Unified Linux Thread

[ShadowVlican]

wow cool thread..... let me start by saying i haven't read anything yet (first thing i did was vote for a unified distro)

i should clarify that i want a unified DESKTOP distro.... a distro for the masses... one to potentially replace windows :)

without reading the thread, i can predict that many here have already debated to death the advantages and disadvantages of a unified distro, so i will go ahead and possibly repeat them from my point of view:

advantages:
a standard is created. there is no standard in linux as of yet... there are to many programs that do the same things (although they slightly differ from each other... which is a waste of programming resources)

collaboration between all programmers instead of each of them doing their own thing

disadvantages:
less choice (though it may help to ease the confusion among linux adapters... we've all heard "which distro is the best" and answered by linux snobs "try them all" or "whichever fits your needs")

i haven't the time to write more, but i'm guessing these views have already been expressed

i'll just repeat that i'd favour a unified linux... or at least less flavours...

[beefcurry]

Coperation is generally better then competition. We need to work togeather to bring down Apple/Microsoft. we can do it :D

[23meg]

That's funny, actually, because with the move to GIT the kernel is anything but unified... a lot of different people have their own trees.

Heck, if the kernel counts as "unified," then everything is "unified." GNOME, X.org, KDE, Samba, OpenOffice, FireFox, etc are all "unified."

Funny indeed. I was just reading an interview with Ben Collins (http://behindubuntu.org/interviews/BenCollins/), Ubuntu's kernel maintainer, and see what he says:

To what extent is there collaboration between the kernel maintainers in different distributions?

The kernel is very distribution specific. Different distributions have different policies on what they do with their kernel. Some distributions will remain as close to stock as possible, even with known bugs, while others are willing to integrate large patches, changing lots of core functionality, varying greatly from stock kernels. Also, distributions have different takes on whether they support drivers in their kernels that are not in stock kernels. This could be proprietary drivers and firmware, or GPL drivers that are not yet integrated in the stock kernel. So while I try to make Ubuntu's kernel as available as possible for other distributions to use, it's very doubtful that our policies will closely coincide with others. One area where we do overlap and collaborate is in kernel security vulnerabilities. Debian and Ubuntu have a common subversion repo to track our outstanding issues, and share patches.

Besides, this whole "unified Linux" (non)issue has been about whether the whole workforce would be better utilized if there were less distributions, or even one (!) "unified" distribution, and I can't relate the assumption that a certain project is "unified" in itself to that.

In other words, what the hell does "unification" mean on a per project basis?

[beefcurry]

A strong definition of unified is required here.

[kornelix]

Some of the comments about the "unified" kernel were better informed than my own, and I am thankful for this. But that only makes the problem deeper.

If you look at the Global State of Linux, you have to admit the following:
* hundreds of distros, perhaps 20 that have significant mindshare
* they are mutually incompatible: you cannot run the same binary apps (the distro publisher does the work of tayloring the most popular apps)
* the apps are 90% overlapping in functionality (a huge waste of talent and money)
* the average user cannot deal with this, and has no desire to do so
* potential suppliers like Dell and HP do not have a supportable platform
* the technical powers that do the work like things this way, because they value freedom over standardization, or because they have more fun inventing new wheels rather than make someone else's wheel roll smoother.

Those who say the current status is OK, or cannot be changed, need to admit that Linux will stay at 1% market share and MS will continue to dominate. Those who wish otherwise are dreaming.

It reminds me of the chaotic state of PCs (hardware and software) before IBM set the standard around 1980. IBM was able to do this because the world was crying for it. I think this will ultimately happen to Linux as well. Perhaps it will take an IBM or Dell or HP to make the plunge, set the standard, and stop the chaos.

[koenn]

I agree with 23meg that this is a non issue so I'm not going to waste my time on a point by point rebuttal of your list.
Just a few comments:
* the apps are 90% overlapping in functionality (a huge waste of talent and money)
* the technical powers that do the work like things this way, because they value freedom over standardization, or because they have more fun inventing new wheels rather than make someone else's wheel roll smoother.
Unless you know of a way to force volunteers to work on projects other than the project they choose, this is a given, and you point is moot

As for "they value freedom over standardization" - I believe that to be incorrect. I made that point before but you seem to have missed it.

It reminds me of the chaotic state of PCs (hardware and software) before IBM set the standard around 1980. IBM was able to do this because the world was crying for it.
I remember those days, and the IBM compatible PC did not become the standard because of cries for a "unified PC". The market evolved towards IBM compatible PC's because of the availability of software and the interoprability between IBM_PCs (and "IBM-clones") . When the IBM pc got into the corporate market, home-users would also tend to start bying an IBM-compatible, if only because it would run the software they got used to at work, and to be able to work at home on files created at work. This is very much an evolutionary scenario : the IBM-pc was the one most fit to the environment, and the non-IBM-compatibles gradually became obsolete.


Those who say the current status is OK, or cannot be changed, need to admit that Linux will stay at 1% market share and MS will continue to dominate.
That's just an opinion - I don't see you offering any proof that Linux's market share and/or MS's marketshare are in any way related to or caused by the fact that there is more than 1 Linux distribution.
So I don't feel the need to 'admit' anything.

[aysiu]

That's just an opinion - I don't see you offering any proof that Linux's market share and/or MS's marketshare are in any way related to or caused by the fact that there is more than 1 Linux distribution.
So I don't feel the need to 'admit' anything. In fact, if Apple has shown us anything, it's that all those "if Linux just did ______, then it would beat Windows for marketshare" arguments have no basis in reality. Apple is unified. Apple has good hardware support (for its own hardware, but it also gets some support from third parties). Macs run commercial software. Macs have all majors tasks point-and-click (no terminal required).

Still, even optimistic statistics put Mac desktop marketshare in the single digits.

Unification, "ease of use," preinstallation, commercial software, third-party support... these are all just red herrings. Windows has one major thing that both Mac and Linux do not have--momentum and market saturation. Those are hard to get.

[euler_fan]

I agree that market saturation and inertia are big advantages and hard to get.

At the same time, I saw this on slashdot:

Linux Starts to Find Home on Desktops (http://online.wsj.com/public/article/SB117374336173334742-0t8WLwdFNFNhmcynmJ8dS7hIX1g_20070320.html?mod=blog s) by WSJ. It raises interesting possibilities.

Insofar as the merits of a unified distro it is not so interesting, but considering where the conversation has turned . . .

[kornelix]

<<
What you describe as benefits of a dispersed linux are the benefits that could be acheived at a much greater volume if a unified distribution were to be created.
>>
I agree with you, but we are in the minority among Linux users, and especially among developers. Being creative or having a different opinion means forking. Its a shame, IMO.

I think we just have to wait and see what happens in the next few years:
- Linux continues to fracture or tends to standardize? (like LSB (http://www.linuxbase.org/) project)
- Linux market share climbs or remains stuck in the 1% region?
- Both fracturing and market growth? via regional or specialty versions?
- Developers are satisfied with 1%, or start to question what they are doing?
(for many, market share is not important)
- Chaos is the right formula and the cream will float to the top?
- Some distro grows into the 600 lb. gorilla and effectively sets the standards?
- Gnu or Ubuntu becomes the basis for other (specialized) distros?
- Dell or HP start selling a pre-loaded distro, and this sets the standards?
(similar to the IBM PC around 1980)

[prizrak]

hundreds of distros, perhaps 20 that have significant mindshare
SuSE, Ubuntu, Debian, RedHat, Mandriva. These are the big desktop ones. Not even close to 20. In fact that's less than Vista versions.
* they are mutually incompatible: you cannot run the same binary apps (the distro publisher does the work of tayloring the most popular apps)
Weird, someone forgot to tell my RealPlayer that.... Also before Ubuntu incorporated KeepassX into the repos I had no problem with an RPM of it. Acerhk module also seems to work just fine even though it's not Ubuntu specific. In short as long as you have dependancies satisfied just about anything will work, the only exception is badly designed apps that depend on a specific directory structure.
the apps are 90% overlapping in functionality (a huge waste of talent and money)
Winamp, Windows Media Player, Media Player Classic, iTunes, BSPlayer, VLC they are all media players for Windows and all do the same ****. MS Office, WordPerfect, OpenOffice, StarOffice they are all office software for Windows all do the same ****. Should I go on?
the average user cannot deal with this, and has no desire to do so
Vista Home Basic, Vista Home Premium, Vista Business, Vista Enterprise, Vista Ultimate, XP Home, XP Professional, XP Media Center Edition, XP Tablet Edition, Server 2003 Web, Server 2003 Standard, Server 2003 Enterprise, Server 2003 Datacenter, Server 2003 Small Business. These are all current MS OS products that can be installed by the OEM. Isn't that confusing? However will the poor [stupid] user choose among all of them?
* potential suppliers like Dell and HP do not have a supportable platform
SuSE, RedHat, Ubuntu, Mandriva - all have a dedicated company that is willing to provide support. All go to great lengths to make the user experience as painless as possible. Dell simply has to choose one of them, doesn't matter which one they choose. Simple truth is that for any newbie w/e Dell includes will work just fine. A seasoned user will simply install his/her own distro of choice with a 100% assurance that it will work. Basically, doesn't matter what they install.
* the technical powers that do the work like things this way, because they value freedom over standardization, or because they have more fun inventing new wheels rather than make someone else's wheel roll smoother.
More like people have different interests. Linux is very standardized, all software uses some open standard (such as ODF), LSB initiative is well under way. Source code is 100% portable between distros.

There is a good reason to not have a single distro. Here is a simple scenario. I have a desktop at work that is only meant to do work. Since it's Windows (not really my choice what gets loaded on them) I have WMP on it, MSN messenger, games, and all the retarded XP effects on by default. This being a work desktop I have no use for any of them. I could (or my IT dept could) customize the install image to not have any of these (well they still get installed just not turned on). It also has the auto update feature that really shouldn't be on a work desktop where the IT dept has to test patches before they go anywhere. It would be pretty nice if there was a stripped down workstation version of an OS that is customized towards general business use and is lean enough to be quick. As you can imagine companies don't change their desktops every 2 years so they are fairly slow.

This is why there are so many different distros. Each does something better than others and if they are mostly duplicate then w/e distro becomes most popular and wins in the end will be the one that was actually CHOSEN by more people.

[beefcurry]

Wow, Prizak. You caught the wind out of my mouth. Face it, Linux is very ready for the Desktop. It may not be perfect, it may not have all the features that we want. But it has a hell lot that we could use. (but sadly, the Linux Kernal has yet to implement any type of Colour Management.)

[Brunellus]

Wow, Prizak. You caught the wind out of my mouth. Face it, Linux is very ready for the Desktop. It may not be perfect, it may not have all the features that we want. But it has a hell lot that we could use. (but sadly, the Linux Kernal has yet to implement any type of Colour Management.)
BEGIN OFFTOPIC

Color management is most certainly *NOT* the job of the kernel. The kernel itself does no graphics.

If you want color management and want someone to blame for its absence, blame Xorg, GIMP, and the various printer driver maintainers--for all the good that blame will do you.

Also, colors are notoriously tough to get a handle on. One extremely popular colorspace--Pantone--is proprietary, with all the headaches involved in implementing that.

[migla]

There is a universal distro, that covers everything: Debian . ;)

I use ubuntu, though, since it is a specialized spoon (not fork) of Debian, with some newer packages (gnome among others).

[bravemosquito]

It can be done - using GoboLinux directory structure and Slackware's packet system. I hate dependency problems, there's no distro that hasn't such... Only Slackware has smallest amount of dependency errors. Also he has and far better scripting system

[prizrak]

It can be done - using GoboLinux directory structure and Slackware's packet system. I hate dependency problems, there's no distro that hasn't such... Only Slackware has smallest amount of dependency errors. Also he has and far better scripting system

Doesn't Slack also have the least amount of users? :lolflag: Slack got dependancy problems like a mofo and quite a few others. Have a friend who used to run it for years before switching over to Ubuntu and now OS X (no clue which version).

The reason why you can't have a unified Linux distro is the same reason you can't have one vehicle that will do everything. A two seater sports coupe won't be able to hold much cargo and more than driver + passenger. It will however be very good at running on a race track.

A sedan on the other hand will not be quite as good at running fast (tho things like Evo's and STi's manage to quite well) but will fit more people. At the same time in case you want large quantities of cargo moved you need a truck.

If you want something that flies you have to have some sort of an aircraft, etc....

Basically the issue is not necessarily that there is too much duplication of resources is that no distro could ever be all-in-one. DSL is meant for different things than Ubuntu. Gentoo has a different target than any other desktop Linux.

You could possibly make a point saying that there is no need for RedHat, SuSE, Ubuntu. and Mandriva to be around since they all concentrate on the same thing. Still you would have a hard time with that since it's basically darwinism at its purest. Whichever wins is the "fittest" in some way.

[bravemosquito]

Призрак, on distrowatch Slackware is placed 10th. For OS maintained from just one person it's a good example :mrgreen: I'm speaking from my experience, anyway I'm firstly *BSD user then Linux

[prizrak]

Призрак, on distrowatch Slackware is placed 10th. For OS maintained from just one person it's a good example :mrgreen: I'm speaking from my experience, anyway I'm firstly *BSD user then Linux
That is quite impressive ;) IMO the more the merrier anyway it's more fun with lots of distros around ;)

[kornelix]

That's an impressive list of stuff from prizrak rebutting all my points. I have to admit this causes me some rethinking. Here is a quick response I will try to keep short.

What "significant mind share" means is arbitrary and not worth debating.

Compatibility: I have several Gnome based apps developed on Ubuntu. They run on some distros, do not run on some others, and run with defects on some others. I have traced the problems to library changes that break a function's calling sequence, directory changes or the placement of files in different directories, and bash command differences. Other problems include different packages in the repositories, making resolution of dependencies something only a dedicated geek would endure. I have found that binaries built on older distros will run OK on newer distros, but not the reverse. I don't understand why, since if the APIs had changed this would not work in either direction.

Linux, with 1% market share, can less afford the waste of duplicated programmer effort than Microsoft and its partners with hundreds of billions of dollars to spend competing in the market.

The argument about different cars does not apply. Would you buy a car that ran on rails instead of streets? An extreme example, but should trigger a thought or two about compatibility and standards.

All the flavors of Vista you mentioned are ABI compatible. Not a good Linux analogy.

Michael Dell is in a quandry about which Linux to pre-install and support. The answer so far is "none".

Having many standards is as good as having none. The power plug is standardized everywhere, and is everywhere different. Useless. LSB is a good idea but not all Linux distros support this and it does not cover everything. Perhaps over time it will get better.

I don't agree that different distros are needed for different needs. Layer the applications you need on top of the standard base, which is the way software has worked for 50 years, and esp. Windows.

Darwinism is often brought up, and obviously competition is good and helps everyone improve. Microsoft needs more of this. I ask if Linux darwinism is only changing which distro gets more of the 1% market share. Do standards and compatibility rule out competition? I don't think so.

[prizrak]

kornelix,
For the sake of space I won't quote you, you know what you said and other interested parties can read it ;)
I wouldn't buy a car on rails but there is a type of vehicle that does run on rails. In fact I take one to work everyday ;) There is an advantage to having something on rails vs something on the road and this is what I was trying to go for with my vehicle (not just car) analogy.

LSB is still in it's infancy that is very true but it's there and is something that AFAIK all major distros are a part of. It's goal is to provide a common base for testing and compatibility between distros independent of what other "niceties" that distro includes. Basically it is there to solve the compatibility issue.

Yes Vista has the same ABI's but XP and server 2003 do not. Yet they are still supported and sold by MS. Vista also takes enourmous amounts of space and power to run because of all the compatibility and "universality" (I think I made up a word) that it strives for. SuSE for one doesn't want to work with my computer at any "usable" speed. SLED couldn't even dl and burn properly. Ubuntu has no such issues and I'm running Feisty. There are other who can't use Ubuntu but SuSE works very well for them.

Layering applications on a common base is basically what LSB is all about. Even then there are still certain things that might require customized systems. For instance a server kernel might be devoid of any multimedia modules to save space and make it more efficient. Render farms and A/V machines will need different things at kernel level and at the amount of load on those every little gain in efficiency counts.

As far as standartization goes, FLOSS in general is alot more standardized than proprietary world. Protocols and file formats are open and anyone can use them, you can introduce improvements into them and they will trickle down to everyone. Also you need to define what compatibility you are talking about. If you mean binary compatibility then you will not likely see it on Linux. Source compatibility has been there since inception.
I ask if Linux darwinism is only changing which distro gets more of the 1% market share
The share is actually estimated to be at 3-5%. Well the incompatibilities you are talking about are brought about by certain changes that distros create in libraries, kernel, etc.... The changes that are beneficial make it to the main tree and eventually integrate into the rest of the distros. The changes that are either useless or making things worse than they used to be don't make it. This is what most people mean by darwinism. The base of all those things is the same it is up to software developers whether to compile their programs for the "official" versions of packages or to use distro specific versions.

I have found that binaries built on older distros will run OK on newer distros, but not the reverse. I don't understand why, since if the APIs had changed this would not work in either direction.
This has to do with backward and forward compatibility. Lets say a piece of software calls "dudewhereismycar" and "dudewhereisyourcar". Another piece of software calls "whereisyourcardude". First two functions have been introduced in badmovie1.0 and the last function didn't make it there till badmovie1.3.9. The application that calls the first two functions will work in 1.3.9 just as well as it did in 1.0 because the function is still there. The application that calls the third function (whereisyourcardude) will work with 1.3.9 and anything above it but not 1.0 because it won't have the necessary function. This isn't FLOSS or Linux phenomenon it's common throughout the industry.

Michael Dell doesn't want Linux on his computers. He never did and was never going to preinstall it. All was said, is that Dells were going to get SLED certified so that when you do get a Dell with intention of putting Linux on it you know that it will work with it. BTW Dell has been selling RHEL servers for quite a while. Somehow they managed to choose a distro for servers I'm sure they can choose one for desktops.

[kornelix]

This has to do with backward and forward compatibility... [/offtopic]

Unless I misunderstand you, this cannot be the problem. The source code was never changed, hence the functions called were never changed. Old binary runs on new system OK. Recompile on new system and it crashes if run on old system. To be specific, these were Ubuntu 6.06 and 6.10. The message to stderr was about missing C libraries. If a new library was used in 6.10 why does the old binary still work?

It should work in either direction as long as APIs were not changed (and no one would be foolish enough to change C library APIs that have been constant for decades). A Windows binary app will run on any Windows system (at least that is my experience). The technology is simple: dynamic linking is through a named transfer vector which is patched into application memory at run time, either when it is loaded, or when the first call to a kernel or library function is made. This should not depend on library versions. Do you know more?

[prizrak]

Unless I misunderstand you, this cannot be the problem. The source code was never changed, hence the functions called were never changed. Old binary runs on new system OK. Recompile on new system and it crashes if run on old system. To be specific, these were Ubuntu 6.06 and 6.10. The message to stderr was about missing C libraries. If a new library was used in 6.10 why does the old binary still work?

It should work in either direction as long as APIs were not changed (and no one would be foolish enough to change C library APIs that have been constant for decades). A Windows binary app will run on any Windows system (at least that is my experience). The technology is simple: dynamic linking is through a named transfer vector which is patched into application memory at run time, either when it is loaded, or when the first call to a kernel or library function is made. This should not depend on library versions. Do you know more?

Sorry I was the one who misunderstood you. I didn't realize you were talking about binaries rather than code optimized for something.

The problem is not the source code, the problem is the binary. When you compile against libc6 (pulling names out of my butt) the compiled binary will look for libc6 or newer to work with. If you compile against libc5 (again pulling it out of my butt) the compiled binary will look for libc5 or greater. So anything compiled against libc5 will run on libc6 but anything compiled on libc6 will not run on libc5 because it doesn't see the library. If you recompile it for libc5 it will work just as happily since as you said yourself the source hasn't changed.

It's true for Windows world as well. It is possible to run 9x binaries on XP (doesn't even work all the time) but you will not for the life of you be able to run an XP program on 9x. The only possible way to get around it is something like Java applications or something that is so simple that it doesn't need any kind of specialized access. To illustrate, 98/Me games will not run on even 2000 and if they do it's usually software accelerated. The game didn't change but the system did.

If you want to contrast it with Windows, most of the stuff for XP does not run natively on Vista. Not even Server 2003 software is 100% Vista compatible (despite Vista being based on it). MS spent a lot of time and money coding an emulation layer for Vista that will run XP programs. There is no way in hades that you will be able to run anything made for Vista on XP. Another issue I actually encountered was MS Office not working on XP SP2 despite it working on Gold/SP1.

