray pigeons for years and suddenly being shown a forest filled with birds of paradise. This great diversity can be compared to a geological epoch called the Cambrian Explosion, when nature designed all the basic body types we see in living creatures today. Evolution weeded out very many of them, selecting just those fit for function.
We're all eager for Linux to have more cross-flavor standards and compatibility introduced, and in time it will. But right now we're in the first few seconds of the "Big Bang" of Linux, and the chemistry is still bubbling and boiling.
I'm in Linux now for the long haul - but right now I'm sitting back and watching the explosion of color - it's refreshing and invigorating after years of M$ dullness...
DC :)
Walt: I think you missed some of my points. My rant against the proliferation of multitudes of distributions is that they all decide to do things differently. I'm talking about basic, simple things like where in the filesystem they put key bits of software. Despite supposed standards (a couple of conflicting ones, anyway) many distributors just do it differently because they can, not for any good or compelling reason.
Ubuntu, for example, puts X modules in a different place than most current distros. Perhaps it's a Debianism or a new standard everyone should follow. I honestly don't know. I do know that with vanilla Ubuntu (or Kubuntu or Xubuntu) that you can't do:
./configure
make
make install
for an X based app and have any prayer of it working unless you tell configure where to find things. Yes, configuration options can fix that issue. A newcomer won't know that, will they?
GNewSense (it's not GNUSense) would be a nightmare for the typical user coming over from Windows. It would put them off Linux, perhaps permanently, if their very normal hardware simply doesn't work. It would be unusable in a corporate server farm as well. I see building a user base and gaining acceptance as more important than what Carla describes as hundreds of half-baked distros.
I don't think you can extend my argument to window managers. I share your dislike for Gnome and KDE, though I do see KDE as probably the best choice for newcomers to Linux migrating from Windows provided they have a system with the necessary horsepower. My last two distro reviews covered Vector Linux 5.8 Standard (Xfce4 desktop) and Xubuntu Edgy Eft (also Xfce4 desktop).
The need for lightweight distributions for older hardware and corresponding lightweight applications and window managers is pretty well undisputed. My article (rant?) talks about "general purpose distributions", not specialized ones with a distinct purpose not filled by Fedora or Ubuntu or Debian or SuSe. I also didn't put a fixed number on how many distributions there should be. I just said a distro for the sake of yet another distro is a really bad idea, particularly when it breaks standards and makes the life of newcomers to Linux difficult.
FSF doesn't believe that 99% free is free enough. They are ideologically driven. They don't believe in having a choice to include or exclude proprietary software. They believe it is all evil. I just don't share Richard Stallman's view of the software world or all of his ideology. OK, maybe proprietary software is truly evil, but for now it remains a necessary evil. That, too, was one of my main points.
DC: Linux is more than 15 years old. We're well beyond the first few seconds of the big bang to use your analogy. We are becoming more and more like UNIX: one name, lots of mutually incompatible distros. OK, maybe not truly incompatible, but incompatible enough for the newcomer.
Oh, and I think many desktop users are aware of MacOS. I think you'd be surprised how many know of Linux but have been scared off, either by FUD or by a genuinely bad experience. Even if that bad experience was a decade ago they often are still very much afraid that Linux is just too geeky and weird for them.
Anonymous: I agree with you but was it really a joke? I still think Linus was half serious and that the goal was worthy. After all Microsoft dominates the world on the desktop now. Why not have Linux usurp it's position? :)
you are delusional if you think a free Unix clone will ever dominate the desktop market in any meaningful way. but i adore your enthusiasm. :)
Cailtyn, I don't think you need to worry so much, because Darwinism is at work here. The 'user' base of those kind of distributions solely consists of fanatics who doesn't really understand how this harsh world works, and that's an absolute minority. After enough time, those communities would die out and the only thing left is a history record. How many of the hundreds of distros are ABANDONED or nearly abandoned? Time already tells the answer.
Remaining users are divided into 2 categories:
1. Joe users and decision makers, whose only concern is whether somethings works out of box or not. At this point, I would say Ubuntu is very clever, though personally proprietary binary stuff is still a worry for me.
2. Technical users and developers would instead pressurize the original effort to incorporate or at least consider some other solution to the object of dispute, though they understand it might not work immediately. There are already enough people giving their public opinion, enough to the point that non-Jue users are aware of the consequences of using binary blob. So I don't have more to say here.
One point worths noting though: the concern about vendor stopping their support of distribution doesn't only apply to proprietary software or driver. For example, even kernel itself drops obsolete feature or driver slowly, and users of those are also left in the cold because they don't know enough to keep development effort ongoing. The net result is exactly the same as when vendor suddenly drops support of something.
When I said that Linux is in the first few seconds of its "Big Bang", I should have made clearer that I was talking about John Doe who lives in suburbia. I privately teach around 100 Windows strugglers of all ages, and almost without exception when I tell them about Linux it's the first they've ever heard about it. Only in the past 1-2 years is Linux being mentioned by ordinary people, not just gurus.
But in 15 years it should have had time to improve, but development has been painfully slow - yes, that's true. Perhaps you could suggest some ways forward in Linux politics (and it is too factional and political for my liking)...
You have to understand - Linux is a Religion.
Face it, Linux will NEVER be big, at least not when every new comer have to waste the next 10 years to find out
- which distro can do what
- this distro supports this computer better while
- that distro supports that hardware better
We need all this in-fighting to keep Microsoft safe. Else with all these programmers concentrating on a single (few) distro, Linux can grow so fast that hardware cannot catch up and not like it's like now (wireless driver not available here, video driver not available there, SHIT!).
Grow up, Linux is just a toy until all these programmers can wake up.
Some things are getting better. Most major distros make it a point to contribute code upstream to the large development communities. For instance, GNOME regularly receives code submissions from downstream distro developers like Fedora, to give but one example out of many.
Also, the Linux Standards Base (LSB) specifications are gaining wider acceptance.
The goal in both cases is to avoid duplication of effort - all distros benefit from upstream code submission, while developers and users get a more predictable environment when LSB specs are followed.
Please, correct the typo/mistake it's no Utoto, but Ututo is...
I think that while confusing for a new comer the number of distros is actually a good thing. Here is why. Large number of distros means that there are a lot of ideas floating around and being developed tested. When they are a good idea and they reach a certain point a major distro picks the idea up and runs with it. What I think is needed. Is a clear way for us Linux users to decide what the major/best/user friendly distro is. An point all home/new/users in that direction. Unanimous or near-unanimous as the Linux community can get. I mean for example do a search for the best/easiest Linux you are going to come up with so many its confusing for a new user. Thats all just confusing. The diversity of the Linux community is a strength a resource a main-stream distro cannot financially afford to touch. Got an idea try it, if it survives the test of time it was good, if its not yet the time and place it will resurrect.
I have to agree who are you ? And who decides which distro is major ? Linux works because of diversity. If we all depended upon the same distro" moslty everybody would still be using microsoft.
It dosen't matter how easy it is to turn Ubuntu into an all-free distribution. If I can get somebody else to figure out what to do (the hard part) and do the work for me then why shouldn't I?
I think you ignore the benefits of having the thousand and one linux distros around.
First, all of these distros are in the end attempts to do something _different_. Puppy Linux was one of the first ones to run completely from RAM, and for that reason it was blazingly fast. Try asking a Fedora or an OpenSUSE admin to include something like that. The mainstream distros are also slow on the uptake for a lot of new software - FC6 still doesn't have Firefox 2.0 for example - and so there is choice for the users, which directly contributes to lack of frustration.
Second, a number of distributions are created and meant only for targeted and small groups of people, for example, programmers in Bioinformatics. There is great benefit in that, because a relatively new user to Linux can simply be told to install this; and his/her required software will run straight off it. Another classic example is Knopmyth for running a media-center out of a live CD.
Third, and somewhat related to the second, is that the burgeoning numbers of distros are directly indicative of choice and freedom of Linux. I have yet to meet someone who comes over to Linux because it looks good or even because it never crashes - a Windows hater can easily go to Mac OS X. What Mac OS X lacks, however, is complete freedom - what drivers you use, basic system principles (try putting your Users (/home) into another partition on OS X and you'll find its quite tough - and may break with an OS update). People come to Linux in the long term because of software freedom - and nothing says freedom like "whatever you want to do - there's probably a distro that does it".
Fourth, and probably the most important - don't you see that these distros mean there is so much of effort in the community to develop Linux? It's infinitely easier to just join a project than start your own, yet people do it - which means Linux development is in good shape for the years to come. If only a few major distros - those supported by companies - are left at some point, I'll seriously start to doubt the community's interest and/or capability of maintaining Linux anymore.
"Ubuntu can be made truly free by removing the restricted modules package and leaving the multiverse packages disabled. That's all you need to do."
False!. The Ubuntu kernel has binary blobs in it. In addition, it seems that the next version of Ubuntu will include the propietary Nvidia driver by default.
Regards!.
The large number of linux distros makes sense. The more the better. Spare me the excuse of "users get confused". The diversity is allways welcome as each user gets to pick the distribution that fits his habits and needs most.
When I abandoned Windows three years ago I wanted a distribution that will look totally different. I wanted to learn something new from the bottom up. I wanted freedom of choice and no "eulas". Many others leaving Windows behind need something that looks similar to Windows. They found the distribution that fit their taste and needs.
I will never go to another Operating System made by a corporation that indulge to brag about "we know what the users want" (no, they definitely don't and will never know, it is a lie).
Let them Linux distros and other (than mswindows) Operating Systems be.
The more the better. So that everybody can use what they like most.
What I need and like most likely is not what you prefer using.
A few responses to bits and pieces of the comments:
First, for those who say Linux will never be "big" and Microsoft will dominate forever: That was once said of the server room as well. Since 2000 Linux + UNIX are the majority in the server room. Microsoft has been stuck around the 42% mark for years. Linux also isn't a UNIX clone. It was build on a totally new code base and nothing in UNIX (except Mac OS) has anything like the desktops and applications available for Linux.