Binary compatibility is very tricky and error prone. That is precisely why Linux [FLOSS] tends to emphasize source compatibility over binary compatibility. In my experience actually quite a few programs work across distros. I have installed rpm's with alien on my Ubuntu boxes before and had no problems (well maybe there would be an odd library that needed to be fetched from the repos). Also LSB compatible RPM's work very well with Dapper and on as far as I know.

[kornelix]

Prizrak - thanks for the technology lesson. It could work better than this. There is no fundamental reason why binary compatibility cannot parallel source compatibility. I guess dynamic linking was just not done this way 20 years ago and nothing has changed.

I found this today in the THG web site: (http://www.geardigest.com/2007/03/01/dell_to_sell_without_os/)
Dell has managed to dodge the issue of selling machines with Linux pre-installed by saying that nobody can agree on which particular flavor of Linux to ship the machine with. Ohh well, consumer choice kills consumer choice.

The same page asserts that Dell is planning to ship computers without any OS pre-installed, for those who use Linux or another copy of Windows they own.

[FernandoMilton]

I found this today in the THG web site: (http://www.geardigest.com/2007/03/01/dell_to_sell_without_os/)

The same page asserts that Dell is planning to ship computers without any OS pre-installed, for those who use Linux or another copy of Windows they own.

That's a lame excuse from Dell. For the linux old-timers, the point of having a Dell with Linux installed, regardless of distro, is the pressure on the hardware companies to make drivers for this one machine. If the hardware works with the kernel, the userland applications doesn't matter, all other distros would work automatically too. For the linux new users, well. It is not if they would complain about what distro is being sold the same way they don't complain about what Windows version is being sold. Sell 98 and they will buy 98, XP and they will buy XP, Vista and they will buy Vista. The same would be true for the distros.

[prizrak]

Prizrak - thanks for the technology lesson. It could work better than this. There is no fundamental reason why binary compatibility cannot parallel source compatibility. I guess dynamic linking was just not done this way 20 years ago and nothing has changed.

I found this today in the THG web site: (http://www.geardigest.com/2007/03/01/dell_to_sell_without_os/)

The same page asserts that Dell is planning to ship computers without any OS pre-installed, for those who use Linux or another copy of Windows they own.
No probs.

I'm not too sure if you could dynamically link to libraries in the way you are talking about. I think Torvalds had something to say about that (in his usual manner I'm sure). If I remember correctly there are actual technological issues with binary compatibility that LSB is trying to resolve. Part of LSB spec is actually a universal packaging system based on RPM (why not deb I dunno) but far as I can see it never took off. I think CnR is the closest to what you are talking about and it is getting ported to Ubuntu and perhaps if it works there to other distros.

I'm also fairly certain that the compiler is the one taking care of the actual linking so dynamic linking would require a major redesign to all the compilers in use in Linux. If ISV's on the other hand do dynamic linking then the problem also goes away on its own since the only problem with alien'azing RPMs is dependency handling. If the software is designed in a way where it will use any library that has required functions then it won't really matter

Dell has sold naked computers for years, they are all business machines and all more expensive than "normal" machines.

They used the same excuse before, it was lame then it is lame now. They have been repeatedly told that it makes no difference whatsoever what distro they go with. If they provide their drivers (even in binary form) they will be incorporated into all major desktop distros (even if you have to manually install them from the repos/CD). I can't really blame them for that, Linux is just too much of a risk for them to take. The margins are razor thin on hardware as it is and there is no guarantee they'll be able to sell em.

The issue needs to be separated into technological part and other parts. From a technological point of view we already have UPM's. There is .deb (apt) and .rpm and CnR (tho that's basically a .deb/apt). All those were meant to standardize software installation in Linux. However RPM crowd didn't find DEB to do what they wanted and vice versa. The issue for the most part is that a UPM is not as critical as it may appear. FLOSS is 100% source compatible meaning that it can always be compiled on just about anything and for the most part desktop distros provide all the packages necessary (so far the only thing I need that I didn't see in Feisty repos is Real/HelixPlayer everything else is there). In the closed ISV world, just about any application for Windows/Mac is provided as fully self contained, meaning it contains all the necessary files within itself and won't be accessing system libraries. No one is stopping ISV's from doing the same thing for Linux. To add to it there are things like Java and Mono, which only require you to have a certain version of Java or Mono framework installed completely eliminating the need to care about libc and such.

Perhaps this is your answer as to why there is no UPM, software development is moving in the direction of what can pretty much be called virtualization thus eliminating the need for packaging.

[Adamant1988]

No probs.

I'm not too sure if you could dynamically link to libraries in the way you are talking about. I think Torvalds had something to say about that (in his usual manner I'm sure). If I remember correctly there are actual technological issues with binary compatibility that LSB is trying to resolve. Part of LSB spec is actually a universal packaging system based on RPM (why not deb I dunno) but far as I can see it never took off. I think CnR is the closest to what you are talking about and it is getting ported to Ubuntu and perhaps if it works there to other distros.

I'm also fairly certain that the compiler is the one taking care of the actual linking so dynamic linking would require a major redesign to all the compilers in use in Linux. If ISV's on the other hand do dynamic linking then the problem also goes away on its own since the only problem with alien'azing RPMs is dependency handling. If the software is designed in a way where it will use any library that has required functions then it won't really matter

Dell has sold naked computers for years, they are all business machines and all more expensive than "normal" machines.

They used the same excuse before, it was lame then it is lame now. They have been repeatedly told that it makes no difference whatsoever what distro they go with. If they provide their drivers (even in binary form) they will be incorporated into all major desktop distros (even if you have to manually install them from the repos/CD). I can't really blame them for that, Linux is just too much of a risk for them to take. The margins are razor thin on hardware as it is and there is no guarantee they'll be able to sell em.

The issue needs to be separated into technological part and other parts. From a technological point of view we already have UPM's. There is .deb (apt) and .rpm and CnR (tho that's basically a .deb/apt). All those were meant to standardize software installation in Linux. However RPM crowd didn't find DEB to do what they wanted and vice versa. The issue for the most part is that a UPM is not as critical as it may appear. FLOSS is 100% source compatible meaning that it can always be compiled on just about anything and for the most part desktop distros provide all the packages necessary (so far the only thing I need that I didn't see in Feisty repos is Real/HelixPlayer everything else is there). In the closed ISV world, just about any application for Windows/Mac is provided as fully self contained, meaning it contains all the necessary files within itself and won't be accessing system libraries. No one is stopping ISV's from doing the same thing for Linux. To add to it there are things like Java and Mono, which only require you to have a certain version of Java or Mono framework installed completely eliminating the need to care about libc and such.

Perhaps this is your answer as to why there is no UPM, software development is moving in the direction of what can pretty much be called virtualization thus eliminating the need for packaging.

Don't Rpath's "appliances" basically virtualize themselves?

[prizrak]

Don't Rpath's "appliances" basically virtualize themselves?

Never herd of them so I will go with an "I don't know"

[Adamant1988]

Never herd of them so I will go with an "I don't know"

rPath.com (http://www.rpath.com)

It seems that almost any application can be made into an "appliance" (including games) that run with a very stripped down version of Linux designed to do nothing but run that appliance, supposedly completely cross platform. "one package to rule them all"

[prizrak]

rPath.com (http://www.rpath.com)

It seems that almost any application can be made into an "appliance" (including games) that run with a very stripped down version of Linux designed to do nothing but run that appliance, supposedly completely cross platform. "one package to rule them all"

Interesting idea, would potentially destroy the OS market as we know since all you would need is some sort of a HAL running to run just about anything.

[Adamant1988]

Interesting idea, would potentially destroy the OS market as we know since all you would need is some sort of a HAL running to run just about anything.

Mmmmhmm.. It turns out that Foresight Linux is actually an "appliance" itself. I'm not sure if this would destroy the OS market or warp it considerably, but the main builder isn't completely open, so I don't think that we have to worry about that..

[jeffc313]

we dont need one unified linux, we just need open standards for everything. that way the real necessacary component, compatibility will be there.

[mediax]

Even if a unified distro was produced, it would be forked again in no time flat.

How many times have you heard that LInux is about choice? Multiple distros is a logical expression of that.

[jonathan21]

does anyone feel there are too many distro's out there.choice is always good but doesn't too much variety make people lost as to what distro to use.I was lucky enough to be introduced in to ubuntu through a friend.but would have fainted if I had been not been intoduced as there alot of disro's out there.am I being to harsh

[Rui Pais]

There are too many people too!

no serious, you don't have to install them all ;)

[3rdalbum]

If someone goes into an electrical store with the purpose of buying a TV, they don't just see that there are a dozen TVs there with different brand names and say "There are too many TVs. This is too hard. I'm going back to reading the newspaper".

No, they ask a salesperson for some information. The salesperson asks questions to discover whether they need LCD, Plasma, CRT; integrated or non-integrated; SD, HD or full HD; Samsung panel, Hitachi/Fujitsu panel, or Toshiba panel. At the end, the customer leaves with a TV. OR, the person does some research on their own and sees which one might be most suitable for their needs.

I don't see distributions as being any different. There may be hundreds of distributions, but new users will usually be steered or steer themselves to one of the top 20. These average joe users have already chosen one antivirus program out of dozens, and one computer out of 300, so why not a distribution out of the top 40?

[ThinkBuntu]

My main criticism is that many distros are repeating each other's work, and only a few are doing real original development while many others are just tweaking pre-existing distros. If there were some mergers (i.e. Mint helping with Ubuntu's UI and design alone) Linux as a desktop OS would probably be better developed than Mac OS X, or at least very close amongst the major distributions. I have absolutely no problem with a wide range of choices, however.

[LaRoza]

It is a different system, Cathedral v. Bazaar.

Remember, Linus Torvalds liked the Unix based system he was using, he just wanted it modified.

[jclmusic]

If someone goes into an electrical store with the purpose of buying a TV, they don't just see that there are a dozen TVs there with different brand names and say "There are too many TVs. This is too hard. I'm going back to reading the newspaper".

No, they ask a salesperson for some information. The salesperson asks questions to discover whether they need LCD, Plasma, CRT; integrated or non-integrated; SD, HD or full HD; Samsung panel, Hitachi/Fujitsu panel, or Toshiba panel. At the end, the customer leaves with a TV. OR, the person does some research on their own and sees which one might be most suitable for their needs.

I don't see distributions as being any different. There may be hundreds of distributions, but new users will usually be steered or steer themselves to one of the top 20. These average joe users have already chosen one antivirus program out of dozens, and one computer out of 300, so why not a distribution out of the top 40?

i agree, and the only reason this is a problem is because big buisiness would rather have us all think the only choices were microsoft or apple, and have pushed people into beleiving as such.

[juxtaposed]

There is never too much choice.

Though if you prefer less choice, restrict yourself to the major distros :P

[zenkaon]

For me this is a question of evolution and natural selection, I view this with a Darwin-type attitude. While many different species (distros) pop up all the time, only those that are strong enough and that adapt to their environment will ultimately survive.

Who will be left in 5 years time? Gentoo? Linsprie? maybe, maybe not. Who remembers OS/2 computers, or SEGA? The innovators and those with market share like red hat and suse are adapting to new demands and are prepared to evolve themselves and are thus likely to be around for a very very long time.

Then you have these new distros popping up like ubuntu, which is aimed at end users (as well as running servers, LAMP's, etc) and it seems like a boiling pot of ideas and innovation. ubutnu seems to grasp this evolve-adapt or die concept and I think that it could be here for the long run. In this world of natural selection, whose to stop a new disto coming along in 2008 and blowing everyone out of the water. Nobody, and thats a great thing.

I know that if you stick Linux into google you get pages of distributions and that may overwhelm people, but Linux is an evolving beast governed by the laws of natural selection. What you will end up with is a set of the most innovative, best os's out there, who will continue to compete and evolve for ever.

Take this world view and compare it to the closed, in-wards looking (inbred) worlds at M$ HQ and macOS.

[hobieone]

sometime to me it seems overwelming. but then again all the choices and competion between the different distros is what is and has been driving llinux to improve it self at a good pace and some distros such as dsl does serve a purpose like have a distro that can run from a thumb drive :). but over all i think it healthy for linix to have all the different distros so linux itself wont get complacent and will always be moving forward at a good pace

[aysiu]

No, there aren't too many distros. Read this for more details:
http://www.psychocats.net/essays/unifiedlinux

[Extreme Coder]

If someone goes into an electrical store with the purpose of buying a TV, they don't just see that there are a dozen TVs there with different brand names and say "There are too many TVs. This is too hard. I'm going back to reading the newspaper".

No, they ask a salesperson for some information. The salesperson asks questions to discover whether they need LCD, Plasma, CRT; integrated or non-integrated; SD, HD or full HD; Samsung panel, Hitachi/Fujitsu panel, or Toshiba panel. At the end, the customer leaves with a TV. OR, the person does some research on their own and sees which one might be most suitable for their needs.

I don't see distributions as being any different. There may be hundreds of distributions, but new users will usually be steered or steer themselves to one of the top 20. These average joe users have already chosen one antivirus program out of dozens, and one computer out of 300, so why not a distribution out of the top 40?

At least all TVs will have AV jacks ;)

[blackspyder]

I dont look at it as too many choices. I look at the original distros from which many have spawned. Ive tried 20+ distros and so far the only 2 that Ive installed on more than one system and upgraded have been Xubuntu and Fedora Core.

[inklein]

Deep down its all linux, and you could do the same thing with any distro, having a whole bunch just lets you chose a distro thats right for the job. I think having lots of distros makes my life easier, not harder!

[DJiNN]

does anyone feel there are too many distro's out there.choice is always good but doesn't too much variety make people lost as to what distro to use.I was lucky enough to be introduced in to ubuntu through a friend.but would have fainted if I had been not been intoduced as there alot of disro's out there.am I being to harsh

There can NEVER be too many Distros. :D Competition breeds innovation which brings new & glorious things for us, the Users, to play with!

Too many Distros...? Pah!!
:lolflag:

DJiNN

[STREETURCHINE]

ah the old linux windows comparisan test

. Lots of distros.
(a) On Linux: The lack of standardization is a huge stumble. Having too many choices is confusing for the end user. Even if it were narrowed down to just Ubuntu and Kubuntu, that's still too many.
(b) On Windows: The Vista editions run to Home Premium, Home Basic, Business, Enterprise, and Ultimate, or you can stay with XP or 2000, and even NT and Me and 98 SE are still out there. Giving a consumer their choice is what makes capitalism work! :lolflag:


no not to many choice is good,there must be one that{woks (oops )}works for you,that is the only one you need,

[3rdalbum]

At least all TVs will have AV jacks ;)

Ahh, but all distros have Bash ;) And the ordinary Jane doesn't want to fiddle around with AV jacks any more than she wants to type commands into the terminal.

[jrusso2]

does anyone feel there are too many distro's out there.choice is always good but doesn't too much variety make people lost as to what distro to use.I was lucky enough to be introduced in to ubuntu through a friend.but would have fainted if I had been not been intoduced as there alot of disro's out there.am I being to harsh

There are too many cars, trucks, refrigerators, styles of clothing, foods to eat, roads to take, things to do, how will I ever decide?

[lazyart]

The only way it becomes a problem is when a newbie can't figure out if they should install the .rpm, .deb or the tar.gz. Or is a particular folder in /usr /usr/bin or the like. Unified distro, no. Less standards? You're on to something.

[dca]

Hmmm, you can't bring Windows into comparison. MS is closed source, GNU/Linux is open source. That's like comparing apples and nipples.

If all the Linux distro(s) applied the same business model of Novell, RH, and a few others and providing subscriptions and tech support/services for something that's basically free just to prove that there is a market (based on what Wall Street determines as a market) for it than I would say there are too many distros...

[kazuya]

No such thing yet. I believe there is not enough yet. As mentioned earlier, all the distros are linux. Using many types expands a user's knowledge and makes one more adaptable to changes and increased innovations. If we all simply used a few or one type of distro or OS, then innovation is dictated upon by standards set by some higher authorities who may or may not promote better innovation. Limiting the number of distro leads to inhibiting innovation and placing the control to a select few.

I am in support of choices and more choices. If one uses all those distros long enough and closely, you start to notice their uniqueness. But if you as a user are just a one track minded user and are limited in your imagination as it comes to desktop or server feel, then you may use either one distro like Ubuntu or Pclinux, or FreeBSD or the commercial OSX, Windows, etc.

The popular types or options of distro comes in five main forms for me:
debian-based
slack-based
arch-based
gentoo-based
source-based

Going to distrowatch - You get quick indication of which distros to try.

[FuturePilot]

Hmmm, if you ask me, I'd say there isn't enough.

[ericfoy]

i agree, and the only reason this is a problem is because big buisiness would rather have us all think the only choices were microsoft or apple, and have pushed people into beleiving as such.

Huh?

[kano]

It's what makes Linux that great, there's tons of options for whatever your need may be :)
It wouldn't make much sense to install a large desktop distro on to a old, headless server when you could use a distro that is more geared towards that, would it? :p

[blackspyder]

I currently have 2 new distros to try. I go through Distro's very quickly. If a distro doesnt fit my needs or expectations (currently image manipulation and music recording/editing) within the first 48 hours I can it and find another. I like lightweight GUI's on my laptop and Full Featured GUI's on my Desktop. I have never had the issue of too many distros, its always a matter of which one to try next.

God I love Linux.

[RAV TUX]

Hmmm, if you ask me, I'd say there isn't enough.

No such thing yet. I believe there is not enough yet.



I agree with both of you, as far as Linux goes and the number of Linux distros to choose from we are in the dark ages, just imagine the number of awesome choices in 10 or 20 years from now....this is only the beginning.

[wiseleader]

Major distrbutions are more or less the same, they're using the same kernel (linux).

If you're desktop user, gui can guide you all the way.

If you're working in command console, some differernces are there, for example, yum in FC, apt in Debian and Ubuntu. Spent more time on the official web site, forum, local documentation can help.

I think we should stick on on distribution, then "explore" the others. Don't make a big jump to a new distribution with all of your working data.

For people who don't like too many distros, I recommend Freebsd, Openbsd, Netbsd, Solaris, only a few choices.

Different users are having different tastes, some like beautiful gui, some like stability and stability, some like fast update and upgrade. Personally, I like all of them, that's why I am here.

[deadguy87]

I think to people who don't know linux the vast number of distros might scare a few people off, but then again if they can find someone online or off for advice then it isn't problem. I was steered right to ubuntu from a linux user group on myspace.

[Adamant1988]

I prefer to think of the Linux distribution situation as a tree. The more branches and leaves feeding a tree, the stronger it is.

[mech7]

Major distrbutions are more or less the same, they're using the same kernel (linux).


That is the problem they are all the same and they all look alike and they all come with the same programs.. it would be easier if they would work togheter and make 1 really good OS instead of many 'ok' ones.

[rai4shu2]

I don't think people are scared off from buying pancakes because of the vast number of different breakfast-oriented restaurants. People have different tastes and services vary to match them.

[Eddie Wilson]

That is the problem they are all the same and they all look alike and they all come with the same programs.. it would be easier if they would work togheter and make 1 really good OS instead of many 'ok' ones.

Why would that even work. That only will work if everybody has the same taste. Other than that there are several very good distros out there. That didn't work for Microsoft so what makes you think it would work for a linux based system?
Eddie

[jrusso2]

I would love to see one of these too many choices people when they hit the cereal aisle at their grocer.

[aysiu]

In case some people are missing it, Linux distros do work together. They're all roughly the same because they all use the same programs developed by the same people. Ubuntu doesn't expend any more energy on AmaroK and GIMP than Fedora and OpenSuSE do.

Read more here:
http://www.psychocats.net/essays/unifiedlinux

In fact, as soon as I dig up the other "too many distros" thread, I'm going to "too many thread" them and unify it with this one.

[mech7]

* nevermind saw thread are merged :p

[prizrak]

I would love to see one of these too many choices people when they hit the cereal aisle at their grocer.

Man don't remind me, it takes me forever to pick a cereal I want just so damn difficult :(

For all of you "Too many" people.
Consider this, right now the only people who install/buy preloaded Linux are either organizations with IT departments or enthusiasts who are either trying out or already have some experience with Linux. It's not the "Which key is the ANY key?" crowd, it's people with certain expertise in computers. I don't think it's too unreasonable for them to do a little research and figure out what exactly they want out of the distro. Those who get Linux installed by friends/family because they feel that it's a better choice generally get whatever the person installing it feels to be the best distro for them.

When it comes to the "ANY key" crowd distro choice is largely irrelevant. Dell is going to be selling their systems with Ubuntu, HP might decide to offer SLED (other than Dell everything is purely for argument sake), Toshiba is gonna go with RedHat, Sony might prefer TurboLinux, and so on. So when a user decides to get a Linux system from one of the big vendors they will be using whatever the OEM decides is the best to use. The choice is basically done for the user.