Fedora doesn't include Firefox 2.0 because it was released before 2.0. Fedora/Red Hat have never been big on releasing new versions just because something new is out there. For them it usually takes a bug or security vulnerability. Other major distros (i.e.: Ubuntu, Slackware, Vector) are very fast to bring out newer versions.
Who am I? Someone who does UNIX/Linux consulting for a living and listens to the concerns of the business user community, the folks who have driven the growth of Linux. Who decides which distro is major? The user community, of course.
A couple of additional responses:
I fixed the link/spelling of Ututo Linux. Thank you for pointing that out.
If people took the time to read the article and my responses you'd know that I was talking about hundreds of general purpose distros that are little more than multiplication of effort, hence "reinventing the wheel". Specialty distros, whether geared for bioinformatics or geared for pen drives and other MTD devices, are specialized items and are distinctly NOT what I am complaining about. I also doubt they are what Carla is complaining about though I can't speak for her.
Yes, Puppy Linux can be loaded fully into RAM making it one of a number of distros that are particularly valuable when using MTD devices. Others include Austrumi, Damn Small Linux (DSL), DSL-N, and Flash Linux. Since I can count the number of such distros on the fingers of my hands and have fingers left over I hardly think they are the problem.
Caitlyn's making a good point. Big deal if there are 1001 and Linux distributions, when most of them are minor, pointless variations of the same thing. I love having all sorts of different specialized Linuxes to choose from for different purposes- tiny ones like Puppy and Pyramid, big fat ones like Debian with over 15,000 packages; Linux for routers, wireless access points, fileservers, web servers; Linux for sound studios; Linux rescue CDs and USB keys; Linux for secure easy Internet kiosks; 100% Free Linuxes (hardly any of those); Linuxes loaded down with all manner of proprietary guff. That's cool stuff.
But we don't need dumb stuff like Red Hat or Debian or Puppy or Knoppix released under different names and with different artwork. Caitlyn said" An application written for Linux should be relatively simple to install on any Linux distribution." Darn right. The Linux Standard Base has been around how many years now? Lots of them. Are we any closer to having a modicum of sanity with regard to binary compatibility? No.
I'd like to see more Hardware Compatibility Lists that name hardware that works well with Free drivers, instead of jumping on the bling bandwagon like it's the only answer. I'd like to see more cooperation, instead of imitating the proprietary world's "not invented here" sickness. I'd like to see more decent howtos, instead of newer differenter lamer GUIs that still don't solve the problem, which is lack of documentation.
I also want to lose weight by drinking beer and eating chocolate, but I'm not holding my breath :)
I'll take Carla's point one step further. A while back I played with Ehad Linux, a localized Hebrew/English version of Mandriva One 2006. Nice idea for Hebrew speakers, but... wouldn't it have been better to include this into the Mandriva project somehow? Couldn't that have been worked out? Now we've got this cool Hebrew version that's based on an older release. Mandriva 2007 has been out for months and still no word on a new Ehad Linux 2007 version.
When I reviewed Vector Linux I complained about internationalization and localization being lacking. What I've done is to contribute to that project. I have packages waiting to go into the testing repository that I've just uploaded that will start to solve the problems for languages I care about. Hopefully other volunteers will do the same for other languages they care about. Tell me: would I have been better off creating a Hebrew or French spinoff distro? I think not. The whole point is to make the distros that are strong now even stronger.
I also agree with Carla about HCLs. However until free drivers are important to enterprise users I fear those of us who consult on Linux for a living are stuck with those binary drivers. Carla's answer, plus more reverse engineering projects, may well be the long term solution.
It's not just about choice, they tend to be good for different things. Each new distribution gives the community something. And there's a really simple solution to making it so everybody can't just up and create a new distribution, close the source, make it so that not everyone can be a developer. I really wonder if any thought goes into the "fewer distributions" argument, as often as it comes up. Lets make a list of people we don't like, or don't think are doing anything significant and tell them to stop making their Linux distro. Novell's friends with people we don't like so they can't. Patrick Volkerding's using an ancient set of install scripts and an unprofessional sounding name like Slackware, so he can't. Red Hat sold out the community, and won't let us download their pre-packaged enterprise product for free any more, so they can't. Gentoo's from source gimmick is silly, so they can't. Debian spends to much time and energy on silly policital and social issues, so they can't. Xandros and Linspire are just trying to replace the de facto standard desktop, so they can't. Ubuntu and Mepis are really just nice repackaging of Debian, so they can't. Live-CDs only address a tiny niche market, no more DSL or Knoppix. I really can't think of a single benefit to the community for reducing the number of distributions. It would be less confusing, or easier to write software? Use RedHat (or whichever other arbitrary distro you decide on), you get exactly the same benefits from only interacting with one distro as you do from only having one. That's really what open source means isn't it, the freedom to customize and share software?
nothing in UNIX (except Mac OS) has anything like the desktops and applications available for Linux.
it's the api and base userland that make an os unix, not the desktop and applications. and if there is indeed a desktop/application that only supports linux, it's a piece of crap and so is it's author.