[Praill]

This idea is contradictory to the ideology of an open-source kernel.
If you unify a kernel into one distribution then you have Windows. If you keep a kernel open then company A can devel their version on it, company B can develop their version, company C can devel a version for kids, etc, and competition can finally exist in the OS industry. Competition is great because it controls pricing, promotes ambition, and prevents a monopoly.

One thing I will agree with is that different distros should agree on standards for package management and other compatibility issues. That way 3rd party software could be easily installed on varying distributions.

[mech7]

Man don't remind me, it takes me forever to pick a cereal I want just so damn difficult :(

For all of you "Too many" people.
Consider this, right now the only people who install/buy preloaded Linux are either organizations with IT departments or enthusiasts who are either trying out or already have some experience with Linux. It's not the "Which key is the ANY key?" crowd, it's people with certain expertise in computers. I don't think it's too unreasonable for them to do a little research and figure out what exactly they want out of the distro. Those who get Linux installed by friends/family because they feel that it's a better choice generally get whatever the person installing it feels to be the best distro for them.


Yeah but this is the problem.. how to make linux desktop ready, and if only IT related people can make that choice it will never get there. Feisty's goals where to concure the desktop market i believe the only way that is possible when it is easy for people what they are getting.

[aysiu]

I was a bit overwhelmed by the choices when I first gave Linux a shot in 2004 (gave up pretty shortly afterwards), especially after visiting Linux.org (lame website, by the way), until I realized my limiting criteria gave me only one choice. My choice ended being Blag because I had a book from the library on Red Hat, but Red Hat costed money, and Fedora appeared to be too many CDs. I wanted a one-CD distro that was Red Hat-based, so I settled on Blag.

When I tried Linux again in 2005 (and ended up staying with it), I looked through DistroWatch to see what was the most popular. I tried Ubuntu and hated it (too much terminal needed and I also didn't know about MD5SUMs, and I had a bad CD that froze at 59% of installation). After doing a little more research on the top five at DistroWatch, I chose Mepis, which I also fortunately found a library book on. I used Mepis for a month, somehow stumbled upon http://www.ubuntuguide.org and http://www.ubuntuforums.org and got hooked on Ubuntu ever since.

Did I end up trying about fourteen distros between Mepis and Ubuntu? Sure. Just out of curiosity. Was I overwhelmed by choice? Not really. Yes, there are hundreds of distros out there. When you're a new user, though, are you really going to go for Pentoo, GoblinX, or Kororaa? Most distros that exist new users have never even heard of. Chances are, they'll start with Ubuntu, OpenSuSE, PCLinuxOS, Fedora, or Mepis--the top five at DistroWatch.

I honestly got more overwhelmed by choice when I used Windows--trying to find "free" applications without spyware or 30-day trials. Synaptic Package Manager totally won me over to Linux--all the applications I need all in one place, all easily installable.

Yeah but this is the problem.. how to make linux desktop ready, and if only IT related people can make that choice it will never get there. Feisty's goals where to concure the desktop market i believe the only way that is possible when it is easy for people what they are getting. prizrak is on the right track:

Businesses will either have a deal in place or will do their research... and there aren't that many obvious choices for the corporate Linux desktop (it's mainly RHEL or SLED right now).

Beginners will likely buy Ubuntu preinstalled from Dell (when it comes out) or have a Linux geek friend/relative install Linux for them... and the Linux geek friend/relative can pick a distro.

Windows power users just need to suck it up and do some research, or--God forbid--take two minutes out of their time to take the Linux Distribution Chooser quiz (http://www.zegeniestudios.net/ldc/).

I wish I had a DistroWatch and LDC for cars, retirement plans, health care providers, and ISPs! Rotten Tomatoes has saved me a lot of time doing research on what movies to see. Choice is good as long as it's informed choice. And I'll probably say it again in two months when someone else has the "brilliant" and "original" idea to merge all the Linux distros together.

[Fundi]

Choice introduces competition, and competition helps things get better. If there was only one linux distro I don't think it would be as advanced as the current choice of linux distro's. Besides I wouldn't say that was more than 7 main one's.

[gashcr]

I think that some standards are needed instead of less distros... at least for basic stuff as package managing. Distros are OK because they fill everyone's specific needs, but it would be nice to create some guidelines, this would be great for development, even would make it easier for hardware companies, as they would be able to make ONE linux driver... not an RPM, SRC, DEB, etc, etc, which is one of the main reasons I think they don't do them right now

[aysiu]

Well, standards are on their way but haven't been fully implemented.

Google LSB and Portland Project.

[gashcr]

Good to hear that!!

[Luggy]

Forget about distros for now, think about applications.
Why do we need so many music players?

We have Rhythmbox, Exile, Listen, Amarok, xmms, VLC and MPD ( if you want to count that as a player) just to name a few.

The idea that if these groups could get together and create a single project sounds like a good idea, but what happens when the Amarok guys start bashing heads with the Xmms guys over the style of player, one with tons of features and another with less and a simpler interface. Or further more, which libraries to use ( Qt vs Gtk ).

You can never please everyone when it comes to which apps to use so it's pretty obvious to see that you could never please everyone on which distro they want.

But that's the best thing about Linux and OpenSource, if you don't like something you can change it or write your own program to do it.

[aysiu]

It's not as if Windows or Mac has only one music player, too.

The great thing about Ubuntu is that it picks one application per task by default. So new users have something to use. If they later want to change up Rhythmbox for Exaile, they have that choice... later. But initially they won't be confused.

[prizrak]

It's not as if Windows or Mac has only one music player, too.

The great thing about Ubuntu is that it picks one application per task by default. So new users have something to use. If they later want to change up Rhythmbox for Exaile, they have that choice... later. But initially they won't be confused.

Actually I think Luggy's point is that it wouldn't be beneficial to have one distro or one music player. Seems like the point is that if people can't even agree on a media player (that pretty much does only one thing) there is no way they would agree on something as complex as a distro.
Yeah but this is the problem.. how to make linux desktop ready, and if only IT related people can make that choice it will never get there. Feisty's goals where to concure the desktop market i believe the only way that is possible when it is easy for people what they are getting.
Linux is desktop ready, from browsing around the forums you can see that the main problem tends to be hardware. If you are getting an OEM install of Ubuntu you eliminate those issues. If you are talking about problems with choice then I basically refer you back to what aysiu said.

[kragen]

I'm half and half on this matter - on the one hand it really irritates me to see loads of separate projects duplicate each others work - there are loads and loads of music players, but for the most part they all do pretty much the same thing and just look different. It's not even like they have different goals.

In the same way it was fustrating to see development split between KDE and Gnome - the two desktop enviroments may have different goals, it seems to me that if the development was combined and worked on one project, things would go so much faster - after all both projects have panels, a file browser, a text editor etc... Why develop two slightly different applications when you could use the combined efforts to develop one application superior to both?

Now I see that in fact it's a little like evolution. You may have lots of variants on a theme - several different web browsers, media players etc... each with slightly different philosophies and methods on how to achieve a goal - (playing music, text editing), but there isn't a need for 15 different media players, so only the strong projects survive. It may sound better to put all the resources into one project, but that's assuming you choose the right project - if you choose the wrong project, you may well end up with something that's junk, no matter how well developed it is.

I see that as one of the beauties of open source programming - if someone sees a better way of working, then they are free to split off and create their own project if they wish, rather than needing to start from scratch, a bit like a mutation or natural variation in biological evolution.

I'm thinking that eventually things will sort themselves out and the weaker projects will be left behind - after all we only have one x-sever, and one kernel (for the most part).

[tenshi-no-shi]

I think that a unified linux would not be a good thing for the simple reason that competition helps with development and hinders it, as well as having different distributions or "flavors" of linux makes it appeal to a larger demographic of people. If there was only a "beginner" distro of linux like debian, etc then the more hands on sorts that like the challenge of setting up linux would probably avoid it. You could try to find an compromise, but that usually results in both parties unhappy with the results.

I like the idea of the many distros as well. It is nice to think that if I do not like the direction that this distribution is going I can find one that fits into my philosophy and lifestyle better.

[Auria]

Perhaps not a single Linux distro

But i really do think apps should be downlaodable for the author's webpage - packaging every single app in thousands of formats and non-standards rpms and debs that work only in one flavour of a distro, etc. is a real pain and besides that repositories are quickly outdated, and also time-consuming for the distro team, while the developers of each app could do it themselves if there was a unified way to do that - that does make installing apps more complicated.
There would be a need for common dependency naming across distros

[aysiu]

Perhaps not a single Linux distro

But i really do think apps should be downlaodable for the author's webpage - packaging every single app in thousands of formats and non-standards rpms and debs that work only in one flavour of a distro, etc. is a real pain and besides that repositories are quickly outdated, and also time-consuming for the distro team, while the developers of each app could do it themselves if there was a unified way to do that - that does make installing apps more complicated.
There would be a need for common dependency naming across distros
Well, the most practical way to solve this is to see how market forces evolve and put one distro on top as the major desktop distro.

Right now, it's looking good for Ubuntu. And if that's the case, any third-party application developers would be sure to package a .deb that's Ubuntu-compatible, and other distros would just have to fend for themselves.

[prizrak]

Perhaps not a single Linux distro

But i really do think apps should be downloadable for the author's webpage - packaging every single app in thousands of formats and non-standards rpms and debs that work only in one flavour of a distro, etc. is a real pain and besides that repositories are quickly outdated, and also time-consuming for the distro team, while the developers of each app could do it themselves if there was a unified way to do that - that does make installing apps more complicated.
There would be a need for common dependency naming across distros

Which in turn would basically require a single Linux distro. You could not run a Feisty .deb on Edgy or Dapper (well there are some exceptions, certain things don't depend on much). For the most part you couldn't run a RedHat package on Debian, the libraries and in fact even the kernel are different.

The reason why you have to make those different packages is because different distro's use different libraries and even in the same distro different versions will use different libraries. The package in some cases can contain all the necessary dependencies but in alot of cases it's dependencies will also have dependencies making it near impossible to package it all.

As Aysiu said, the only way to deal with the "problem" is for one or two distros to emerge as market leaders with everyone developing for them.

[forrestcupp]

The funny thing is that the same people who preach "freedom to choose" are the same ones who bash Windows Vista for having too many shut down options to choose from.

We need choices, but do we really need thousands of choices with really only 3 or 4 major variations on how things are set up?

Standardization efforts are good, but they will never be totally implemented because of too many stubborn people. It's true that the best scenario is for one or two distros to be clear leaders, and that is pretty much how it is. But for how long?

[mech7]

not only the hundreds of same distro is the problem also they the different parts don't work togheter is a problem. When there is a bug there is no central way to report it, you have to go to the individual project :(

[prizrak]

The funny thing is that the same people who preach "freedom to choose" are the same ones who bash Windows Vista for having too many shut down options to choose from.

We need choices, but do we really need thousands of choices with really only 3 or 4 major variations on how things are set up?

Standardization efforts are good, but they will never be totally implemented because of too many stubborn people. It's true that the best scenario is for one or two distros to be clear leaders, and that is pretty much how it is. But for how long?

It's not about being stubborn, it's about having something that works for you. There are only 300 distros or so btw not thousands. There will always be a clear leader (or 2) in anything. You have to realize that alot of distros are highly targeted. You get Gentoo that is 100% customizeable and is basically aimed at the nerds (not even geeks at this point), then you have Debian that tries to provide the most stable environment possible (great for servers). Ubuntu, RedHat, SLED and Mandriva are all fighting for desktop dominance (although the first 3 offer servers). There are things like Puppy and DSL that are aimed to basically be run off a USB stick. Then we have basically a troubleshooting environment that is Knoppix. That's not counting Linux that runs switches/routers/firewalls/vpns, then other Linux that runs on PDA's and smartphones and so on.

There is no one-size-fits-all solution, this is why Windows servers are absolute nightmare. There will always be 100's of distros but only a handful of them will ever be relevant. You will see leaders in desktops (most likely Ubuntu and SLED right now), and in servers (RedHat is popular, Debian is too), the rest will be highly specialized.

It's survival of the fittest, Mandriva used to be THE desktop Linux distro but now Ubuntu and SLED has overtaken them. Evolution requires alot of doubling of work, after all you won't argue that there shouldn't be a Tiger if a Wolf also runs on four legs ;)

[aysiu]

People always say "there are hundreds of distros," but most people never really try more than 3-15 of them. A lot of people will just look at the top ten and try to whittle it down from there.

P.S. A lot of the hundreds of distros have nothing to do with regular desktop users.

I got more confused about picking a restaurant to celebrate my birthday at than picking a Linux distro to start with.

[23meg]

People always say "there are hundreds of distros," but most people never really try more than 3-15 of them.

And they don't even hear about more than 20-30 of them, ever. It's little beyond a tautology: "There are hundreds of Linux distros, so there's a lot to choose from". Well, do you know all the thousands of species of flowers that grow in far continents? Do you have to? Not unless you're a botanist. You know a couple dozen of them, and pick one out of the few dozens available at the florist.

Same with Linux distros: there aren't 59434 separate distros all aiming at you, that you have to know about. At most, you'll have to read about twenty, try five, choose one, forget the rest.

[prizrak]

And they don't even hear about more than 20-30 of them, ever. It's little beyond a tautology: "There are hundreds of Linux distros, so there's a lot to choose from". Well, do you know all the thousands of species of flowers that grow in far continents? Do you have to? Not unless you're a botanist. You know a couple dozen of them, and pick one out of the few dozens available at the florist.

Same with Linux distros: there aren't 59434 separate distros all aiming at you, that you have to know about. At most, you'll have to read about twenty, try five, choose one, forget the rest.

I don't even know 20 distros of the top of my head and I have started with RedHat 6.2 (I think that's what it was it was around 2000). Just for fun here is a list of distros that I can remember right now.
1 - Damn Small Linux
2 - Puppy Linux
3 - Yellow Dog Linux
4 - RedHat/Fedora (same crap really)
5 - CentOS - didn't know it existed until trying it at work for a project a few months ago
6 - 8 - Blag, Xandros, PC Linux OS - heard of them but never seen.
9 - Knoppix
10 - Debian
11 - Ubuntu
12 - Slackware
13 - Gentoo
14 - ASP Linux - Russian RH clone
15 - ALT Linux - Russian Mandrake/iva clone
16 - *spire
17 - SuSE
18 - Arch
19 - Sabayon

Umm I actually think that's all, was going to put Korroraa in there but it's pretty much dead. Now out of all of those 6 - 8 I heard about from Aysiu. The Russian distro's have a pretty specific market non-Russians are unlikely to know them. Knoppix, DSL, Puppy, Yellow Dog - specialty distro's that desktop users don't need to be concerned with. Slack, Debian, Gentoo, Arch, Sabayon - "Geek" distros, that a normal user should not be concerned with. This leaves only 5 distros to choose from, that's a smaller number than Vista editions. Out of those there are many distros that are basically clones of each other so really if you try one you tried all.

[forrestcupp]

I'm not against different distros as much as I am against differences in directory structures and packaging. Of course different people have different needs. All a distribution really does is package a bunch of software that targets a specific crowd; it's not really its own OS. But why can't the structure be standardized? It can, and people are trying to do it right now, but it won't fly because of stubbornness. Some people like yum and rpm's and they won't give it up, some people like debs and apt, and they won't budge. If there were one good way to package software, then people would have their choice between compiling from source or just installing binaries. Differences are great, but the foundational things should be standard.

People have argued that you can't have one package for different distros because different distros use different versions of dependencies. So my question is, why in the world aren't the dependencies made to be backward compatible? It shouldn't be that hard to upgrade versions of things when necessary.

At least using the most popular distro increases my chances of software being packaged for me.

[prizrak]

People have argued that you can't have one package for different distros because different distros use different versions of dependencies. So my question is, why in the world aren't the dependencies made to be backward compatible? It shouldn't be that hard to upgrade versions of things when necessary.

First I'd like to point out that backward compatibility is pretty good in Linux land but over fairly short periods of time (year or so).
Having said that, there are some problems with backward compatibility idea. It introduces bloat - functions are rewritten, redesigned, reimplemented all the time to make them more efficient, stable, fix bugs and so on. In order to provide backward compatibility it would be necessary to basically keep the old function with the new. This would create huge libraries and a fairly complex way of linking to them as software would be looking for different functions in the same library. It would make the libraries more complex, which means they would be more prone to bugs.

Another problem is inability to introduce new techniques. One of the biggest reasons for updating the libraries as often as it is done in FLOSS world is because there are new techniques/technologies arising all the time. This is something that's been holding Windows back. The latest infamous problem that spanned accross every single Windows version was handling of animated cursors. It's not because MS couldn't rewrite that library but because it had to keep a stable API for all the 3rd party software that expected certain behavior.

It of course introduces slowness, as applications search through libraries (That are huge) for their functions they have to slow down. It might not mean much with a single app but a bunch of different ones, especially when they are using older and slower functions will really bog a system down.

Backward compatibility has both good and bad sides but for the most part things are backwards compatible unless we are dealing with a major version change. On the other hand the problem is forward compatibility. You could not run a Feisty deb in Edgy because it is looking for newer versions of libraries. Same problem across distro's. Though you can add to that that distro's actually may use different kernels and different branches of the same library.

As far as filestructure goes it really doesn't matter. A package manager should be the one handling where the files go, of course that would require adding some handling code to the manager and the package itself but hardly as difficult as getting all distros to conform to one file system.

[aysiu]

More BS from our friend Alexander Wolfe:
Too Many Linux Distros Makes For Open-Source Mess (http://www.informationweek.com/blog/main/archives/2007/07/too_many_linux.html)

This is the man who brought us such wonderful gems as Ubuntu Linux's Achilles' Heel: It's Tough To Install On Laptops and Dell's Ubuntu Offering Shows Why Linux Is Its Own Worst Enemy

I generally scoff at the idea of bloggers existing as paid employees of Microsoft just to spread anti-Linux FUD, but Alexander Wolfe might change my mind about that...

[23meg]

If you'd rather not click the link and give Information Week and Wolfe another hit with an ubuntuforums.org referrer, the full text is below:

Too Many Linux Distros Makes For Open-Source Mess
Posted by Alexander Wolfe, Jul 18, 2007 11:09 AM

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.

Precisely how many distros there are is probably unknown, since anyone with a some spare time on their hands can create one. (Check out Linux from scratch, which "provides you with step-by-step instructions for building your own customized Linux system entirely from source.")

Let's agree that the authoritative figure is 359, since that's the number of distros on the list maintained by DistroWatch.com.

True, some distros are more distributed than others. Ubuntu, which is clearly the flavor of the month (who says publicity doesn't matter when it comes to Linux?) is out front of everything else, according to DistroWatch. OpenSUSE, PCLinuxOS, Fedora and MEPIS round out the top five. (Since this is an enthusiast site, one must assume that Novell and Red Hat are way unrepresented, so one should add those guys into the top tier.)

The existence of some number of multiple versions of Linux makes sense, on the grounds that there are different kinds of users who need different distros? But 359?

DistroWatch's stats page has as apt an explanation for this phenomenon as you find anywhere:

"A Linux distribution is like a religion. If you've ever tried to suggest to another person that his or her choice of a distro might not be the best, then you know what I mean." [The "I" is probably site founder Ladislav Bodnar.]

Ah, so Linux is like a religion. I seem to have heard that one before. 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

Here's the way that meme was framed in a 1999 post, entitled "Fear of Forking":

Linux won't fork because the fork-er has to do too much work for no payoff: Any worthwhile improvements he makes will be absorbed into the main branch, and his fork will be discarded/ignored as pointless.

One defense-- I'm sure I'll receive comments in this regard -- is that this comment is referring to forking of the kernel rather than a multiplicity of distributions. It is indeed true that the kernel hasn't forked in any significant way, thanks to Linus Torvalds' control.

It's also true that the few deviations which can fairly be called forks are very valuable, in that they are patches or shell add-ons (the latter are not really forks) to support real time and load balancing. (Real-time Linux merits is a worthy endeavor, meriting a whole, separate discussion.)

So I'll grant readers that, if there's anything amiss with my argument, it's that I've dragged the "f" word into the discussion. I should've just said that there are way too many distros, and left it at that. But then I wouldn't have been able to close with the thing I can't help but think, no matter how many times open-source supporters tell me that what they're offering is so much better than the OSes pedaled by Microsoft. It's this:

There's no other way to put it: Linux is a forking mess.

[saulgoode]

Bad puns only work if they arise naturally out of a conversion. They fail completely if you have to manipulate the discussion in order to inject them.