The point most of you seem to be missing or else totally ignoring is that a confusing multiplicity of *similar* general purpose distros, many of them poorly implemented and poorly thought out, hurts Linux and those of us in the Linux community who would like to see wider acceptance and adoption of Linux. I don't want to take away anyone's freedom. I do want people, including would-be distributors, to think about the consequences of their actions.
I see a whole lot of missing the big picture in the comments so far.
I totally get what Carla and Cait are saying and it does make sense. You do not need to close the source nor limit the different distros going about. What you do need is to have set standards so that a new program can install on just about all of them.
time is wasted in the corp enviro when so much time is spent hunting down wayward files/folders. The home user, looking to install something neat, use it for a time, and then head to soccer practise will just opt to either not install something they saw as useful. this leads to that person telling others that Linux is just for hardcore geeks and no one else.
So what? Well this person knows the CEO of a small company that does about $30 Million USD a year. The IT staff at the company has been pushing for a linux server or perhaps linux on the desktop. The problem occurs when the CEO takes the word of the "informed friend" who has never steered him wrong over the nicely written papers on why Linux would improve the workspace. Hiring just "hardcore geeks" is viewed as being more expensive than just about anyone who has supported Windows. So the CEO makes a poor decision and the IT group pulls their hair out in frustration as multiple projects take longer to complete and suck up more resources.
I have dealt with the above. Yes, I left there and will never go back to another shop that is pure MS. I did manage to have two linux servers (RH 7.X) up and running by the time I left. They immediately formatted them and installed Windows 2000 server on both boxes.
Linux does not have to be just like Windows in every aspect. However, there is a need for some standards and that could lead to even more growth. Will Linux ever overtake MS on the desktop? I have no idea but I have seen stranger things happen. Nothing lasts forever.
Wait..we (or you) are missing something, are we talking about linux from scratch or forking? Because if out there are 100 buddies, with 100 great ideas of developing a distribution, and managing it in a total different way of the rest of 99, then I'm more than happy to see 100 distros. But if we are talking about forking (Ubuntu), thats just a bad idea.
Ms. Martin wrote above: "Who decides which distro is major? The user community, of course."
Actually, in a larger sense, the distro "deciders" may not just be the amorphous "user community", but are more likely to be a blend of end-users, developers, and maybe not so surprisingly, market forces.
As the saying goes, "Competition makes the product better."
The recent Dr. Dobbs' Journal online article 'How To Tell The Open Source Winners From The Losers', href=http://www.ddj.com/dept/opensource/197003071?pgno=1 sheds further light on this principle of competition and how it improves open source software (and presumably the resultant quality of Linux distributions!)
The 9-point checklist in the DDJ article highlights the key factors to consider when evaluating a successful open source project such as a Linux distro, and is reproduced here:
--- Begin quote ---
A 9-point checklist before considering using open source.
A thriving community. A handful of lead developers, a large body of contributors, and a substantial--or at least motivated--user group offering ideas.
Disruptive goals. Does something notably better than commercial code. Free isn't enough.
A benevolent dictator. Leader who can inspire and guide developers, asking the right questions and letting only the right code in.
Transparency. Decisions are made openly, with threads of discussion, active mailing list, and negative and positive comments aired.
Civility. Strong forums police against personal attacks or niggling issues, focus on big goals.
Documentation. What good's a project that can't be implemented by those outside its development?
Employed developers. The key developers need to work on it full time.
A clear license. Some are very business friendly, others clear as mud.
Commercial support. Companies need more than e-mail support from volunteers. Is there a solid company employing people you can call?
--- End quote ---
SOME of the 1000+ Linux distros and their variants will fade and disappear outright over the next five years, just as SOME of the 139,000+ open source projects on SourceFourge lacking one or more of these 9 key factors will also fade and disappear outright over the same time-course.
To the person writing this particular reply, Caitlyn, Carla and the above commentators taken together seem to be expressing in each of their own ways, many (if not most) of the same basic points as this 9-point checklist.
oye: Most new distros are created by forking. GNewSense is a perfect example. It is, as you say, just a fork of Ubuntu created for ideological purposes.
elibnotlibc: I agree with most but not all of your quoted nine points. For example, a distro doesn't have to have paid support. Debian doesn't and is in no danger of disappearing. It certainly qualifies as a major distro. There are, of course, any of a number of consulting firms who will happily support Debian or Slackware or any other major distro where commercial support is lacking for a fee.
Good discussion.
Apparently elibnotlibc likely spotted nycace36's bottom-most Linux-Watch Talkback at http://www.linux-watch.com/cgi-bin/board/UltraBoard.pl?Action=ShowPost&Board=talkbacks&Post=212&Idle=0&Sort=0&Order=Descend&Page=0&Session= Very good! :-) Perhaps you and Ms. Martin would also be interested in viewing the Linux-Watch Talkback at http://www.linux-watch.com/cgi-bin/board/UltraBoard.pl?Action=ShowPost&Board=talkbacks&Post=170&Idle=0&Sort=0&Order=Descend&Page=0&Session=?
In this latter Talkback, the points are emphasized that Slackware and Debian have different models as far as Linux distributions are concerned.
Model A: The benevolent dictator, e.g., Patrick Volkerding of Slackware Linux
Model B: The squabbling "free-for-all" that has characterized Debian since Ian Murdoch and associates cobbled together one of the first GNU/Linux distributions.
Can we take these two contrasting A & B models even higher than the nine-point checklist and even this original blog's subject of 'So Many Distros, So Little Time' ?
How about applying these models to actual UNIX-based Operating Systems themselves ?
In the A model of GNU/Linux, Linus Torvalds has been the one benevolent dictator of the Linux kernel itself since its inception, despite the multitude of resultant distros developed on top of it.
In the B model of BSD variants such as FreeBSD (highlighted at http://ezine.daemonnews.org/200104/bsd_family.html), there is a group of Committers and a Core Team in charge of maintaining the entire BSD tree.
The closest constructs to Linux distributions in the BSD world are the three operating systems FreeBSD, NetBSD and OpenBSD themselves. Obviously three BSD operating systems is VASTLY different than the 1000 Linux distros ranted upon and defended in this blog.
These differing models each have their benefits and drawbacks, as Ms. Martin and others will readily attest.
There are number reasons linux is like this, main reason is it's simply bloatware. Simple really, linux should invented the loose lead system. Way I see it the problem arise in the kernel itself, no matter how many times it has been updated. It still lacks the fundamentals to rival competition, aside from that having an open source nature it lacks time and care. People seem to find it more rewarding to make craft dinner over and over again, not really filling in the loop holes the original Michaelangelo started out with. It's ugly really takes the cake when you find that, aside from having a gui front end its rear end is still showing lol. Dressing the cake really doesn't do it any justice, it still behaves like its something left to dos era. On the corporate end it is actually as expensive to to license Linux as much as Microsoft's Windows, only difference is seating license is more open in Linux than Microsoft.
Soon technology will surpass us so far ahead we will loose the keys to the doors we created!!
I'm a Linux n00b, and frankly glad that there are hundreds of distros to choose from. That, to me, says that the culture of the community works.
There's bound to be something (or multiple things) that I don't like about each distro, so it's very nice to know that I can modify it and make it my own. I'll likely do so, and publish the distro for the rest of the community to see. Many won't like the changes I've made, and even more won't ever even bother to look into it. That's completely fine. I didn't make it for them, I made it to enhance my skills and to suite my needs. And yet, even though I'm in it for myself, the Linux community and message is served -- it's all about freedom.
ok,first off i admit i am just above bieing a n00b, i started using linux with puppylinux and knoppix and a few other liveCD distros, and i can say that getting linux to do what i want it to do the way i want it to do was not much different than windows, only without constant threat of malicious code getting into my machine, so now i use fedora 6 zod 90% of the time, and windows the other 10%. i used M$ products my whole life and could definitely be considered a windows power user, but the biggest reason i migrated to linux is that i simply got bored with windows.
so to sum it up, windows is like a base model chevy cavileer, sure, it drives but it breaks down alot and is high maintence, while linux is like, well, a pile of high-preformance racing parts, i spend 3 days setting up fedora 6 to do most of what i need, something you put together and customize vs something that works right out the box. that is the main flaw that holds linux back from bieing serious contender to m$, however m$'s days are numbered, since vista's release that i am convinced. everyone i know who has installed or bought a new pc with vista pre-installed oem has hated it, and done one of 3 things.
(1)those with alot of money to spare have gone to mac's
(2)those who just want something that works got rid of vista in favor of XP or 2000
(3)those who have better knowledge and are not dependent on a certain program for work have gone to linux or bsd unix
so i see good proability that apple will make a version of its os to work on x86 and 64 platforms in the near future, that will offer more compatibility for linux code, therfore making more software available to run on nix
one more problem is see is that i hate to say this, most programmers and users in the linux community are too driven by ideology, ther must be some way to make your efforts pay off or somehow make enough to sustain your work, and also, ogg sucks, i prefer mp3 over it anytime, which yes, its patented but also did not fraunhoffer say that they would not take legal action against makers of free software, well the whole proprietary drivers and restricted format 2 bit encryption is bullshit anyway. but then you got distros like linspire, who the hell wants to pay for something like linspire that hasnt bult a reputation for itself since practically nobody used it yet.l cant really comment on it since i am not willing to shell out the $40 or however much the cost is only to find its not much different than something i can use for free, you think they might be more successful if they allowed private individuals to use it for personal use free of charge and only required those who use it as a cheap alternative to windows when building new PC's and those who will use it for their business.