[stinger30au]

Does Linux really need another distro?
If the "linux camp" wants to get serious about knocking Microsoft off its pedestal of its desktop dominance, then all distros need to start working together rather then keep breaking up in to smaller fragments. I honestly believe Linux can knock M/S off its throne, but untill the "Linux Camp" stand united, your pushing asking for the impossible.

I believe the "Linux Camp" can do it and win.

[aysiu]

Does Linux really need another distro?
If the "linux camp" wants to get serious about knocking Microsoft off its pedestal of its desktop dominance, then all distros need to start working together rather then keep breaking up in to smaller fragments. I honestly believe Linux can knock M/S off its throne, but untill the "Linux Camp" stand united, your pushing asking for the impossible.

I believe the "Linux Camp" can do it and win.
The Linux distros are working together and fragmented. That's how open source works.
Linux Doesn’t Need a Unified Distro (http://ubuntucat.wordpress.com/2005/09/15/linux-doesnt-need-a-unified-distro/)

[stinger30au]

Linux Doesn’t Need a Unified Distro (http://ubuntucat.wordpress.com/2005/09/15/linux-doesnt-need-a-unified-distro/)

i dont think Linux needs a unified distro at all. but i think there needs to be less fragments and more support for current distros, there is enough of them already. im amazed at how many people look at a distro say "wow thst great but it needs x,y,z to be better so i will make a entire new distro rather then add to whats already there"

[Nunu]

Can it be done?

[smoker]

The Linux distros are working together and fragmented. That's how open source works.
Linux Doesn’t Need a Unified Distro (http://ubuntucat.wordpress.com/2005/09/15/linux-doesnt-need-a-unified-distro/)

excellent article, well written and frank, and i completely agree :)

[loell]

i dont think Linux needs a unified distro at all. but i think there needs to be less fragments and more support for current distros, there is enough of them already. im amazed at how many people look at a distro say "wow thst great but it needs x,y,z to be better so i will make a entire new distro rather then add to whats already there"

:popcorn: you have some point in there, however, you can't stop those distro developer wannabes , using a reconstructor , add and minus some default programs/ packages then name the Distro with their name initials and then release it to the public :lolflag:

[init1]

There are different distros for different people. Gentoo is for experts and Ubuntu is not. If they were combined, either one or both of the groups would complain. I'm not saying experts can't use Ubuntu, but it isn't optimized for experts like Gentoo is.
That's why unifying Linux would be bad (and can't be done).

[handy]

Great minds think differently!

Which works wonderfully with Open Source & Free Software.

[RebounD11]

It would be great to have a distro that is compatible with any other distro (like one where U could install RPMs and DEBs from any other distro without compatibility problems) and a distro that takes all the good stuff from all other distros and work (stable) on any computer - the only problem is that such a distro is:
1. Improbable/impossible - but such a project might evolve into sth very nice
2. very large and a lot of stuff will be unused (reason I quit Sabayon Linux which is probably the most complete distro of the moment) and some people use their OS for one/few certain job(s) and everything else just takes up space on disk and time to get rid of it.

In my opinion it would be a great initiative - and it would add another choice to the beautiful diversity that is linux :D

[@trophy]

How can having 50 , 100 , 1000 distros harm Linux ? ...The old saying Many hands make light work

I think what he was getting at is that when you have people duplicating efforts, it divides your resources, and each team gets less done than they could have done working together. They do this because the tradeoff is that they get to make their project do *EXACTLY* what they want it to.

Sometimes, though, this tradeoff isn't worth it. For example, if what I want to do with it is "linux for desktop computers, and it should have Frozen Bubble installed by default" then it probably wouldn't be worth starting my own distro. Consequently, I use ubuntu, and the first thing I do every install is grab Frozen Bubble!

For that, and other reasons, I think there's probably a number of linux distros which would be optimal in terms of using our resources wisely. It's much larger than 1, but much smaller than the 1000s currently out there.

[aysiu]

I think what he was getting at is that when you have people duplicating efforts, it divides your resources, and each team gets less done than they could have done working together. They do this because the tradeoff is that they get to make their project do *EXACTLY* what they want it to.

Sometimes, though, this tradeoff isn't worth it. For example, if what I want to do with it is "linux for desktop computers, and it should have Frozen Bubble installed by default" then it probably wouldn't be worth starting my own distro. Consequently, I use ubuntu, and the first thing I do every install is grab Frozen Bubble!

For that, and other reasons, I think there's probably a number of linux distros which would be optimal in terms of using our resources wisely. It's much larger than 1, but much smaller than the 1000s currently out there.
Since Linux is open source, and most of the programs made for it are also open source, there is very little duplication of effort.

To take your example, you could create a Linux with Frozen Bubble installed by default, and it might be worth it. Just install and use Reconstructor, and you can modify the default set of programs on the Ubuntu .ISO and create your own Ubuntu called Frozen Bubblebuntu. The whole process might take you an hour.

People do it all the time. Ubuntu Christian Edition. Ubuntu Satanic Edition. Ubuntu Ultimate Edition. Ubuntu Studio. Linux Mint.

Reconstructor:
http://reconstructor.aperantis.com/

Ubuntu and other distro developers don't create their own Gnomes from scratch, duplicating the work that Gnome developers put in. There's one Gnome, and all the Linux distros that use Gnome use the same Gnome (maybe they change the wallpaper or default set of applets). Likewise for just about any included program. Same AmaroK. Same Audio TagTool. Same Audacity. Same Skype. Even when Debian creates its own version of Firefox (IceWeasel), it isn't duplicating efforts. It did not build IceWeasel from the ground up. Debian developers just took Firefox and removed all the trademarked material (Firefox logo, etc.).

It's closed source developers that are duplicating efforts. In open source, you can take what's out there, make one tweak and call it your own. In closed source, you have to start your program from scratch always... or find an open source program with a BSD license (as Apple did for Mac OS X).

[@trophy]

Since Linux is open source, and most of the programs made for it are also open source, there is very little duplication of effort.

I know it's the same GNOME and all that, but resources are still being divided. Community resources, for one: "I only use Ubuntu/Red Hat/Gentoo/etc" Financial is another: Any money spent on Red Hat branding doesn't produce a matching benefit on our end like money spent on Red Hat coding *sometimes* (though not always) does.

And another resource that gets split is user enthusiasm. I didn't like linux for a long time, because there was too many choices I had to make about everything. I've since realized that the benefits outweigh that negative, but I still believe that there's an optimum amount of choice that's good, and past that it's a bad thing.

It's like that episode of Seinfeld where the carpenter was going to redo Jerry's cabinets, but he never got anything done because he kept asking what kind of hinges, what kind of handles, what kind of screws, etc that Jerry wanted.

[aysiu]

I know it's the same GNOME and all that, but resources are still being divided. Community resources, for one: "I only use Ubuntu/Red Hat/Gentoo/etc" Financial is another: Any money spent on Red Hat branding doesn't produce a matching benefit on our end like money spent on Red Hat coding *sometimes* (though not always) does. Well, there will always be something wasted. If you didn't waste resources on separate branding, you'd waste resources trying to coordinate the efforts of people with different visions into one narrow goal and end product.

And another resource that gets split is user enthusiasm. I didn't like linux for a long time, because there was too many choices I had to make about everything. I've since realized that the benefits outweigh that negative, but I still believe that there's an optimum amount of choice that's good, and past that it's a bad thing. And I happen to think the choice is at its optimum. I can name probably a good five to ten distros that are aimed at the home desktop user. The rest aren't even relevant choices. When it comes down to the average Windows power user, Gentoo and Slackware are out of the picture. So are about a hundred distros most people can't name from memory (and need to consult DistroWatch to find the names of).

It's like that episode of Seinfeld where the carpenter was going to redo Jerry's cabinets, but he never got anything done because he kept asking what kind of hinges, what kind of handles, what kind of screws, etc that Jerry wanted. It's exactly like that. Instead of having one kitchen you have to keep asking questions about or use your best judgment on, you have as many kitchens as you want, and Jerry can pick the one he likes best without having to be consulted for every little thing.

[@trophy]

It's exactly like that. Instead of having one kitchen you have to keep asking questions about or use your best judgment on, you have as many kitchens as you want, and Jerry can pick the one he likes best without having to be consulted for every little thing.

Perhaps we should just agree to disagree then.

[Yfrwlf]

Hi,

besides Canonical being Ltd., there is one more thing that drives me crazy - why do we the hell have that many linux distros around?

Isn't is different folks developing parallely exactly the same things? This greatly decreases the productivity of the whole community. I mean sure a lot of distros use the same packets etc., but for example the PROMOTION, which linux really really need in order for OEM to ship new pc with LINUX INSIDE, instead of that Microsoft crap beeing in every god damn mobile device now (got a PocketPC recently - Win mobile 2003 - and I am about to spit fire and throw it against the wall if I wont get linux inside soon!!!

I would be pretty happy with one universal distro around, beeing nicely boosted with features for everyone... I think Ubuntu goes this way and thats what i love it for. Until someone in Canonical screws up...

Jan

This thread is huge, and I don't have time to read through to see if this was already mentioned, but I just had to comment.

What you always hear most everyone quoting as being a good reason for having so many distros is "choice". A major problem that that ignores though is the need for standardization. Linux could remain free but become so fragmented like Unix that you HAVE to have version X of program A to run version Y of program B. This fragmentation can turn Linux, a free OS, into a very *restricted* OS, when all the collaboration is removed. Again, while having everything "free" and open source means that you can fairly easily port things, I think there is a limitation there too, mainly time/effort/money being wasted.

This needs to be prevented in order to allow work to be focused and time not wasted. While there are some great examples of groups understanding the need for standards and pushing for them, I believe that Linux could and should try to push interoperability to the limit. Most everyone here probably understands the frustrations when you go to download a program that *isn't* in the repositories and they only offer RPMs, or you download the binary or the source and it fails because of a multitude of problems. Then you have Windows and Mac users who go and grab a single file and install it often without a flaw. I don't mean to focus on package standardization though, but in general I believe that all things need to be pulled in to follow *standards* of some kind or another. If developers thought ahead about what interfaces each "module" they were creating and how it could be standardized so that any other "module" could take it's place, Linux could be completely modularized, even allowing an easy swap of one kernel to another as long as it met the specifications for the standard?

My point is that Fedora has some nice X and Y features should not prevent Ubuntu from having them as well. Standardization is as important, IMHO, as having the software under a free license. If you can't USE the software, you are restricted from it, and free software has lost it's meaning.

What I think may be neat is if you combined all the servers from all the different repositories together and had one massive set of packages, and then you could have a tiny net installation procedure in which you would specify the *template* for the packages you want. If you want Ubuntu you select the Ubuntu template, for CentOS select this other template. If everything was inter operable then competition would allow the best packages to shine *regardless* of the distribution they happened to get bundled into.

I know this is a dream but I believe more developers should start forming ways of making things standardized and ways of working together so that no work is lost with everyone making their own custom distros which often have most features overlapping with one another. Or hell, even if you wanted to leave the KDE stuff on the KDE servers, and the Gnome stuff on theirs, and the Fedora stuff on theirs, etc, at least make an easy way to get *everything*, a universal repository system standard and make sure that all the basic things are standardized so when I did go tell it to get me a kernel and X and a desktop environment, it'd all cooperate and play along nicely. THAT is what is good about Linux, is when the software gives you freedom, AND it works! Sometimes I swear I'm close to crying fowl when certain big players in Linux insist on not playing with others, probably for monetary reasons. Canonical may be no different, but at least they haven't made their own version of deb packages that aren't compatible with regular ones or something stupid. >.<

</rant>

[ExSuSEusr]

i totally agree with jan, Linux would be overwhelmly better than windows by now,

Ububtu already IS. :)

For the average person - who may only be interested in surfing the web, checking email, creating a 'powerpoint' / other document, listening to music and watching video.... Linux - regardless of distro (within reason) already works 'out-of-the-box.'

I really like SuSE, but the SuSE community sucked. Sorry if I offended anyone, but people of the SuSE community weren't very inclined to help each other much. I know this after two years of use and spending a great deal of time on the boards. *One* of the main factors that brought me to Ubuntu was I saw just how much support the community offers. You ask a question here, and you'll general have an answer in a few moments. Ask a question on a SuSE board and you're lucky if you get an answer in a few days - if at all. Now I'm not saying SuSE is a bad distro, it's great. And, in fact, if you install it - you can pretty much do everything the average computer use does without really having to configure much at all.

However, there are those of us who use our systems for much more. I installed Ubuntu less than a week ago and I have everything I'll need (WoW, and other Windows 'only' programs) up and running. I like to think of myself as a bit more tech savvy than the average user, though. Therein lies my point. For the average user, SuSE, Ubuntu, Debian, Gentoo, all these distros work - and work well. For those who are needing more, chances are they can manipulate (whether such involves compiling their kernel or not) as they need.

It's a matter of choice - which exactly what Microsoft does NOT offer. If we as a community compiled all our distros into one - then we will have become what so many of us despise. An open-source version of the same crap we avoid. I would venture to say that Linux becoming "one" is exactly what Microsoft would want. What keeps MS worried (and yes, they're worried about Linux - despite what you hear and read) IS the fact there are so many different choices.

Standardization isn't necessary in Linux, it is after all, everything our community stands against.

Such is my opinion.

[Yfrwlf]

Standardization isn't necessary in Linux, it is after all, everything our community stands against.

Tell that to the Linux Foundation, freedesktop.org, OpenGL, and other groups who are trying to standardize/modularize Linux to allow software to be actually created for Linux and USED by us, the users. You can't make a game for Linux if there are 10 graphics API standards out there, that means 1/10th of users would have access to a program if they were all used evenly, and it means more work and wasted effort. Standardization is a good thing when it allows room to grow and can adapt to new ideas while not causing the standard to fail in their compatibility. Standards means a set of common languages in order to allow communication between modules. There is nothing wrong with this, and it in no way hinders the creation of new or alternate programs to replace the programs that use the standards. If the standard needs to be updated for whatever reason, this can be done. Everyone still has the freedom to create whatever the hell they want to, that will never be taken away, but all developers should seriously consider using standards when they program so that their software CAN BE USED. If they find the standards that exist to be too restricting, they should help update them or write their own, sure, but for god sake look at the big picture before you run out and start programming kernel L33tRandomness that doesn't work with anything else, maybe you should think about putting your effort in places that actually contribute and improve the existing systems instead of wasting your effort if there really isn't a reason to do so.

To help out and pull with the group takes a lot of maturity, I know, but if it's possible and you don't have a good reason to start your own faction then please, for the sake of all of us Linux users in the world, please don't. Help the community to get more software that is compatible with their systems, or help improve the standards if you feel they need updating. Linux needs to be a cohesive platform that anyone can learn about in school, study for, and then program for and help out, not some random, fractured mess that you have to learn on the fly that can't have any documentation for it because no one has decided on some basic standards to set for it.

[Chymera]

I voted for other so i think i have to state my personal opinion.
NO, it cant be done, think of the hundreds of thousands of devs having to communicate with one another it would be like 15 TB of mailing list records every year.
YES, its an idiotic idea, linux, at least from where i look at it, means choice. I don't want a gaming distro, i don't care about games, in fact i would be bothered by preinstalled packages destined for gaming. I dont want flashy presets, I like to build my computer's look myself to an extent of 100%. On the other hand there are tons of people who wouldn't touch any os with little to no gaming support, and don't have the desire to build pixmaps for their themes and would instead need some flashy presets.
Just an example of how hard it would be to unify everything into the "GOD" distro.

[flarkit]

Hi

I'm sure this has been rehashed a few times, but I'd like to hear some views from the community in-case I missed the discussion.

I'd just like to better understand why the world's brilliant Linux skills are scattered across so many distributions. If that mental power was focused on a smaller group of OS's, wouldn't we be able to take over the world (figuratively speaking)?

It may seem a little boring, but surely each of these would benefit from a concentrated effort to be great at what they are intended for. Instead there's this ongoing "power struggle" between the distro's. This does potentially lead to improvement as better features and functionality is thought up, but it still seems (to me) as if the community resources are being squandered in different, diverse directions rather than in a concerted effort to make Linux greater.

I'm not trying to combine the whole of Linux into a single OS that is bogged down by irrelevant functionality. IMO there should be 3 or 4 directions, catering for various "specialisations":

Linux Enterprise, Linux Studio and Linux Home/Office. The Kernel would still be at the core, naturally.

I like the notion of "Linterprise", LinStudio and "LinHoff"
:)

Surely it would go a long way towards delivering a package which more people and companies would trust. Still adhering to the OpenSource vision, but in a less scattered way.

Am I missing a key justification for having so many brilliant minds working in different directions? Is there no smooth, natural way to bring these resources together in a more focused direction? Or is it simply the natural way for people to prefer diversity?

Just a tiny rambling notion

[loell]

to be concise,

main reason.

People don't agree on which is better over the other on many different things, hence the fragmentation.

[Dark Star]

http://tuxenclave.wordpress.com/2007/10/30/major-linux-distros-shootout/

Check thismight help you :)

[flarkit]

to be concise,

main reason.

People don't agree on which is better over the other on many different things, hence the fragmentation.

I understand that and agree entirely, with one caveat. An endeavour like this, to have a consolidated effort, should be done under the vision of an architect (looks at the Torvalds's and Shuttleworths out there).

It shouldn't dampen creativity, since this is necessary for innovation, but surely we are all part of the Linux effort, not so?

[loell]

I understand that and agree entirely, with one caveat. An endeavour like this, to have a consolidated effort, should be done under the vision of an architect (looks at the Torvalds's and Shuttleworths out there).

It shouldn't dampen creativity, since this is necessary for innovation, but surely we are all part of the Linux effort, not so?

this is where the part where it gets repeated, we might end up in the recurring discussion area, since as you know this has been discussed over and over.

so i'm trying to avoid repeated reasons.

but another thing is concise. its all about Freedom...

[flarkit]

so i'm trying to avoid repeated reasons.

but another thing is concise. its all about Freedom...

I agree again.

My point-of-departure (which I failed to mention, apologies), is the fact that whilst the resources are scattered across many distros, Linux itself (as an OS) will not be able to take the market share from "the behemoth". I believe that individuals and companies will remain hesitant to adopt it, since there are so many competing factions in the community.

I fully agree with you that the freedom side is an important aspect of Linux too. But is this not indirectly preventing us from attracting the majority of users?

[popch]

main reason.

People don't agree on which is better over the other on many different things, hence the fragmentation.

Main reason,

people can not agree on what's best because people (and organisations) in fact need different things.

[jpittack]

Frankly, I love the idea of multiple distros. I have only tried Ubuntu, but when I get some old desktops that people don't want anymore, I will install and try as many distros as I can. Why? I'm freaking addicted to trying to new stuff and learning as much as I can about anything computer, engineering, graphic, or science related. I will at one point want more distros. Sadly, I don't think I will ever be able to create my own, as I am pursing a college degree and career that will not allow me to. At least I can alpha test a few.

[tubasoldier]

I like installing different distros as well.

However, I can understand where he is coming from. It seems there is a new distro every month. That is a major waste of developer time and manpower. Just go to distrowatch.com and see how many there really are. I seriously think that if a distro is not in the top 100 then the developers should spend their time working towards making more mainstream distros better. Of course there are distros specified to do different things that people may need.

Vixta recently came out. It will help prove my point. It is a Fedora based distro whose only difference is that it has a theme that looks like MS Vista. What a damn waste. That kind of crap should just stop. Just make an ugly theme for everyone and move on.

[loell]

Main reason,

people can not agree on what's best because people (and organisations) in fact need different things.

maybe, maybe not, but we are talking about development and resource fragmentation here,

for example in ubuntu circle, things aren't fragmented, there are variance to fit a specific need.

its possible to have variations without fragmentation.

[flarkit]

Added a poll, to see what the Ubuntu visitors think
:KS

[PartisanEntity]

I have mixed feelings on this topic, on the one hand, obviously having competition speeds up innovations and improvements, but at the same time, having too many distros might also squander resources.

It's one of thos topics that keep coming up. I also agree with popch, needs will lead to various distros and projects.

[loell]

I fully agree with you that the freedom side is an important aspect of Linux too. But is this not indirectly preventing us from attracting the majority of users?

depends, cause though difficult to prove, it is possible to have one thousand distributions and all of them is easier to use than any windows OS and just works ;)

i really don't believe in the notion that too many choices creates paralysis its really the lack of information that creates it.

[frup]

I think it is good that distros not only compete with other OS's but also each other I think this pushes a release to be good (and will do so more in the future).

[Dixon Bainbridge]

Hi

I'm sure this has been rehashed a few times, but I'd like to hear some views from the community in-case I missed the discussion.

I'd just like to better understand why the world's brilliant Linux skills are scattered across so many distributions. If that mental power was focused on a smaller group of OS's, wouldn't we be able to take over the world (figuratively speaking)?

It may seem a little boring, but surely each of these would benefit from a concentrated effort to be great at what they are intended for. Instead there's this ongoing "power struggle" between the distro's. This does potentially lead to improvement as better features and functionality is thought up, but it still seems (to me) as if the community resources are being squandered in different, diverse directions rather than in a concerted effort to make Linux greater.

I'm not trying to combine the whole of Linux into a single OS that is bogged down by irrelevant functionality. IMO there should be 3 or 4 directions, catering for various "specialisations":

Linux Enterprise, Linux Studio and Linux Home/Office. The Kernel would still be at the core, naturally.

I like the notion of "Linterprise", LinStudio and "LinHoff"
:)

Surely it would go a long way towards delivering a package which more people and companies would trust. Still adhering to the OpenSource vision, but in a less scattered way.

Am I missing a key justification for having so many brilliant minds working in different directions? Is there no smooth, natural way to bring these resources together in a more focused direction? Or is it simply the natural way for people to prefer diversity?

Just a tiny rambling notion

Alot of linux zealots post threads like "what can be done to increase linux market share/uptake," and "how can we spread the word and convert MS users...."

That's when I swtich off. Its not a religion, its an OS. A tool to achieve whatever goals you have. Linux is not a cult you buy into and need converting too, because the message Linux is selling is so much better than the others (the heathens!) etc. Its just an OS. I don't understand why people spend so much time thinking about doing this.

In terms of innovation, I don't think consolidating distro's would achieve anything significant. I don't see the need to consolidate from an innovation and technical point of view, or a market uptake point of view. Linux is just linux. Free distro's no real tangible business model.

[flarkit]

<duplicate deleted>

[flarkit]

Thanks for the input. I guess this idea took hold after thinking about some of the policies and methods which MS uses to keep people locked into their huge part of the market. I'd really like to see people freed to enjoy the benefits of OSS (and maybe even improving it).

If the vision for Linux is not to provide a solid, free alternative OS to the costly, virus-plagued favourite, but instead to be an open platform for encouraging geeks to explore their creativity, then it's succeeding very well. This is not a criticism in any way either. Merely my attempt to understand Linux's purpose.

BTW: the poll-results have it... :)

[mosaic2s]

development needs creativity and this is best if all directions are explored. then pick n choose. best to go this way for now. and keep on adding the stable features in all distros.

[EdThaSlayer]

Are you telling us to do something we don't want to do? I get your idea though :D and before I would have agreed with you, until I tried many different distros.

[Kowalski_GT-R]

Alot of linux zealots post threads like "what can be done to increase linux market share/uptake," and "how can we spread the word and convert MS users...."

That's when I swtich off. Its not a religion, its an OS. A tool to achieve whatever goals you have. Linux is not a cult you buy into and need converting too, because the message Linux is selling is so much better than the others (the heathens!) etc. Its just an OS. I don't understand why people spend so much time thinking about doing this.

In terms of innovation, I don't think consolidating distro's would achieve anything significant. I don't see the need to consolidate from an innovation and technical point of view, or a market uptake point of view. Linux is just linux. Free distro's no real tangible business model.

I don't see zealots around. Relax and don't switch off :)
flarkit has well argumented his concerns, which are also mine.

Its the drive to have a better open source OS.
When I approached Linux I was after security, stability and the dreaded user-friendliness.
As much as I appreciate my Linux, I must admit that most of the times, in these aspects it failed to deliver.

Distro-hop is not a sport for everyone...

O/T: my compliments to flarkit on a well argumented thread. It's very good to share opinions with like-minded people. This forum is something special.
ciao,

Andre

[beniwtv]

Some people seem to forget an important point:

If we would create a "unified" distro, it wouldn't help us much anyway.

For example, the devs from one distro wouldn't want to work on another distro, because they want their OS to be different.

Hence, this is why they created the distro in the first place.

So cutting down the distros only would cut of developers. Not to mention, even if there are like 400+ distros out there, most of them (or nearly all) develop GPL software, which any other distro can benefit from.

The important thing about Linux, and FOSS in general is that you have choice, you can use a tool that does exactly what you want.

I always wonder why people say Linux has to much choice, when we have like 100+ car companies, making diferrent cars, but we don't complain about that? Or having hundreds of motherboard manufacturers? Mobile phones?

I think the only thing we could do to improve this is to make as much as documentation available, so users can make a choice, knowing their options. Because choosing an option out of hundreds which you don't know of is hard.

[speedwell68]

The thing abou Linux is freedom of choice. I have settled with Ubuntu because it works well on my hardware, out of the box. I was using Suse and gettin nowt but problems. I have a mate in China and he swears by Suse, his choice for his reasons. Consolidated distros would soon become Windows.

[Kowalski_GT-R]

you can use a tool that does exactly what you want.

Err.....no?

I think the point of the thread is: "can we have a better tool, please?"

it makes sense....see all the problems with gutsy. (I also had with DSL and Puppy).

4 distros as in the original post here can be a bit too much, but I agree that fragmentation, as much as I understand it, prevents the upcome of Linux, and ultimately, of open source. (and ultimately, of a better desktop OS which puts the big names on their cards)

I hope I made myself clear, whereas my english becomes inefficient :)

[beniwtv]

Err.....no?

I think the point of the thread is: "can we have a better tool, please?"

it makes sense....see all the problems with gutsy. (I also had with DSL and Puppy).

4 distros as in the original post here can be a bit too much, but I agree that fragmentation, as much as I understand it, prevents the upcome of Linux, and ultimately, of open source. (and ultimately, of a better desktop OS which puts the big names on their cards)

I hope I made myself clear, whereas my english becomes inefficient :)

Gutsy had no more problems for me than Feisty had - this depends on your hardware I guess...

I think open source is currently doing very well. For instance, with Ubuntu I'm able to have all the tools I need for my job easily.

But this is because Ubuntu is configured by default the way I like it, needing only a few small changes.

Other distros don't behave by default the way I like it, for various rasons, and a unified distro wouldn't either. Windows also doesn't fit my needs.

And the most important thing is that you can't please everyone with the same thing (see my post above about car companies).

Agreed that some other distros can please much people out there, but not me. Not everyone. So I certainly would need to change them until they fit my needs, hence a new OS is born...

See my point? Even if we create a "unified" distro, people would complain, eventually starting a new distro, and then all the "fragmentation" starts again.

And it certainly doesn't slow down open source development. It may slow down adoption, maybe.

If we really want more people to adopt Linux, we will need to show and explain them their choices. Much people aren't even aware that Linux exists. I even know computer stores that don't know what Macs are.

That being said, I could not be happier with Gutsy right now, it does just everything I ask it for -- :popcorn:

[Kowalski_GT-R]

Gutsy had no more problems for me than Feisty had - this depends on your hardware I guess...

The gist of this thread is right HERE.

maybe flarkit can clarify.

Installation / distro-hop is a sport, and a b....y good one at it: i enjoy its benefits myself :), but,

my rig is mission-critical. I need it every single day.
A hassle-free open source experience is something which I feel somehow the world is still lacking.

This leads to the mumbling which originated this very thread.
BTW I voted yes. I think flarkit has a point.

ciao all,


Andre

[flarkit]

Good to see some debate going on. :)

I wouldn't compare the plethora of Linux distro's to the choice of car manufacturers and models, since one can't really customise and update your car with the ease that you can change your OS.

I think the point of developers working on different distros was born from the idea of being free to customise your Linux as you like. And perhaps a (pipe)dream would be to retain the freedom to customise your Linux, whilst benefitting from having a more unified effort into improving it's strengths. Also, to somehow present that in a way which does not seem as if the developers are competing with each other.

I still feel that concentrating the large number of Linux developers (and this refers to technical and creative skills), would improve the stability and features of Linux. At the risk of sounding like a stuck vinyl record.

[original_jamingrit]

You're correct in assuming that this idea is a re-hash of what other people new to the linux community have suggested. Here are the reasons that make sense to me:

A: As mentioned many times above, it's about freedom, freedom to choose and all that.

B: Almagamating all distros probably won't make FOSS guys work harder, because there is not as much duplicated effort taking place currently as you might think. Guys working at launchpad come up with fixes, and when those fixes might apply to different distros, people from other distros take note. Many wonderful men and women work on different apps that are used in different distros, and the features/bug fixes and things they make and release are given to every distro.

C: And we (or hopefully most of us) are not here to take over the marketshare. We're not here to make everyone in the world use GNU/Linux. It would be nice if everyone preferred GNU.Linux, but not everyone does. When I switched to Linux, I was not thinking to myself; "Hey, I'm going to help an OS different than Windows take over the marketshare!". I was thinking :"Hey, this looks neat, and different."

[Thyme]

I very much like the fact that there a many distributions out there, and that one day I myself may even be able to have a distribution of my own (one day .. !). I think that the ideas behind each distribution are great and having many different distributions with different styles just make everything more interesting and fun.

I can't imagine Patrick Volkerding's "keep it simple" philosophy on Slackware being included in any maintream distribution. Also, having multiple distributions will improve competition tremendously within the open source community and also maintain high enthusiasm among participents.

[Bloch]

There is also the point that companies that might want to support linux are faced with 3 or 4 main distributions. For example when you go to a website to download software you see 1 windows .exe, 1 OS X .dmg file .. ... and then for linux a .deb, an .rpm, another deb for ubuntu, a .tar.gz installer PLUS a few lines on what hacks you might need.

example
http://www.pandanet.co.jp/English/glgo/download.html

So hardware makers creating drivers, sites with streaming media, software makers etc would need to have a tech guy proficient in 3 distros to be their linux-man.

I think in the past year Ubuntu has *almost* taken on the role of the standard linux system to test on. Just from browsing around, whenever a site/product mentions linux compatibility they will mention ubuntu.

This is a good thing.

[fuscia]

i doubt puppy users and pclinuxos users would ever find a combo distro as satifying as those two distinct distros. it's the same as the kde vs. gnome discussion: they are different enough, with huge followings, that there needs to be both of them. better to have a lot of different bunches of happy people than one big bunch of somewhat disappointed people.

[Dixon Bainbridge]

I don't see zealots around. Relax and don't switch off :)

I wasn't implying there were linux zealots in this thread, I was speaking in general terms. Hang out on most linux forums and the same recurring arguments about increasing market share come up. I find that talk pointless.


We all want a better tool to work with. I just don't think that consolidating will necessitate that end. It doesnt matter that there are hundreds of distros of linux, as long as learning is shared, who cares?


Consolidating learning across distro's is more important and relevant than consolidating the platform itself. That's my point.

[aysiu]

I'm sure this has been rehashed a few times, but I'd like to hear some views from the community in-case I missed the discussion. Don't worry. You can catch up on the discussion. I've merged your thread with the other discussions on the same topic.

[kopinux]

1. the 86pc is linux playground, so i think it will stay like this forever, it cant be done, until the 86pc dies or something.

different people, different distro, different hardware, different problems, different solutions.


2. though there is someplace where linux will really excell, its called other standard hardwares/platforms, embedded or closed/fixed devices.

different people, different distro, same hardware, same problems, same solutions.

[flarkit]

I've only recently had the notion about combining the efforts of so many Linux developers. So I needed as much input as possible to see more aspects to this. I realise that there are indeed many standard features and there are many shared tools and methodologies which are cross-pollinated between distros.

It may very-well happen that a new generation of tech-savvy users will prefer OSS + Linux to Windows, but at the same time, I'm quite sure there's an overwhelming number of entrenched Windows users who will stay with their platform, because it seems to be easier to use than the complicated-seeming, command-line-interface-driven FREE alternative.

Linux is slowly improving that side (indeed, I firmly contribute much of this to Ubuntu), but it hasn't quite nailed it. It's understandable that this is very difficult to achieve whilst retaining the freedom to customise as one prefers. The challenge also has another big aspect, which is innovation. People need to be wowed with incredible new, fun and easy ways to do things, to draw their attention. Things like Compiz have started Linux on the path to a whole new interface and I reckon new desktop hardware will play a role in that. But the guys on the proprietary side of the fence are working just as hard at this.

So, my concern is that an evolutionary adoption of the OSS platform(s) is not going to happen fast enough to prevent people from staying with the company that they've grown accustomed to. I think someone with the required insight and presence needs to stand up and get the community organised in a more focused way. (looks pointedly at Mssr's Torvalds and Shuttleworth)

Perhaps it is undeniable that the freedom-to-do-your-own-thing has to be retained, to keep all the free thinking supporters of OSS working on Linux. Perhaps it will grow towards the platform that convinces the unconvinced companies and individuals that they should migrate. BUT, can we depend on human nature to see the light? Or do we need the architects to step up and give clearer guidance to the efforts of all the enthusiasts who are scurrying around in their own little camps?

[I'm not sure where in my head the above all flowed forth from, but I do feel strongly about it, even after reading the excellent counter-arguments.
:)]

[chrismine]

I have read somewhere there is about 500 Linux distro's or flavour's available.

There must be a lot of brainpower out there - what will happen if these people together to work on just one distro - I think Linux will progress faster and we will have better programs and issues will be sorted out more efficient.

Or will it be a bloated inefficient OS?

What do you people think? Be interesting to hear your opinions..

[conehead77]

I have read somewhere there is about 500 Linux distro's or flavour's available.

There must be a lot of brainpower out there - what will happen if these people together to work on just one distro - I think Linux will progress faster and we will have better programs and issues will be sorted out more efficient.

Or will it be a bloated inefficient OS?

What do you people think? Be interesting to hear your opinions..

They would argue which would be the best method to make a good OS and we would end with nothing but windows.

[chrismine]

They would argue which would be the best method to make a good OS and we would end with nothing but windows.

ROFL!!!!

[de_valentin]

Yes there is a very good reason they are not working together...they all have other wishes and demands.
Even if they could work together well, they would have to compromise on everything, and that would probably leave you with something that might be a lot like windows it works for the masses but it's not that good. It's comparable to the EU, it's great to work together but the more member the more opinions you get the less efficient you become.

The best way would be to find the distro that suits you best and help them become the best they can be...But i think that's where we are now. Even the reason most people are on this forum at all.

[aysiu]

I've merged this with the other thread on the same topic.

[hutxubix]

Linux and Open Source software is an amazing phenomenon, but it is also amazing the number of distros on the air [I just realized in watchdistros.com].

Ok, I think I know that one of the reasons for that is that for different needes, different distros, but I still think that there are efforts made 10 times.

Don't you think that a unification of distros to just, let me say, 5, will be a great advantage for the Linux world?

Instead of 50 developers that solve the same problem on their own, we would have 10 (still a nice number) and the other 40 could do more things in all the projects that don't advance as they could.

What do you think?

[Kosimo]

I'm agree with that..

We have a common enemy, and we should fight together.

Kernel folks are doing the same kernel for everybody by the way, and this, is the core of all distros.... So, we've got some very important thing in common.

Then, GUI's... The most popular are just two GNOME & KDE, and in this area I think is good to have a good and fair competition between them, so, everyone is trying all the time to be the best.

And about distros....

Probably yes, but kind of impossible...

reason? I don't really know...

[bvanaerde]

We have a common enemy, and we should fight together.

I know where you're going with the "common enemy"... but I have to disagree.
Just choose whichever distro that suits you best (even if this is Windows), and be glad with having a choice.

[23meg]

but I still think that there are efforts made 10 times.

That there are 10 times the amount of distros than actually needed (which would be an argument hard to substantiate) doesn't mean 10 times the effort is made, because most of the code is reused, and most of the work is done upstream.

Instead of 50 developers that solve the same problem on their own, we would have 10 (still a nice number) and the other 40 could do more things in all the projects that don't advance as they could.

Distribution developers don't develop most of the software that make up the distro; their job is to put software developed by upstream projects together in a way that makes sense for the intended audience, make sure it all works together, and maintain it. And even that effort is being duplicated less and less due to the proliferation of derivatives.

What do you think?

Here's more of what I think:

http://ubuntuforums.org/showpost.php?p=2200966&postcount=477
http://ubuntuforums.org/showpost.php?p=1339459&postcount=28

[23meg]

Kernel folks are doing the same kernel for everybody by the way, and this, is the core of all distros.... So, we've got some very important thing in common.

http://ubuntuforums.org/showthread.php?p=2291161#post2291161

[mcduck]

Reason? How would you tell everybody that Linux is no more free operating system that you can develop into your own distribution as you wish, but instead from this point on you'll just have to work on one of the biggest distributions doing tasks somebody else assigns to you?

As long as people have the freedom to build Linux into what they want, they are going to do that.

Also it's worth to notice that most developers aren't "fighting any common enemy", they are just building programs and tools and distributions that they like to build and that people like to use.

[aysiu]

They are working together. For more details, read the thread I merged yours with.

[Iandefor]

I have read somewhere there is about 500 Linux distro's or flavour's available.

There must be a lot of brainpower out there - what will happen if these people together to work on just one distro - I think Linux will progress faster and we will have better programs and issues will be sorted out more efficient.

Or will it be a bloated inefficient OS?

What do you people think? Be interesting to hear your opinions..I'm thinking that even if you managed to cause the developers of all 500 of the Linux distros to agree to work together on a single project, the project would fork until there were 500 Linux distros again.

The reason there are that many distros available is because each distro has its own goals and aims. Putting all the developers under one roof just ensures that you won't get any developers' needs met.

The Linux Mint guys would get in catfights with the Ubuntu guys over otherwise illegal codecs and the Ubuntu and Debian people would get into fights over restricted drivers and things like Sysvinit vs. Upstart and Debian and Fedora would have fights over which versions of packages to include and packaging policy, and everybody would fight over the release schedule. The only way to make it work would be to institute a dictatorship (benevolent or otherwise) and make everybody work towards the same goals.

At which point, you will have killed Linux as a Free operating system.

[RAV TUX]

Many can Talk about a Unified Linux, others put it into action:

http://cafelinux.org/ueosc/

Keep an eye out this is just the tip of the iceberg.

;)

[pjkoczan]

One of the great beauties of Linux is that it's chaotic. Hundreds of distros mean hundreds of ideas, and hundreds of ways to improve upon existing ideas. Not all of them work, but Linux would not have evolved as quickly as it did. Not to mention that more competition leads to a better product.

Look at the BSD community. There are relatively few versions of BSD, and development is slower and more deliberate. Stability is usually better than a comparable Linux distro, but it isn't as cutting edge. Installation procedures are much less-straightforward (in my opinion), and it takes a lot more work to get all the nice new things you have mostly by default in Linux to work in BSD, especially with regards to slightly unsupported hardware.

[Iandefor]

EDIT: Nevermind.

[Paradoxfox93]

Sounds terrible and SHOULDN'T be done.

The OP has links to breakmonopoly and yet suggests a unified linux distro? Now maybe because I'm fresh I don't know enough about linux to see the other side of the agruement (practically of effort to results) but it seems to me that that's just working in the opposite direction of the very purpose of linux. Moreover, it also seems me, correct if I am wrong, but that all of linux does work together in a shambling sort of way. Programmers sharing codes, experience, techniques and putting their work out their for all to see, improve and learn from. To me, that suggests a beautiful anarchic unification versus microcrack fascism.

[Yfrwlf]

Yeah, I yap a lot. :)

I think the key is this: having software selection and configuration options is what matters. As long as you have that, it shouldn't matter what distro you are running.

If you want package selection A, and your friend wants selection B, great. You should both be able to get them, and also download program C and use it.

If you want pre-configuration A and your friend wants B, fine. Just be sure that you can both run program C if you want it.

In other words, what is the harm in multiple distributions? Nothing. Not if they are merely a selection of packages that you can install on any other distro.

If we had ONE repository that all Linux installers could download from, and "metapackages" or "styles" or "groupings" of packages and configuration options we could choose from, it would greatly simplify things for everyone. Wait, I'm not finished.

Imagine if there was no longer a segregation of distros. If all there were were *programs*. THE Linux community, as well as other OS communities, would surround and talk about programs. That's where the emphasis should be. If everything was completely modular so that we always had the choice to install the Red Hat Printer Utility, or so that we could install nm-applet (the Network Manager) on Linux, meaning regardless of our selection of other packages, it would give us all much more freedom. That's what Linux is all about.

Dependency issues? Sure, they may be in there, but those are problems to be solved. There is no reason why you can't allow two programs to exist side-by-side, and if for some reason they can't be worked out immediately, then temporarily declare them as incompatible, but I think if all bugs get worked out so that everything is perfectly modular and beautiful, that everything should be installable.

You think there's a problem with some programs needing a different library version for every single different program version? Maybe finally moving into a world that no longer demands constant recompilation will make developers start actually forming standards for their libraries, so that every version doesn't break compatibility, but still allows room to grow and improve. Even if this were true, there should always be a way to install different versions of different libraries side-by-side until some kind of standard API is created for them.