the makers of linux need to get together for some common protocols for compatability for installing new software and drivers at least, and it really be nice if i could figure out how to keep my desktop, personal documents and files in such a manner that i could run more than one distro that all use the same desktop, and i am sure there is a way, i did it on a dual boot windows 98/xp machine.
i think the future lies in these crazy "derivitives" of unix, after all, doesnt M$'s software have soom roots in unix, its just one thing compiled from the next, that is unless bill gates's dirty tricks and money-power win in the end and tryany will reign supreme. seed novell/suse's "open-source" code with proprietary code and overturn the validity of the gnu/gpl and therfore own linux and do away with "free software" as we know it, now i am going to bed
[quote]"He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me. That ideas should freely spread from one to another over the globe, for the moral and mutual instruction of man, and improvement of his condition, seems to have been peculiarly and benevolently designed by nature, when she made them, like fire, expansible over all space, without lessening their density at any point, and like the air in which we breathe, move, and have our physical being, incapable of confinement or exclusive appropriation."
Thomas Jefferson, in Writings of Thomas Jefferson, vol. 6, H.A. Washington, Ed.,1854, pp. 180-181.[/quote]
After reading these posts, it seems that nobody realizes that everyone here is in the
As a Windows user for over 10 years, I was having a conversation with a friend last night, while he was installing X version Ubuntu in his laptop. I was kind of interested in moving to Linux, because I displease Microsoft, and I hate Windows Vista and refuse to move.
But after going to look for a version and found all this "distros" to choose from I just become scared. Call me single minded, call me a siple person that cannot be as versed in Linux or "religious Linux fanatic". Call me not so intelligent ,I don't care...
But it's true, I will have too keep with Windows XP, because I can simply find what I'm looking for and run all my hardware without spending my whole life trying to find the right "distro". Funny name, sounds like "bistro".
I would love to see Microsoft to be just a sad shadow in History, and I would love to see that "Big Bang" happen with Linux.
But what I've learned is that it is usual for people working on Linux, that they are so brilliant, intelligent, or geeky, that at the same time become ARROGANT, and start missing the point within their own genius. And preted that they a carriers of the only turth about Linux, their religion, and makes averyone else "us, the common people" feel unwelcome.
Just a simple fool guy like me... asking you to unify your efforts and crash Microsoft. Is that too hard for you Linux community?
Oscar: Linux will never "unify" nor should it. Choice is a good thing. Half-baked distros (short for distributions) are not and that is what I was complaining about.
I think if you tried any of the major, well supported, popular Linux distributions you would be quite happy with it if you give yourself a little time (a few months) to learn and become comfortable. The distribution really doesn't matter that much. Think of it like McDonalds, Burger King, and Wendys: they all serve hamburgers. Which is best is a matter of taste.
Good user friendly distros would include Kubuntu (just Ubuntu with a very friendly KDE desktop), Mandriva, SuSe, or Fedora. Pick one. If you have a friend running Ubunutu who can help you when you have questions then Ubuntu or Kubunutu are probably good choices for you.
Obviously simplicity is too big a challenge for many people.
Ubuntu is why I use Linux.
Next question?
chriss
I realize that I'm decades late (in Internet time) with this comment but I just finally found this article.
I have been using Linux since November 1991. Back then the only "distro" was HJ Lu's Boot/Root 5.25" floppies. A friend of mine wrote the first, in when would eventually become known as HOW-TO's, on getting Linux to boot from the HD. It involved using a hex editor on the MBR and some other fun hacks.
I remember the very first "Linux" book ever published (Running Linux by Matt Welsh. I remember the first magazine, Linux Journal (my subscription started with issue #2). I attended the first Linux specific track at a conference (OpenSource/FedUNIX in '93, if I remember right) and was even a presenter of a session at the following year's conference Linux Track.
What does all this mean? How does it fit in the topicality of this post? I've seen Linux grow from it's birth. I remember the sheer exultation of when the first 1.0 kernel was released. I remember MCC Interim, TAMU-2A, SLS Linux, Yggdrasil and the very beginnings of multiple distributions. I remember the very first "distro war" when Slackware came out to challenge SLS.
Now we have anywhere between 300 and 500 choices on distros. There are the specific ones targeted for specific audiences and some handful of niche distros but the landscape for general purpose server and/or desktops is slightly inundated. It's good to have choices but the point Caitlyn Martin is making about resource usage is very valid. There should be a balance between having, as Carla Schroder was quoted, "a half-dozen good solid ones" and the freedom that one has for creating their own personalized distro.
Just because you can do something dosen't mean you should do it.
I concur with your views : http://screwgoth.blogspex.com/2007/04/16/drowning-in-distros/
I'm a sysadmin living in Montpezat (South France), and currently in charge of public libraries and town halls in eleven villages or small towns around here. My job consists of installing servers and desktops there, and networking everybody, everything being 100% GNU/Linux and 100% open source, from the OS to the public library management software.