So you want choice, even, more choice in the Linux realm, and you think making different distros matter less may be one key to that? I completely agree. We will have more choice if this happens.

And what of Ubuntu? Well, Ubuntu will be merely a selection of packages and configuration files for those packages. But then you could subtract a few, or add more of course, if you wanted to, at installation time. "Rolling your own distro" would be a cakewalk, as it's merely a selection of packages and configuration files. The point is you'd have access to a much, much larger software repository if you wanted it, among other benefits.

Could something like this ever happen? I sure hope so. Developing more standards to make sure programs have the ability to stay cross-platform is one really important step toward distro-freedom. It would also allow more competition between programs, so the "best" ones can come out of the woodwork more easily. Once programs are more modular and freely transferable, this will get much closer to being able to have happen.

One small team working on one large repository (and it's mirrors), testing package installation and fixing bugs, making everything function with everything else. Actual Linux program unity. If programs were smart, if libraries used intelligent APIs, if thought was put into programs to ensure compatibility and modularity, the Linux world would be unstoppable.

Differences can exist when we have the freedom to install what we want. Playing in the same ballpark only gives us all more freedom to pick and choose. United we are strong.

You can go off and make your own kernel and the rest of the programs to run on top of it. If the argument against all this is that it will give you more freedom, it won't. You will have things just how you wanted them, or rather, how you made them, but you will not have all the programs the rest of the community has made. A community can only work together when they can all access each other's work. The easier you can do that, the easier it will be working together. Instead of making your own kernel then, how about you come play with kernels compatible with the Linux/Gnu OS? Improve them, add features to allow you to turn on and enable those things that you want. Don't like the compatibility system? Help improve it. Don't like any part of the kernel at all, or feel it'd be easier starting from scratch? Go ahead. But, you might consider making yours compatible with the Gnu OS, so then you can run all those programs which exist. Then, everyone can have a choice, to run Linux, or run your kernel, if you let them, but they and you will still be able to run everything else you all love too. That makes everyone more free.

Standards/frameworks/APIs/ABIs can co-exist with having everything the way you want. They can be very flexible if created intelligently, and can really help in many ways, the most important of which to grant more freedom. If you don't want to use them, you aren't forced to. For everyone else though, they need the freedom to choose them, to reap the benefits they provide.

One last thing. If Linux was unified enough to make so programs, program groups, and configuration groups were all that mattered, it would not mean Canonical would lose it's income source, it wouldn't mean Ubuntu would lose it's flare, it would simply mean the Ubuntu Package Selection would be what was supported, and it would still mean they could have a revenue. But, it would mean their packages and selection of packages would have more competition, because with just a few changes in packages, many, many more Ubuntu-like selections could come about more easily. Right now there are things like Ubuntu Mint and others, and those would become a bit more numerous possibly. Though, at the same time, maybe they wouldn't because with being able to install packages from within your package group, or outside of it, only being a few clicks away, perhaps more users would just choose "Ubuntu" and stick with it because they know it. So I hope Canonical, Red Hat, Novell, and all other Linux-based companies recognize the value in pulling together, and in accomplishing this task, because anything that makes Linux strong will make them strong too. The day Linux companies start trying to fragment and make proprietary Linux to try to get an "exclusivity advantage" will be a sad one, if that day hasn't already occurred. It's what companies like Microsoft would love to see happen.

Rebuttals and other ideas are appreciated.

[IsawSp4rks]

The idea sounds bloody awful and should never even be entertained.

Linux is fundamentally about FOSS, once you start bringing all the different interpretations and ideas under one umbrella or banner called Linux you both water down what FOSS represents and streamline what it practically does and therefore FOSS (and derivatives), GPL (and derivatives) and GNU/Linux (as ethos and practical application) each become hollow and pointless.

Not to put to fine a point on this, but streamlining Linux is the dream of desktop Windows users who wish Linux would engage in catch up with MS so that they could in effect get Windows - or something almost exactly like it - for free.

:popcorn:

[aysiu]

The Linux community does share programs. In Ubuntu, right-click on the Network Manager applet, and you'll see both Red Hat and Novell had a hand in making it (the icon for the applet was created by a Red Hat employee; I'm not sure what exactly Novell did).

Even though Novell is the main force behind Evolution, Evolution is a part of Gnome, and any Linux distribution can use it. GIMP on Ubuntu is pretty much the same as GIMP on Mepis or GIMP on PCLinuxOS. OpenOffice is sponsored by Sun, but it works on all Linux distributions the same way.

The whole point of open source is that we share. I don't see what you're talking about Linux distributions needing to share the same applications--they already do.

[Yfrwlf]

The Linux community does share programs. In Ubuntu, right-click on the Network Manager applet, and you'll see both Red Hat and Novell had a hand in making it (the icon for the applet was created by a Red Hat employee; I'm not sure what exactly Novell did).

Even though Novell is the main force behind Evolution, Evolution is a part of Gnome, and any Linux distribution can use it. GIMP on Ubuntu is pretty much the same as GIMP on Mepis or GIMP on PCLinuxOS. OpenOffice is sponsored by Sun, but it works on all Linux distributions the same way.

The whole point of open source is that we share. I don't see what you're talking about Linux distributions needing to share the same applications--they already do.

My point was they will get shared more, and by those who aren't tech-savvy, and they won't have to rely on a bunch of distro packagers and devs to do it for them. It would make everyone's life easier. I'm talking about anyone easily installing any so-called third-party software, i.e., outside the repos. I don't want there to be such a thing as third party software, but merely "unsupported" software by certain companies. It'd be great to be able to easily have anyone install software outside the repos by either downloading a package that just works, or by adding someone's repo link so you can get all future updates. Sure, as long as there is at least source compatibility, distro packagers will be able to share, but we want things to be better than that, we want users to be able to share and have the packages they want regardless of what "distro" they're running (what should be a simple selection of packages).

Ordinary users will be able to easily install a lot more programs if the repositories are bigger. The repositories would be bigger if distros combined their efforts. If Linux had a standardized ABI, third-party installation would be a snap. Try telling Ubuntu_Newb_52385 that they can just easily install programs outside the repository. If this was true, you would NOT need to install any specific distro, because distros wouldn't matter. Due to the problems with incompatibilities, dependency issues, compilation errors, and in general because Linux isn't as standardized as it should be, Linux programs are not easily transportable the way they should be. I'll say it again, you shouldn't have to use distro SuperCoolLinux5000 just to get program X. If a "distro" was a mere grouping of packages and different configuration files, and it was completely modular, you wouldn't have to get stuff from Ubuntu's repos, you could connect to any repos and add them all together.

Quite simply, we (Linux users) want to be in control of our applications and to be able to install the ones where and when we want to, and to make backups and archives and to share them online with friends, we don't want to have that done by distro packagers, waiting and depending on them to gain access to our programs just because Linux isn't "compatible enough". We do not want to be restricted by stupid incompatibilities just because we chose a different "flavor", because that's not what they are, they are different OSes, and it's much harder for Linux to survive like that. Linux will be much better off once it's all "Linux", once everything is compatible and modular.

If you can't grasp how pulling everything together could be more beneficial, perhaps you don't understand the power of communities, and what makes Linux as great as it is right now. If we continue that and make that stronger, Linux will be stronger. I believe the OP had a very good point in that Linux, further united than it is now, would be more powerful. Competitors would love to see Linux completely fragment, and we shouldn't let it, we should make it do the opposite.

Software Installation on Linux: Today, it Sucks, Part 1 (http://ianmurdock.com/2006/12/15/software-installation-on-linux-today-it-sucks-part-1/)
Part 2: Tomorrow, it Won't (http://ianmurdock.com/2006/12/19/software-installation-on-linux-tomorrow-it-wont-with-some-cooperation-part-2/)
http://www.linux-foundation.org/en/LSB
http://www.linux-foundation.org/en/Packaging
http://kerneltrap.org/node/4006

[cardinals_fan]

What an appalling idea. Everyone's needs are different, and there cannot be one distro that satisfies everyone. With that said, the idea of a unified package format is interesting. One trait that I think WILL become almost universal in Linux distros over the coming years is a rolling release cycle. There are too many advantages to rolling release to list them all here, but the primary one is that a distro can always be up to date with one command/click in a gui.

[Foster Grant]

Hi,

besides Canonical being Ltd., there is one more thing that drives me crazy - why do we the hell have that many linux distros around?

Isn't is different folks developing parallely exactly the same things? This greatly decreases the productivity of the whole community. I mean sure a lot of distros use the same packets etc., but for example the PROMOTION, which linux really really need in order for OEM to ship new pc with LINUX INSIDE, instead of that Microsoft crap beeing in every god damn mobile device now (got a PocketPC recently - Win mobile 2003 - and I am about to spit fire and throw it against the wall if I wont get linux inside soon!!!

I would be pretty happy with one universal distro around, beeing nicely boosted with features for everyone... I think Ubuntu goes this way and thats what i love it for. Until someone in Canonical screws up...

Jan

Well, I wouldn't. Different needs require different toolsets. We have the big corporate distros (Red Hat and SUSE). We have the advanced-user distros (Debian, Fedora, Slackware). We have the userfriendly distros (Ubuntu, Xandros, PCLOS). We have an enterprise-strength distro that's community-based for smaller companies and non-profits (CentOS). We have small distros for older PCs (DSL, Puppy) and even old Macintosh hardware (Yellow Dog) and the PS3 (Yellow Dog again). And we have Freespire ... decide for yourself what niche it fills. :-P

Besides, it's been tried and found to be a failure. UnitedLinux was abandoned after version 1.0 (there were plans to abandon the project even before Caldera/SCO went Benedict Arnold on us).

http://www.unitedlinux.com/
http://en.wikipedia.org/wiki/United_Linux
http://news.zdnet.com/2100-3513_22-5146194.html

What we have is what seems to work.

[Zyphrexi]

don't forget damnsmalllinux, linuxfromscratch, and gentoo.

dsl - excellent small distro for fixing stuff
lfs - for the truly insane
gentoo - for people with too many cycles

[DeadSuperHero]

While I do think that dipping from a common codebase is good, I don't think most Linux distros do that enough. It'd be great to standardize on things such as libraries, dependencies, and all that. But, even that could be disasterously impractical.

[Tuxoid]

There is specific user-level advantages to the separation of Linux.
if you want to learn deeper stuff about Linux; Use Slackware, if you're a developer/tester; use Fedora, if you want a great desktop distro; use Ubuntu, OpenSuSE, or Linspire (note: this is only my opinion)

This also applies to Desktop Environments. Where I would suggest KDE to a power user (due to its easy customization), I would suggest GNOME to an average user doesn't want extra features that they say they won't use.

When it comes to choices like, package managers, update methods, shell options, etc. many computers users won't usually notice the differences in the developers ideal flawless ratification of the application. What windows or OS X user is going to notice differences between dpkg and rpm. I'm not saying that the differences are always unnoticeable, as Slackware's package manager doesn't even resolve dependencies by default. These differences are useless to the end-user (and really, just cause useless confusion).

So in turn, there are user-relevant choices, and user-irrelevant choices affecting the user experience in positive and negative ways.

[wdaniels]

I don't think the "can or can't be done" aspect makes much sense here - surely this is only a question of everyone agreeing to just work on a single distro, the answer to which is obviously a resounding "no", so it can't be done, but only because a significant proportion of Linux users and developers value precisely the opposite such that the means would defeat the end by definition.

It is not necessarily the objective of everyone in the Linux community to cause all computer users to run Linux, only perhaps to achieve sufficient market share to warrant decent support from hardware manufacturers and software developers, and in the same way raise awareness of the freedom to choose for the benefit of those who value it or most need it. But this can't come at the cost of shoehorning everybody into a single way of doing things, or there is not a lot of point.

[bannik]

so guys do you think its time for linux to fight back against microsoft.

with all the hype linux is getting i think it time for the making of the best linux ever, something that has to fight against micro and mac.

what i don't get is why do all the linux developers (opensuse, fedora, pclinuxos etc) unite and make 1 linux to rule them all, i mean they all believe in free software and do the work for free so wouldnt it be easier to all of them to join.

plus it would benefit linux, one of the biggest issues with linux is the fact that there are so many options to choose for a distro its getting silly now, we need 1 linux

[Methuselah]

Mandating a single linux distribution is contrary to what linux is about. Furthermore, it's unenforceable. Linux being GPL, anybody can fork it or package the kernel with a different suite of supporting software.

I kind of agree with the concern behind your statement though. It's good to have as much compatibility between distributions as is reasonably possible. I think this is what the Linux Standard Base is about.

[Mateo]

so guys do you think its time for linux to fight back against microsoft.

with all the hype linux is getting i think it time for the making of the best linux ever, something that has to fight against micro and mac.

what i don't get is why do all the linux developers (opensuse, fedora, pclinuxos etc) unite and make 1 linux to rule them all, i mean they all believe in free software and do the work for free so wouldnt it be easier to all of them to join.

plus it would benefit linux, one of the biggest issues with linux is the fact that there are so many options to choose for a distro its getting silly now, we need 1 linux

if google or some other established software giant wanted to challenge windows, they could. but most of the major players don't consider operating systems to be an important technological trend. The O.S. is essentially a tool to get you to the web. The web is where the money is. So see instant-on OSes (like Asus') become more popular in some segments.

[cardinals_fan]

what i don't get is why do all the linux developers (opensuse, fedora, pclinuxos etc) unite and make 1 linux to rule them all, i mean they all believe in free software and do the work for free so wouldnt it be easier to all of them to join.

plus it would benefit linux, one of the biggest issues with linux is the fact that there are so many options to choose for a distro its getting silly now, we need 1 linux
Oh, we need one Linux and we need the Linux developers to unite. We is such an easy word to use, but in many cases it really means I.

The reason your suggestion is bad is because everyone is different. You may like some things, but I like different stuff. For example, let's take a quick quiz:

Do you like the following:
Compiz?
GNOME?
OpenOffice?
PulseAudio?
Graphical Configuration Tools?

You likely answered "yes" to most of those questions. I answered "no" to all of them except the last one, where I say "rarely". A distro including Compiz, OpenOffice, or GNOME is not one that I can enjoy.

[SunnyRabbiera]

Really having one single Linux is a bad idea, it goes against the very concept of open source.
Though maybe more co operation between distros instead of the unification of them would be good.

[hardyn]

I also understand what you are saying, But having one linux would in many ways stifle creative flow. If you had a super-dooper amazing idea, but had to ask permission for it to be included "linux"... you would not be really impressed.

[cjnkns]

I believe not one linux is a good idea, but more cooperation betweeen the distributions and applications. The biggest problem right now I think (this is probably different for everyone) is the vendor support and applications. No one is makeing apps for linux OS's. There is just no money for commercial developers to make cash in it. Simply because the install base of linux is not as wide as that of Windows. I really wish that people would start building some good apps for linux that would support music players like my ipod and include music stores into banshee or rythmbox that had a good selection.

Until then ... guess I have to use Windows too. Simply becuase I am not good enough at programming to build linux apps. :(

[SunnyRabbiera]

I believe not one linux is a good idea, but more cooperation betweeen the distributions and applications. The biggest problem right now I think (this is probably different for everyone) is the vendor support and applications. No one is makeing apps for linux OS's. There is just no money for commercial developers to make cash in it. Simply because the install base of linux is not as wide as that of Windows. I really wish that people would start building some good apps for linux that would support music players like my ipod and include music stores into banshee or rythmbox that had a good selection.

well this is starting to change, with vista out there a lot of software developers are turning to linux.
Linux has a bright future ahead for it, each day there is something new for it and things are always improving.
I do feel though that a unified linux project of some sort should happen, where developers from all distros would get together and share developments instead of the bickering and in fighting we have now...
Debian, Mandriva, Ubuntu and anyone else who has not sold their soul to microsoft should pull together to at least collaborate with eachother
to make linux in general better, but stay independent too.

[jrusso2]

One Linux to rule them all!

[perce]

Everybody have different tastes and different needs. The main advantage of Linux over proprietary operating systems is that you can tailor it to suits your tastes and your needs. We don't need to loose our diversity, we need to stop all the silly religion wars like emacs/vi, gnome/kde, opensource/free software.

[SunnyRabbiera]

We don't need to loose our diversity, we need to stop all the silly religion wars like emacs/vi, gnome/kde, opensource/free software.

I agree, the in fighting does hold us back.

[Kinst]

Yeah let's just discontinue gnome, it's obsolete now that Qt is open source. :)

[SunnyRabbiera]

Yeah let's just discontinue gnome, it's obsolete now that Qt is open source. :)

Not for commercial use so GTK still has its place.
why do you think nero or AVG for linux use GTK? No fee :D

[akiratheoni]

Everybody have different tastes and different needs. The main advantage of Linux over proprietary operating systems is that you can tailor it to suits your tastes and your needs. We don't need to loose our diversity, we need to stop all the silly religion wars like emacs/vi, gnome/kde, opensource/free software.

Agreed, I don't like all of the free software in-fighting when we should instead be fighting Windows... but of course to fight Windows, we need to unify, which means agreeing on software, which just causes more in-fighting... it's a never ending cycle :(

[qazwsx]

Not for commercial use so GTK still has its place.
why do you think nero or AVG for linux use GTK? No fee :D
Skype, Google-Earth, Opera,...?

+QT allows non-free static builds and works better in cross platform (non-free GTK apps for Windows and OS X?)

Best part of it: Non-free software pays OSS development.

[dburnett77]

I, myself, consider myself to be an enthusiast.

If there were this grand war, I'm sure they're'd be losers. Oh, well...

I don't encourage dribble-drabble. If you like MS so much, us it and make money. Which you can also do with linux; ie, specific: Ubuntu Studio.

Yeah, use that and create a sound track the will be used in the next Block Buster. Make millions. If you think you're so important with what you have to say: Try pod casting. Several actors have begun careers this way.

ITunes will list you for free, and you will eternally have prescence...

[cardinals_fan]

Agreed, I don't like all of the free software in-fighting when we should instead be fighting Windows... but of course to fight Windows, we need to unify, which means agreeing on software, which just causes more in-fighting... it's a never ending cycle :(
Ah, the proverbial 'we' has returned. Why exactly do 'we' need to agree on software? And why exactly do 'we' need to fight Windows?

[Kinst]

Not for commercial use so GTK still has its place.
why do you think nero or AVG for linux use GTK? No fee :D

Hm, you're right, let's discontinue KDE, that would be easier.

Either way ubuntu's pretty much just a front-end to debian sid, we don't need both, let's kill ubuntu while we're at it.

[pjkoczan]

Linux (and OSS in general) has advance greatly in the past few years in no small part due to internal competition. Different projects (GNOME vs. KDE, MySQL vs. PostgreSQL) and different distros (Red Hat vs. Debian vs. Gentoo vs. Novell, et al) push each other to meet and exceed others' features and performance.

I could stand to see Linux less fragmented, but a completely unified Linux system, I believe, would be counter-productive.

It's fine to debate GNOME vs. KDE or vi vs. emacs or Linux vs. BSD or OSS project A vs. similar OSS project B, as long as those debating realize that their side wouldn't be nearly as far along without the other side.

[aysiu]

I've unified this discussion with the other discussions about a unified Linux distro.

[cardinals_fan]

Linux (and OSS in general) has advance greatly in the past few years in no small part due to internal competition. Different projects (GNOME vs. KDE, MySQL vs. PostgreSQL) and different distros (Red Hat vs. Debian vs. Gentoo vs. Novell, et al) push each other to meet and exceed others' features and performance.

I could stand to see Linux less fragmented, but a completely unified Linux system, I believe, would be counter-productive.

It's fine to debate GNOME vs. KDE or vi vs. emacs or Linux vs. BSD or OSS project A vs. similar OSS project B, as long as those debating realize that their side wouldn't be nearly as far along without the other side.

I've unified this discussion with the other discussions about a unified Linux distro.
+1 on both

[sicofante]

This thread is much too long (that's what happens when you start merging threads...) so I'm not sure if it's been said before:

I think there's no need to have just one unified distro, but there is a bare need for a unified platform. That means compatible binaries between distros, unified installation procedures, unified directory and system files placement, etc., etc., etc.

The fact that any developer must develop for a given distro is plain idiotic and won't allow Linux to move outside the catacombs for many many years (until a particular distro becomes a de facto standard and then everyone will develop for it only). I know many will say they can compile from source in any distro, but not everything is open source and not everyone is preaching for it to be. What's worse: most people loathe dealing with the command line, let alone a compiler. I can't find a single good reason why basic system things can't be shared by every distro, the same way the kernel is shared by everyone.

That's what's really needed: a unified platform and as many flavours (distros) as people wish.

[Superkoop]

I think Linux should just have standards, so that rather than projects having to create tons of different package installers, they only need to create one. This is the type of unifying that needs to be done. Making it so that all Linux is compatible with all the distros.

Different Distros is fine, having different programs and such. And GNOME and KDE should try to get together, and make it so that things can somewhat work more interchangeable with them. They should stay separate, but at least try to be compatible with the other almost somehow, like program level.
This would make things a lot easier for all the end users, being able to have everything working with whatever you use.

[cardinals_fan]

The fact that any developer must develop for a given distro is plain idiotic and won't allow Linux to move outside the catacombs for many many years (until a particular distro becomes a de facto standard and then everyone will develop for it only). I know many will say they can compile from source in any distro, but not everything is open source and not everyone is preaching for it to be. What's worse: most people loathe dealing with the command line, let alone a compiler. I can't find a single good reason why basic system things can't be shared by every distro, the same way the kernel is shared by everyone.

Give examples of what you mean. The 'basic system things' could be almost anything. In any case, I've never had trouble installing, say, Google Earth on any distro - using the binary from Google.

[sicofante]

Give examples of what you mean. The 'basic system things' could be almost anything. In any case, I've never had trouble installing, say, Google Earth on any distro - using the binary from Google.
Check the procedure for making Ethernet bonding in Fedora and Ubuntu, for placing of system files and administrative procedures.

Check the availability of DEBs vs RPMs at any major closed-source app.

If I were to develop a closed source app (which I might, BTW), what should I choose to package my app? Which distro should I try to test my app in? Do I wait until Ubuntu dominates and Debian derived distros are the standard or should Debian based distros, RedHat and Novel sit down and decide on using standards for all three?

Isn't it pretty obvious we need a platform instead of this mess?

[pjkoczan]

Check the procedure for making Ethernet bonding in Fedora and Ubuntu, for placing of system files and administrative procedures.

Check the availability of DEBs vs RPMs at any major closed-source app.

If I were to develop a closed source app (which I might, BTW), what should I choose to package my app? Which distro should I try to test my app in? Do I wait until Ubuntu dominates and Debian derived distros are the standard or should Debian based distros, RedHat and Novel sit down and decide on using standards for all three?

Isn't it pretty obvious we need a platform instead of this mess?

I believe that's the point of Linux Standard Base (http://www.linuxfoundation.org/en/LSB).

[sicofante]

I believe that's the point of Linux Standard Base (http://www.linuxfoundation.org/en/LSB).
Yes, but there's no serious commitment from the big three to LSB and if that doesn't happen, it's useless.

[cardinals_fan]

If I were to develop a closed source app (which I might, BTW), what should I choose to package my app? Which distro should I try to test my app in? Do I wait until Ubuntu dominates and Debian derived distros are the standard or should Debian based distros, RedHat and Novel sit down and decide on using standards for all three?
If there was a unified package standard, I would vote for Slackware's .tgz packages. They're very easy to create, reducing the strain on packagers.

[decoherence]

Well, I haven't read through the entire thread, but since world + dog has put in their two cents, I figured I would too.

Something non-programmers often don't realize is that a lot of linux development happens just for the fun of it. That's right: there are people who code for fun and they made most of the system you're using now.

So in many cases, a developer will purposefully duplicate another's effort, maybe because they don't like the other's code base, or attitude or chosen programming language, or maybe they think they can do better starting from the ground up, or maybe (as is most often the case for me and i think for most hobbiest programmers) they just want to know how something is done so they do it. Some times those little self-edifying projects turn in to mainstays of major distributions.

The reason this works is because there is no central body of management or regulation. One distro can use software X and another distro can use software Y and the general public gets an idea for which is better.

Don't get me wrong -- I think interoperability and standards are very important, and I would love it if I could sit down at a SuSE machine and not have to acquaint myself with all the things it does differently from Debian. However that can never happen through one body (even a consortium) enforcing a policy. The solution will be a technical one, arising organically from the efforts of programmers. No artificial regulation is needed. Only patience.

freedesktop.org project has made great strides in offering a set of standards that competing technologies can adhere to (or not) thus helping to break down the technical and usage related barriers between distros, desktop environments and so on.

[bks]

I don't know enough to say whether or not it can be done, but I don't think it should be done. I enjoy having all the different distros to choose from. Besides, some work better than others and all for different people (if that makes any sense)! I do, however, much prefer Ubuntu, but that's for another thread.

[Yfrwlf]

Yes, but there's no serious commitment from the big three to LSB and if that doesn't happen, it's useless.

That's fine. If they don't adopt the capabilities that users want, less users will use their distros. What I want first and foremost is the basic ability to install any programs I want to install, easily, meaning with no compilation. After the LSB Packaging API is finalized (see the Burgdorf API (http://www.linuxfoundation.org/en/Burgdorf_Packaging_API)), you'll be able to install any software package on any distro.

If there was a unified package standard, I would vote for Slackware's .tgz packages. They're very easy to create, reducing the strain on packagers.

Standardization isn't and never was about forcing any particular software on any particular distro. They can and should be able to use the software they want, and you as a user should have that same freedom. The standards are the things in *between* software. It's the communication APIs that allow all software to communicate. As far as the end user is concerned, this helps them have free access to software because of that modularity and interoperability. That is good for everyone, and will cause software to get more use. On the developer's side, they know this, and thus they are also interested in using some of the popular APIs/interfaces that are out there so that their program they develop won't be a complete waste. If they find that interface/API to be restrictive, then it should be *improved*. The best APIs are ones which are easily extensible to adapt to new features in always-improving technology. OpenGL, for example, has a huge interest in making sure that it's as power as possible so that it sees use. If you want to though, make your own standardization layer if OpenGL doesn't completely meet your needs, that's fine, as long as there is a need for it. Chances are though that the existing APIs can take care of you though as an API should be as small and unintrusive as possible so that it can be as flexible and extensible as possible.

Standards doesn't mean force everyone to use software X, standards mean use a public and open communication channel so that your software can actually reach the masses and play nicely with their systems and with other programs. It's helpful to you for your project in gaining popularity through users and more developers, and it's helpful to others because they can actually get access to your software. Software freedom has no meaning if it's completely gimped somehow because it's so proprietary that it must be run in a specific set of conditions that no one has anymore. That's the way it was with many of the Unix's, and that's one of the reasons they died out.

Something non-programmers often don't realize is that a lot of linux development happens just for the fun of it. That's right: there are people who code for fun and they made most of the system you're using now.

So in many cases, a developer will purposefully duplicate another's effort, maybe because they don't like the other's code base, or attitude or chosen programming language, or maybe they think they can do better starting from the ground up, or maybe (as is most often the case for me and i think for most hobbiest programmers) they just want to know how something is done so they do it. Some times those little self-edifying projects turn in to mainstays of major distributions.

The reason this works is because there is no central body of management or regulation. One distro can use software X and another distro can use software Y and the general public gets an idea for which is better.

Don't get me wrong -- I think interoperability and standards are very important, and I would love it if I could sit down at a SuSE machine and not have to acquaint myself with all the things it does differently from Debian. However that can never happen through one body (even a consortium) enforcing a policy. The solution will be a technical one, arising organically from the efforts of programmers. No artificial regulation is needed. Only patience.

freedesktop.org project has made great strides in offering a set of standards that competing technologies can adhere to (or not) thus helping to break down the technical and usage related barriers between distros, desktop environments and so on.

Right, and that solution is to create more APIs/common languages/hubs/etc like FreeDesktop.org has helped to do, so that Linux will become more modular. This modularity will give users and develoeprs the ability to install and run software more easily. It's the lubrication between the gears that will allow the many gears of Linux to function together, so helping this by making better and better lubricants will help interoperability and thus truly help software freedom for everyone.

Of course developing software is for fun many times (I hope it always is), but that doesn't mean you should sit down and try to think of a way to make software that is completely incompatible with everything else unless there is a need to do so. If I'm going to develop a 3D game for example, I'm going to look around to find the best systems to use for doing so, and modularity and user-accessibility are huge factors to consider in doing so. Even if I were to, say, build the entire 3D engine myself because I thought I could make a better one, I'd want to make sure that it could interface with OpenGL so that others could use my engine and thus help out with improving my engine. I'd want to make sure that I was using a packaging format that allowed everyone to install my game. Right now on Linux, that can't be done except to release your game as a raw binary or source or go through the pain of trying to package it for as many distros as possible, but after the LSB Packaging API is standardized, that too will be possible. I'll be able to pick DEB, RPM, ebuild, or any of the existing formats that I want to use and use it, knowing that it'll be installable on any distro that uses the LSB standard. I will know that a) it'd be silly for a distro to not want to use such a feature to give their users access to more software, and b) users will want this access and will choose distros that supply it. This API/software is tiny, and is basically in a sense a protocol that allows access to more software packages. Think of it like installing a tiny program to allow you to install plugins to play any types of movies. Suddenly, you have access to all these different formats. I mean, you'd be stupid not to want something like that, so, my point is that these small standardization layers / APIs are just there to make things modular. You can install or uninstall them, and as far as the LSB Packaging API is concerned you'll be able to install this "layer" on old systems that came out long before the API was standardized, so fear not Ubuntu 6.06 users. =D

I don't know enough to say whether or not it can be done, but I don't think it should be done. I enjoy having all the different distros to choose from. Besides, some work better than others and all for different people (if that makes any sense)! I do, however, much prefer Ubuntu, but that's for another thread.

It's called choice of software, and no one should ever take away this choice. Standardization isn't about trying to force everyone to use a specific program, like I said, modular programming allows the following:

1) Modular programming allows programs to directly compete with one another and to play nicely and use other programs around them that they interface with, so that users and developers can have more freedom in their choice of software.

2) It chops up programs into smaller, more manageable sections so that development can be easier. It's very difficult when you have a massive blob that you're trying to diagnose a problem with. If it's too big to wrap your head around, it turns off developers.

3) It allows less duplicate effort. Instead of having to rewrite tons of the code in order to simply add on feature X, and fork the project and all sorts of crazy stuff, you can just use modular programming, like a system of plugins, to allow features to be easily added. This means the users who don't want them for whatever reason don't have to have them, and the users who do will be able to easily add them, and the entire project didn't need to get forked just because of this difference in preference/need. If some program that your program interfaces with isn't good enough for you, then sure, build a better one, build a better Xorg or a better Compiz or GTK+ library or help just to improve them, whatever floats your boat. Standards simply help give you easier access to technology.

4) Not only does it mean you don't have to duplicate effort above your program, but below it as well. You can actually use Xorg and such instead of writing your own GUI engines.

All this stuff is taken for granted. Everyone here uses HTML, CSS, FTP, ghostscript, ODF, and many other standards every single day, so there is no question about whether they are bad or good. These things don't force anyone to use them, and anyone can come out with better standards. The important part is that everyone has access to being able to use them if they want to. That's what standards mean.

[STG_Fridays]

Unifying Linux would be like ice cream makers going "Screw all of these different flavors, stop making them right now. Let us concentrate on making one flavor! One flavor that will rival the gods!!!"

Insert whatever you want in place of ice cream.

It still would be tragic.

[sicofante]

Unifying Linux would be like ice cream makers going "Screw all of these different flavors, stop making them right now. Let us concentrate on making one flavor! One flavor that will rival the gods!!!"

Insert whatever you want in place of ice cream.

It still would be tragic.
This is pure nonsense and a typical misunderstanding of the idea of "unifying Linux". If Linux was unified we would have standard libraries, standard file hierarchy structures, standard package management, etc., etc., etc., and THEN we could have lots of flavours in terms of what a distro would include or not (applications, themes, services, desktop organization, you name it). We would have different flavours of a unified platform.

Ice creams are indeed a unified platform already. You can't do whatever you want that looks like an ice cream and call it an ice cream, or you'll have the public health authorities chasing you with good legal arguments and other ice cream makers accusing you of cheating on customers. That doesn't prevent ice cream makers from producing flavours.

Many will say the kernel is the platform. Lots of others (like myself) happen to disagree and believe that a platform needs a lot more than a kernel to be called that. That's unified Linux.

[sicofante]

Something non-programmers often don't realize is that a lot of linux development happens just for the fun of it. That's right: there are people who code for fun and they made most of the system you're using now.
And there's a persistent confusion between what anyone can do in their garage and what the industry should try to accomplish.

There's nothing wrong with lots of people making experiments, researching and producing all sorts of distros. But there's something definitely wrong with a handful of companies producing an operating system they call "Linux" and that can't even share the same binaries. It's just ridiculous.

[lukjad007]

Please note: I have not read all 71 pages of this thread. I may be saying the same things that others have said. (Looking at the poll leads me to this conclusion also.)

I will now list the problems.

First off: Who?

Who is going to make the one "authentic" Linux Distro?

Secondly: What?

What are they going to install on it? Vi or emacs? Are they going to have KDE or GNOME or Fluxbox, XFCE, LXDE...? What will the file manager be? Nautilus or Rox?

How? How are we going to enforce this "one size fits all" Linux? Change the license and install M$ code of ethics? (Also, from experience, one size fits all fits no one.) Will it be user friendly or not? What is user friendly? Will there be a command prompt? Will it be based on Debian, Red Hat...?

Most importantly: WHY?

Why would we want that? I like having a lot of choice. Linux survives by being flexible, fast, and ever evolving. If we stop that, we stop Linux. It would be like taking away the only use of Linux. After a while, this entity will think that it is the sole OWNER of Linux. It will attempt to through its weight around and become a soulless corporation.


One Distro to rule them all, one Distro to find them, one Distro to bring them all, and in the darkness bind them.

Ummm... No.

[sicofante]

Please note: I have not read all 71 pages of this thread. I may be saying the same things that others have said. (Looking at the poll leads me to this conclusion also.)
Well, not reading the 71 pages is one thing, and not reading the last few posts is quite another. All your questions are pretty well answered in them.

[lukjad007]

Please note that I am responding to the OP who said:
besides Canonical being Ltd., there is one more thing that drives me crazy - why do we the hell have that many linux distros around?

Isn't is different folks developing parallely exactly the same things? This greatly decreases the productivity of the whole community. I mean sure a lot of distros use the same packets etc., but for example the PROMOTION, which linux really really need in order for OEM to ship new pc with LINUX INSIDE. [SNIP]

I was just saying that I had read the first few posts and then launched into my argument. I don't have the time read all of the posts. I read at least the first and then post my answer. I haven't the time to much more while I am on the phone and eating lunch.

My argument, against the OP is still the same. It matters not what others have said, I am not rabbit trailing. I answered the poll first and these are my reasons why.

Also, when I read something I do not start at the last page and read backward unless it is a manga comic book. (That part freaks me out so I don't read them anyway.)

[Canis familiaris]

WHy the hell do the guys who pit for a Unified Linux distro start investing thier time and effort with OpenSolaris. OpenSolaris is free and open source and is very promising as well as unified.

Linux == Choice.

[aysiu]

First off: Who?

Who is going to make the one "authentic" Linux Distro? We don't know yet. I'm betting on Ubuntu, though.

Secondly: What?

What are they going to install on it? Vi or emacs? Are they going to have KDE or GNOME or Fluxbox, XFCE, LXDE...? What will the file manager be? Nautilus or Rox? If it really is Ubuntu, I would guess Gedit, Gnome, and Nautilus.

Nothing would prevent people from installing the alternatives, but those would be the standard, as they are the default, and most non-Ubuntu Forum member computer users stick with Windows defaults, generally, and would probably stick with the Ubuntu defaults as well.

How? How are we going to enforce this "one size fits all" Linux? Change the license and install M$ code of ethics? (Also, from experience, one size fits all fits no one.) Will it be user friendly or not? What is user friendly? Will there be a command prompt? Will it be based on Debian, Red Hat...? Healthy competition. If one distro becomes the de facto standard for home users (laptop/desktop/netbook), all other consumer-targeted distros would be stupid not to base their distribution on the dominant distro and make sure to have 100% binary compatibility with that dominant distro. As I said before, no one knows what that dominant distro would be, but it very well may be Ubuntu.

Most importantly: WHY?

Why would we want that? I like having a lot of choice. Linux survives by being flexible, fast, and ever evolving. If we stop that, we stop Linux. It would be like taking away the only use of Linux. After a while, this entity will think that it is the sole OWNER of Linux. It will attempt to through its weight around and become a soulless corporation. Well, if "unification" happens the way I think it will, your choice won't be taken away. There will still be hundreds of Linux distros, but only hackers and network administrators will use most of them, and the consumer-oriented distros will all be fully compatible with the dominant distro. Think of it as Firefox and Iceweasel/Flock.

One Distro to rule them all, one Distro to find them, one Distro to bring them all, and in the darkness bind them.

Ummm... No. Um, it could happen. It wouldn't bind anything, though, since the source is all open and GPL'ed.

I'm with sifocante on this one - it's more important to have standards and binary compatibility than to have only one distro.

Of course the "one distro" to be compatible with will likely be determined through healthy competition and not any kind of supreme ruling from on high.

[estyles]

Same disclaimer as lukjad007. Sorry, but I can't read a 71 page thread to see if my comments have been made. Similarly, I understand this limits the chances that my own points will be read...

It's a nice ideal. If all of the parallel toil and tribulation could go into making one, or a handful, of superior products, then I think we'd end up being better off. Linux is different things to different people, true, but what if the distros that are the same thing were merged? At least to some degree? I know it won't happen. There are too many egos and different ways of doing things, and people who want the recognition, earning potential, or just the experience that goes with creating your own popular distro. So it won't happen. But I personally think it would be great to have, say 5 distros.

1) Most stable - emphasis on stability. Everything must work perfectly. Should support all hardware where possible, and be tested and certified. Using non-supported hardware should throw a warning, letting you know if the hardware is unsupported, partially-supported, or supported but not tested.

2) Legacy - emphasis on efficiency. Support old hardware and provide as much functionality as possible with limits on resources used. Aim for a particular target minimum spec machine - say x486, 64MB RAM, 120 MB HD. Or whatever is determined to be a min spec PC.

3) Damn Small - emphasis on size. Create the smallest GNU/Linux distro for use on flash drives or other limited space applications. This will necessarily overlap with the Legacy distro, but with the caveat that this distro doesn't have to worry as much about running quickly on old equipment - it should expect to load everything into RAM and so could include some CPU hogs if necessary and if it can fit them into the space needed.

4) User-friendly - The easiest possible linux distro. Everything should be set up by default if possible. This doesn't mean there is no terminal, but the average user shouldn't have to touch it.

5) Bleeding-edge - The newest and greatest. Newest (mostly)stable versions of everything, eye-candy, multi-media, the works.

On top of those 5 distros, you could expect to add additional components as necessary. Ideally, for most distros, several options would be included in the install DVD, and let the user pick on installation. The primary choice would be desktop environment - why should there be a different distro for each environment? Let a person choose their environment on top of the distro. Other choices in this area would be where developers could redirect their creative energies instead of making new base-level distros. All of the stuff that is different about current distros could be installable packages on top of the generic distros.

But anyway, yeah, I know it won't happen. Certainly not like that.

What we *do* need, IMHO, is more standardization of certain things, some of which are already being worked on. Standard GPL drivers, GUI configuration utilities, compatibility layers between different distros, and most importantly:

A standard package manager. In my fantasy world, the standard package manager would have a binary mode and a source code mode - the source code mode would be designed to be compatible with all GNU/Linux and *BSD OSes, whereby it resolves dependencies and builds everything from source without user interaction; IMHO this would give the best of both worlds - ease of use of a .deb package, and the portability and optimization of compiling source. Even if that fantasy isn't realized, I think it's extremely important for GNU/Linux to have a standardized package system. Using RPM's with Alien is a good start, though.

[sicofante]

Linux == Choice
This thread is much about the meaning of that simplistic statement.

[Canis familiaris]

...
Do you realise this is really unlikely?
Even Ubuntu for that matter has tons of Choices. Ubuntu, Kubuntu, Xubuntu, Edubuntu .......
Also people would resist for a "de-facto" distro.
Tell a KDE user to use GNOME and you'll know by their response.

[Canis familiaris]

Using RPM's with Alien is a good start, though.
RPM as a standard = FAIL (IMO)

[aysiu]

Do you realise this is really unlikely?
Even Ubuntu for that matter has tons of Choices. Ubuntu, Kubuntu, Xubuntu, Edubuntu .......
Also people would resist for a "de-facto" distro.
Tell a KDE user to use GNOME and you'll know by their response.
If by people you mean current Linux users, then, yes.

If, however, Ubuntu ever does go mainstream (at least as mainstream as Firefox, anyway), you'll get a whole ton of Linux users who don't really care about KDE v. Gnome or vim v. emacs. They will (as they did with Windows) stick with the defaults.