I started Linux in 2001, with a Slackware 7.1 CD on a battered 486. After that, I went through some serious distribution testing. Must have tested something like 20 distros, if not more (Debian, Gentoo, Arch, LFS, Crux, SuSE, Mandr{ake,iva}, KateOS, Vector, Zenwalk, Fedora, and so on), I got quite proficient with most of them, but I eventually returned to Slackware for a few years.
I fully agree with your argument. Most of the time I can't help thinking that I'd wish that *all* the half-baked distributions would just fall into oblivion and disappear from the face of the earth. Hell is paved with good intentions, a french saying goes, and I would add: it's also paved with shitty Linux distributions.
My choice for the remaining ones, the ones good for getting Real Work(tm) done: 1) CentOS 2) Slackware 3) Debian. Plus eventually 4) Vector, if only they did some internationalization effort.
Our main database server is running CentOS 5.1. The client desktops in the libraries are also running CentOS 5.1, with the XFCE desktop and a handful of carefully selected apps. Been running for a year and a half now, and not a single complaint.
@Niki: Internationalization in Vector Linux has improved greatly in 5.9. All the aspell dictionaries and Mozilla language packs for Firefox, Sunbird, and Thunderbird are in the repository now. Fonts for just about anything in the world you may need are there. Like Slackware there is still no easy GUI tool for changing language settings but at least everything you need is there.
@Caitlyn. Thanks for your response! I admit it's been quite some time since I gave Vector a shot. I have a spare laptop here that I always use for quick tests, so I'll follow your suggestion.
Currently I'm running CentOS 5.1 on both servers and desktops. CentOS has a very clean implementation of XFCE (latest: 4.4.2) in the [extra] repository. The text-mode installer ('linux text' on boot) is ncurses-based and much less demanding in terms of RAM (128 MB instead of 512 for full-blown graphics). I've done quite some tweaking to CentOS, which permits me to be as flexible as in good old Slackware times, but with considerably less work.
What appeals me: 1) it's an Enterprise-class Linux 2) Yum has some really nifty options (especially groupinstall and groupremove) 3) It's very easy to create your own Yum repositories 4) Packages that are not included in CentOS are either in the RPMForge repo (very clean RPMS), and if they're not, they can always be rebuilt from a Fedora SRPM: never failed to build a single package. And 5) it's not as bloated as people say. In idle mode, my server needs only 16 MB of RAM, my XFCE-based desktop only 70 MB.
cheers from the sunny South of France!
The fact that there are a billion linux distributions (slight over-estimation), and possibly as many if not more open source or variant software projects out there is heart warming, I am no longer a hacker so I have no voice. But, I have always believed in unity in diversity, and this must work in the business world and the computer world and the physical human world of politics as well. Freedom doesn't just mean, you can change the code if you wish but only if you change it in the way that we want you to change it. In England there is an evil that has perpetuated society since the 1990's it was called poltical correctness, what it amounted to regardless of the ideas of unity, diversity and acceptance that it grew out of, was thought control, you may only think in this way, not that far from you may only code in this way, or the earlier promise that, any individual who knows more about a computer and it's wonderfull expanse of network than the manufacturers or the governments is a hacker and therefore a criminal. I say no. Let Linux like FreeBsd and other smaller unix mutations carry on expanding, the universe is infinite, I like to see an infinite variety of wild life walking around this planet, I think it is healthy if in the virtual environment of computers, there is also an infinite number of choices, although I think more pressure should be put on the evil empire to free the code of all of it's operating systems as well, any computer code that is closed, is only closed because of a law or a piece of paper in a patent office, even linux distributions insist on liscences instead of acknowledgements of authorship. Open all the code that ever was closed, and keep up the evolution, unfortunately as to making linux work better, I have a moral.
Once when I was still old, I switched from beos r5 or 4 and a complicated tangle of x fvwm unix self built non discript and torturous confusion which was possibly an operating system but had been bunged together over many years and barely kept the old machine it was on running using celotape and possibly glue, tisk tisk, windows 95 had also found it's way onto the machine, and I know that to some that is an evil word but I didn't pay for it so, who knows whether I can be truly moral about that. But I built a new computer. I bought new hardware, I found linux, and spent several years installing tweaking polishing, uninstalling, getting annoyed and hacking rc after rc file, until one day I found mandrake 7. something, it was nice, it was simple, not all hardware worked with it but the kernel which is really what linux actually is was improving beyond my expectations. about a year later, new software could only be installed by self compiling, the old ./configure make make install etc, we should all know the drill by now, I had no problems with this. Until, some new software that I downloaded, source code of course (which I needed to use for something I was doing), could not compile due to two significant factors. KDE and GNOME, QT and GTK, two sides of the same coin warring against each other and themselves and not peacefully coexisting but being counter productive. My highly hacked and personalised mandrake 7. something was out of date, that doesn't mean that the kernel was, but everything else, oh it will still work, I will be able to write this or that gimp 1.2 is a very good program etc, now I am using opensuse 10.1 I refuse to use yast, so do everything by hand or nano and it is shaping up nicely, still no favoured kde 1 or gnome 1 to mix and match, no kfm to be used as a file manager wherever you want to with whatever interface you are using, no. Instead? konqueror the file manager that can be a web browser and other stuff or nautilus which can do as much but not view extract or edit the contents of a simple cpio archive, no we can't quite do that yet. 7zip an open source program for windows can. I was not being sarcastic. and here is the moral, has the operating system improved, yes the kernel was good strong robust capable of running many forms of hardware, but was a bit buggy on usb 2.0 devices using an old usb 1.0 hub, a microsoft answer would be buy a newer hub, but I am an environmentalist as well and will keep old technology running for as long as it works. But and here is the nub or summit, I may as well have been windows xp or windows vista when I got it and first installed it, and I wonder whether we are all missing a point here, I have seen and used ubuntu, I disliked it, I liked it's philosophy, I just didn't want to use something that thought looked and felt like a microsoft product, auto update after auto update etc. and all new linux distributions are painfully trying to copy the big and simple xp and vista way of doing things, even the all important manuals are getting thinner and thinner, and dummer and dummer. And I know the horrific reason for this dumming down, poach more disatisfied windows users. Whilst never really thinking of the consiquences. I see the consiquences every where I look. People use computers every day for work and pleasure, but more and more of them have no idea how they work or why they work.
And the worst is this, I have actually watched a check out girl in a shop use her barcode reader to read the wrong barcode and thus overcharge someone for it and then when confronted with the mistake her answer was THAT'S WHAT THE MACHINE SAYS THE PRICE IS, MACHINE'S CAN'T LIE. This is a quagmire. Because now we have to reducate the world. A computer is just a tool, linux is just an operating system, the more versions you use, the more you will know, the more complicated the software the more you will know, the more confusing the choise, the more alive your brain will be, never stop learning, never stop changing, never stop questioning idealogy of any kind, they are all suspect. Linux used to stand for intelligence and a bunch of other stuff that gnu and mit and such as concepts also stood for, I t would be nice to see it standing for such things again sometime in the near future before the gurgling brains of an entire generation ask, dur, how do you switch the ipod on grandad? Sorry for ranting.
I've used windows all my life and I've never had a problem. I install onto a hard drive, install drivers, and omg it's instant! No repositories, no Synaptic, I don't receive 5 million error messages telling me why over one hundred thousand linux developers ,that are "smarter" than Windows dev's, can't make drivers that you double click and click next about 5 times. My PC can play the Half Life 2 engine at about 140 frames per second on max settings. But my PC can also watch linux try to figure out what a video game is, then cry because the developers are too wrapped up in competing with each others distros to even care about the migration eager users. The bottom line is Windows XP Pro SP2 costs nothing, Vista is out and big bad Bill doesnt give two shits if you pirate it. Linux loses due to too many distros that are too complicated.
@goatboy: I've supported both Linux and Windows professionally for more years than I care to admit. I want no part of the virii, trojans, spyware, and assorted other malware that is endemic to a Windows environment. I don't want to deal with all the security issues. If you seriously believe that Microsoft doesn't care about software piracy you are seriously deluded. Linux, for me, just works. Anyone can pick one of the major distributions and just stick with it.
Windows just works? Tell that to all the people who installed Windows Vista SP1 and got a truly nasty surprise when their systems no longer worked. As has been widely reported for the first time in many years Microsoft is losing some of it's share of the desktop market to a combination of MacOS and Linux. The Asus EeePC subnotebook has been a huge success and has spawned a ton of imitators. Linux loses? I don't think so.
Look, nobody is forcing you to run Linux. If you're happy with Windows that's wonderful for you. Don't expect to convince many people on a Linux blog that Windows is just wonderful. It isn't.
@Nick: It's probably no surprise to you that I disagree with you about Linux and I also disagree with you about political correctness. Being respectful of people and not being deliberately offensive is a GOOD THING. Trashing political correctness is used as an excuse for all sorts of bigotry nowadays. I'm NOT accusing you of any such thing. I'm just saying political correctness does have it's place. Can it be taken to extremes? Sure it can and if that is what you meant to rant about then I'll probably end up agreeing with you. I don't believe in throwing out the baby with the bathwater.
Linux used to equate with intelligence? With MIT? Perhaps in some ways, yes, but I still believe Linus Torvalds half-joking goal of world domination is a good one. Linux can be and should be easy to use for ordianry folks and yet it can still offer the ability to get under the hood, customize, and do special things for those who are more skilled. It's not an either/or proposition.
I still believe having tons of half baked distributions is off-putting to those not in the Linux community and not particularly helpful to anyone in the community. If someone has something truly innovative then of course I want it developed.
Does that always have to mean yet another distro? If a new distro is truly innovative I'll champion it. I've written reviews that were mainly positive about distributions like Wolvix and AliXe. I'll probably end up doing a similarly positive one about GoblinX. For every distro like those there are probably five that offer nothing new or special. That's the main point. Who benefits from that?