Do you know what I see when I look at my co-workers' Windows computers? Luna theme. Horizontal taskbar at the bottom. Desktop icons everywhere. Hidden inactive tray icons. Hidden file extensions. You know why? Those are the Windows XP defaults. People who are not power users generally stick with the defaults.

If Ubuntu really took off, most people would use Gnome, Nautilus, Ubuntu-compatible repositories, two task bars, and the Human theme. Only power users, programmers, and other tinkerers would care about trying out KDE or Konqueror or messing around with themes.

[Canis familiaris]

If by people you mean current Linux users, then, yes.

If, however, Ubuntu ever does go mainstream (at least as mainstream as Firefox, anyway), you'll get a whole ton of Linux users who don't really care about KDE v. Gnome or vim v. emacs. They will (as they did with Windows) stick with the defaults.

Do you know what I see when I look at my co-workers' Windows computers? Luna theme. Horizontal taskbar at the bottom. Desktop icons everywhere. Hidden inactive tray icons. Hidden file extensions. You know why? Those are the Windows XP defaults. People who are not power users generally stick with the defaults.

If Ubuntu really took off, most people would use Gnome, Nautilus, Ubuntu-compatible repositories, two task bars, and the Human theme. Only power users, programmers, and other tinkerers would care about trying out KDE or Konqueror or messing around with themes.

I agree only Power Users care about KDE or GNOME or whatever.
But the way Linux distros spread among people would create differences among the average users.

Say Hari was frustrated with Windows. Asked his friend Alice what to do. Alice uses Mandriva on her laptop and installed Mandriva in his PC. He is a very happy user now.
Another person, Charles wanted an alternative OS, he searched, went to Distrowatch, selected Ubuntu (which is at top) and ended up with Ubuntu.

Here we see two new users thrown into different domains of Linux.
Is it possible to ensure each new user ends with Ubuntu or any other distro? I think not. At least not yet.

Of course if we can ensure each new user starts his/her journey with Ubuntu or any other distro, then what you think will easily happen.

[estyles]

RPM as a standard = FAIL (IMO)

I just mean that being able to use RPM in Ubuntu using Alien is a nice start. I definitely prefer the .deb package format, for as much as I know about it...

But I think a unified package format would have to be different from either of them, though ideally it should be able to support either.

[Canis familiaris]

I just mean that being able to use RPM in Ubuntu using Alien is a nice start. I definitely prefer the .deb package format, for as much as I know about it...

But I think a unified package format would have to be different from either of them, though ideally it should be able to support either.

If Ubuntu should support RPM seamles, OK but which RPM? :D
Maybe Fedora's.

But if this is the case Fedora should also support DEBs.

[cardinals_fan]

RPM as a standard = FAIL (IMO)
Why?

[Canis familiaris]

Why?

RPM has far too many flaws to be a *standard*.
I have used RPM based distros in the past and they were much more painful than Debian Based Package Management Systems.

[cardinals_fan]

RPM has far too many flaws to be a *standard*.
I have used RPM based distros in the past and they were much more painful than Debian Based Package Management Systems.
I doubt that RPM is the problem. You probably just didn't like yum(ex), urpmi, etc.

[aysiu]

I agree only Power Users care about KDE or GNOME or whatever.
But the way Linux distros spread among people would create differences among the average users.

Say Hari was frustrated with Windows. Asked his friend Alice what to do. Alice uses Mandriva on her laptop and installed Mandriva in his PC. He is a very happy user now.
Another person, Charles wanted an alternative OS, he searched, went to Distrowatch, selected Ubuntu (which is at top) and ended up with Ubuntu.

Here we see two new users thrown into different domains of Linux.
Is it possible to ensure each new user ends with Ubuntu or any other distro? I think not. At least not yet.

Of course if we can ensure each new user starts his/her journey with Ubuntu or any other distro, then what you think will easily happen.
Well, the way I'm thinking of a particular distro becoming consumer-market-dominant changes that whole scenario of being "introduced to" a distro, as I envision dominance being achieved through successful OEM sales.

Even with the "popularity" of the Eee PC, I still have seen very few of them on the streets.

If a netbook with Ubuntu preinstalled really takes off (meaning "everyone" knows at least two or three people who have one), then Ubuntu will be in a good place to be the dominant distro. If the same thing happened with Mandriva or Fedora, it very well could be Mandriva or Fedora. It could be Xandros. Who knows?

[sicofante]

If Ubuntu should support RPM seamles, OK but which RPM?
I don't know much about RPM, but if you're implying that it's not even one only thing, it's much worse than I thought... :-k

[Canis familiaris]

I doubt that RPM is the problem. You probably just didn't like yum(ex), urpmi, etc.

Maybe. But RPM has nothing special in itself to earn the rights to be a standard.

[Canis familiaris]

I don't know much about RPM, but if you're implying that it's not even one only thing, it's much worse than I thought... :-k

Yup RPM in SUSE != RPM in Mandriva != RPM in Fedora

Note: This is not sole problem of RPM itself. Even DEBs are all not compatible
Ubuntu DEBs are not totally compatible with Debian DEBs, And Xandros is neither to them.

Welcome to the World of Linux Package Management. :)

[cardinals_fan]

Maybe. But RPM has nothing special in itself to earn the rights to be a standard.
Does deb? Does tgz? Does fpm? Does ANY format?

[Canis familiaris]

Well, the way I'm thinking of a particular distro becoming consumer-market-dominant changes that whole scenario of being "introduced to" a distro, as I envision dominance being achieved through successful OEM sales.

Even with the "popularity" of the Eee PC, I still have seen very few of them on the streets.

If a netbook with Ubuntu preinstalled really takes off (meaning "everyone" knows at least two or three people who have one), then Ubuntu will be in a good place to be the dominant distro. If the same thing happened with Mandriva or Fedora, it very well could be Mandriva or Fedora. It could be Xandros. Who knows?

Yes most non-tech savvy users buy preloaded PCs, so this is the best way for more people to get introduced to a particular distro.

Unfortunately Dell Preloaded Linux (Ubuntu) != Eee PC (Xandros)

Recurring Problem this. :(

[Canis familiaris]

Does deb? Does tgz? Does fpm? Does ANY format?

No. I do not support a single Package Management system in all distros.

I propose consistent commands (which would call the appropriate package manager commands) and consistent set of GUIs which would manage software in a distro without the user even knowing which package manager the distro uses unless he/she wants to know thatis.

[insane_alien]

so to sum up, problems with one linux is

1/ you are never going to convince EVERY linux distributer to work on the same project, differnt opinions, different methods etc. etc.

2/ not every user wants only the one thing, so you'll end up getting customised versions for the major groups, people will make customized versions of those and eventially we will be back where we started.

3/ linux is already one project. the linux kernel. different distros are just the kernel with some preinstaled apps on top of it, most with patched kernels to suit their own needs.

Trying to contain linux to such a limited situation will only result in it branching and getting back to the way it is now. ubuntu is one of the most widely used distros and is the closest thing we are likely to see to a 'unified linux'. ever.

the best way to describe linux is as a gas, it will expand to fill every available niche and any attempts to capture it all and squash it down will only result in it leaking and going back the way it was.

[Canis familiaris]

I will again state that a free, open source unified *nix and even more closer to Unix based OS which seems promising is OpenSolaris.
I hope OpenSolaris gets more things working its way.

[lukjad007]

so to sum up, problems with one linux is

1/ you are never going to convince every linux distributer to work on the same project, differnt opinions, different methods etc. Etc.

2/ not every user wants only the one thing, so you'll end up getting customised versions for the major groups, people will make customized versions of those and eventially we will be back where we started.

3/ linux is already one project. The linux kernel. Different distros are just the kernel with some preinstaled apps on top of it, most with patched kernels to suit their own needs.

Trying to contain linux to such a limited situation will only result in it branching and getting back to the way it is now. Ubuntu is one of the most widely used distros and is the closest thing we are likely to see to a 'unified linux'. Ever.

The best way to describe linux is as a gas, it will expand to fill every available niche and any attempts to capture it all and squash it down will only result in it leaking and going back the way it was.
+1

[patrickaupperle]

I hate the idea. I really like yum, apt, and pacman. I want them all to develop. I like the different bases red hat, debian, and arch; I want them all to develop. I have no problem with them exchanging data more often and getting a little closer, but I want them to develop separate distros with separate goals and separate ideas as to what should go where.

[STG_Fridays]

This is pure nonsense and a typical misunderstanding of the idea of "unifying Linux". If Linux was unified we would have standard libraries, standard file hierarchy structures, standard package management, etc., etc., etc., and THEN we could have lots of flavours in terms of what a distro would include or not (applications, themes, services, desktop organization, you name it). We would have different flavours of a unified platform.

Ice creams are indeed a unified platform already. You can't do whatever you want that looks like an ice cream and call it an ice cream, or you'll have the public health authorities chasing you with good legal arguments and other ice cream makers accusing you of cheating on customers. That doesn't prevent ice cream makers from producing flavours.

Many will say the kernel is the platform. Lots of others (like myself) happen to disagree and believe that a platform needs a lot more than a kernel to be called that. That's unified Linux.


You make a good point, I'll give you that, but I would like to ask you a question in return.....

What about dippin' dots? :lolflag:

[PrimaryMaster]

I wish there was only one linux distro... be it ubuntu... i returned to linux thanks to ubuntu... i now can read and fix most of my problems.. i wish there was only one distro coz then linux could really beat the hell out of mswincrud!

[cardinals_fan]

I wish there was only one linux distro... be it ubuntu... i returned to linux thanks to ubuntu... i now can read and fix most of my problems.. i wish there was only one distro coz then linux could really beat the hell out of mswincrud!
No, you really don't. That one distro would be Slackware, and there would be no worthless tools for graphical system configuration.

Wait, you mean not everyone agrees with me in every way?

[Yfrwlf]

so to sum up, problems with one linux is

1/ you are never going to convince EVERY linux distributer to work on the same project, differnt opinions, different methods etc. etc.

2/ not every user wants only the one thing, so you'll end up getting customised versions for the major groups, people will make customized versions of those and eventially we will be back where we started.

3/ linux is already one project. the linux kernel. different distros are just the kernel with some preinstaled apps on top of it, most with patched kernels to suit their own needs.

Trying to contain linux to such a limited situation will only result in it branching and getting back to the way it is now. ubuntu is one of the most widely used distros and is the closest thing we are likely to see to a 'unified linux'. ever.

the best way to describe linux is as a gas, it will expand to fill every available niche and any attempts to capture it all and squash it down will only result in it leaking and going back the way it was.

Linux needs the ability to allow users to do certain things if they want to do them. Some system allowing hard drives to be mounted permanently when a user buys a new one and puts it in their computer is a legitimate need and there needs to be a mechanism to make that process easy for Linux users. Period.

Linux needs several other features as well, mainly a cross-distro packaging solution so developers and users are finally very free to share programs and will never be forced to compile, but that's another topic.

[quirks]

I am referring to the very first post (should there be a unified Linux?):

The answer is "No". I love to have some choice as to what OS I use.

I voted for "Yes" though, because Linux is more than just an OS: it is a huge collection of applications, like OpenOffice.org, GIMP, Firefox, Evolution, etc. And what I would really like to see is LESS FORKING of these apps for FEW BUT BETTER APPS. Let me explain: I used about three burning programs in the last few years: GnomeBaker, Nautilus CD Burner and the relatively new Brasero. All of them work reasonably well, but they are all still lacking essential features. If there weren't that many distinct projects, but developers worked on only very few ones together (as it is with Gnome and KDE), the result would be a much better app. But sadly, it is very common that people just dump a project and start from scratch or create a fork for minor reasons. And this is one of the main things about Linux, which drives me mad. :mad:

Everyone who takes part in the poll should ask himself: "Do I want a great variety of software of which none really works, or do want only few choices but therefore elaborate tools?"

Unfortunately, nobody gives a sh*t about my opinion, but still thanks for bringing up this topic!

[neighborlee]

I am referring to the very first post (should there be a unified Linux?):

The answer is "No". I love to have some choice as to what OS I use.

I voted for "Yes" though, because Linux is more than just an OS: it is a huge collection of applications, like OpenOffice.org, GIMP, Firefox, Evolution, etc. And what I would really like to see is LESS FORKING of these apps for FEW BUT BETTER APPS. Let me explain: I used about three burning programs in the last few years: GnomeBaker, Nautilus CD Burner and the relatively new Brasero. All of them work reasonably well, but they are all still lacking essential features. If there weren't that many distinct projects, but developers worked on only very few ones together (as it is with Gnome and KDE), the result would be a much better app. But sadly, it is very common that people just dump a project and start from scratch or create a fork for minor reasons. And this is one of the main things about Linux, which drives me mad. :mad:

Everyone who takes part in the poll should ask himself: "Do I want a great variety of software of which none really works, or do want only few choices but therefore elaborate tools?"

Unfortunately, nobody gives a sh*t about my opinion, but still thanks for bringing up this topic!

I totally agree with this sentiment.

Linux , if its to succede ( and dont we all want that for our friends and others we dont even know yet? ) on the desktop and be sold on computers and used by not just us geeks ( yes I am one too ) but by average peopole with average incomes some of them, then that Operating system needs to be as good as it can, and Ive seen far too many 'iffy' releases from all major distros , for this not to be a obvious issue with Linux.

I love foss just as much as anyone of the rest of you, and while I am sure some of you will SCREAM about 'choice must prevail' , you need to consider that choice doesn't necessarily need to be about the core OS itself, but how we theme it and use it . What is so hard about the 'greater' linux community as a whole , getting together like the cool mature lot we all are , and deciding that a ONE base that is solid and competent is good enough for all of us, and let those that find that not good enough, - customize their own for their own needs and forget forking. If windows has managed to do this ( and god knows aren't there geeks in the windows world too and demand customization, and somehow find windows enough ? ) why cant the foss world do it ? ;)

I truly think alot of the magic of windows for better or worse depending on your viewpoint of that, is the huge base of users they have that collectively like a hive mind, relay 'bug' reports en masse back to headquarters instead of separate armies doing so with 500 different channels, and the ensuing mess you can imagine that being. I can say with full certainly that I've seen far too often sadly the usual, ' well during testing we never had bug reports about that and so it never made it to the updates ', and that means with a smaller userbase resulting from a hugely fragemented core , that such things happen much easier because in that pool of users you have of course a resulting much smaller base of represented hardware.

Choice isn't always better you know, sometimes it just confuses &/or makes more work of the issue at hand ;) ( or so I hear: http://www.youtube.com/watch?v=Iy8R5TZNV1A )

cheers to all
nl

[lisati]

Different distros suit different people. I'm sticking with Ubuntu because (a) I'm kinda used to it, (b) I've managed to get it to work on my different machines a lot more easily than some other distros, (c) these forums are great, and (d) I probably don't have the patience to lean another distro right now.

[Grant A.]

Linux distros can't be unifed because everyone knows 99% of forks of open source projects start because one group thinks "I can do it better". And with such a large user base in an OSS project, sooner or later that utopian linux would fork like crazy, due to conflicts between developers, Free vs. free, etc.

[tvtech]

needs money giving linux away for free doesn't help this lots and lots and lots of money money money money. okay so this is a little fud but seriously to get what your asking for would take a lot more money than most people are willing to put behind this. plus, that's half the idea of linux, free open and everyone. if you have the knowledge you can make it 100% compatible with YOUR hardware so..... learn have fun and enjoy

[Yfrwlf]

I am referring to the very first post (should there be a unified Linux?):

The answer is "No". I love to have some choice as to what OS I use.

I voted for "Yes" though, because Linux is more than just an OS: it is a huge collection of applications, like OpenOffice.org, GIMP, Firefox, Evolution, etc. And what I would really like to see is LESS FORKING of these apps for FEW BUT BETTER APPS. Let me explain: I used about three burning programs in the last few years: GnomeBaker, Nautilus CD Burner and the relatively new Brasero. All of them work reasonably well, but they are all still lacking essential features. If there weren't that many distinct projects, but developers worked on only very few ones together (as it is with Gnome and KDE), the result would be a much better app. But sadly, it is very common that people just dump a project and start from scratch or create a fork for minor reasons. And this is one of the main things about Linux, which drives me mad. :mad:

Everyone who takes part in the poll should ask himself: "Do I want a great variety of software of which none really works, or do want only few choices but therefore elaborate tools?"

Unfortunately, nobody gives a sh*t about my opinion, but still thanks for bringing up this topic!

Also mostly agree. While I just think that having that "way out" is important because it's all open source, I'm sure there doesn't need to be as much forking as there is and totally agree that duplicate work if it's truly duplicate work is stupid/unfortunate. Instead of forking over things like someone wants this feature, and someone else doesn't want it but wants another one instead...it's called use a freaking plugin system and have them available as options, or use modular programming, somehow in some way.

Take KDE 4 for example. Many users hated it because it came with the slab menuing system etc etc and was just a different layout than KDE 3 was. Microsoft has done some things that were good, and one of those things was to allow users on Windows XP to go back to the "classic menu" of Windows 98/95. Now, before someone tries to blast me with the argument that keeping more options in it adds to software bloat, first off hard drives are massive now days and cheap so I think performance is a bigger issue than file size bloat, but that aside, you don't have to even include it in the program. If one developer wants to have a different GUI layout, or a certain option, or whatever, you can have all of that be easily installable by using something called modularity.

Modular programs allows your code to be improved and used by other projects as well as your own. Take Empathy for example, it uses a few different libraries that can be easily adopted by other IM client programs, the "program" part of Empathy is really small because you're basically left with just the code for the GUI and not a whole lot more. Nautilus plug-ins are a great example. You can install which ones you want, they get added to your context menu, so users that don't want those features and think of them as just clutter don't have to have them, but it empowers users who do want those features.

That's how all programs should be, plain and simple. That way you don't have to fork all of Gnome, you can just help make it more modular so that users have more freedom to easily enable the features they want. Unless of course the programmers refuse to make things more modular or whatnot and you have to fork it in order to add in the feature you want, but I think, hopefully, that's a rare situation.

Any way, Linux needs to be about choice, and here I'm not using that argument for pro-forking, because the reality is that forking can give Linux users less choice by causing less features in programs, as you mentioned. Choice means modular programming to allow competition to occur directly ideally without causing duplicate effort (i.e. your software that does the same thing competes with another's system, but the surrounding software doesn't need to be forked just because of that, and both systems are still compatible with the surrounding software because it's modular and perhaps APIs and such were used or whatnot). Choice means NOT being forced to use massive software stacks, and instead being to pick and choose any software you want to. This is why I have a problem with "distro mentality". You should not be making a program for a distro, you should be making a program for Linux, and Linux should not be segregated and forked up, but should merely have different programs installed by default and be software bundles, and if a program requires certain programs/features you should be able to easily and quickly install those dependencies. All a distro should be is a collection of software packages. Once Linux users are really free to install and remove whatever software they want to, distros will be nothing more than a bundle of software that may be a good starting point so they don't have to do as much installation work after their OS is installed.

That having been said, I think Linux in some cases does an OK job of making things modular. It re-uses lots of libraries, for instance. There are just a few big sore sticking points left to clear up though, and modular/intelligent programming needs to be utilized much more. Have to remember that parts of Linux are still quite old and modular programming is a somewhat newer phenomenon. Ultimately it's up to programmers to program intelligently, so I hope the future will be brighter. :popcorn:

[utnubuuser]

If we're talking about a unified Linux, for the purpose of promoting, I'd say promote Gnu, because hey, lets face it, if you're using/promoting the graphical Gnu/Linux desktop, your promoting Gnu, not Linux.
Linux doesn't need to be promoted, it's already in charge of everything. But the beloved Gnu could use some love/praise.
Also, as a Gnu user, I'm very fond of Gnu's good friends Terminal and Bash.

For example, when promoting Windows PCs, are you promoting the graphical desktop called MS-Windows, or the underlying MS-DOS (or whatever it is)?

Linux is like the grass on the Savanna, Gnus love it and have to gEdit to survive.

One of Gnu/Linux's best features is it's diversity. I love the idea that I can go ahead and make a distro of my very own if I so choose. -- I've heard it said that that's the point of Free and Open Source software.

[Kevbert]

This has been done. One version is Ultimate (http://ultimateedition.info/ultimate-edition-19/) which uses KDE, Gnu and Xfce applications.

[flarkit]

I'm with jan on this topic. I completely understand the motivation of having a wide freedom of choice and it is a wonderful thing. This is so entrenched in the core design, that I doubt a unified distro would be realized.

However, I think that there is a pool of awesome talent in the community, who are duplicating efforts and spending energy on distro-specific features, which could be combined to enhance a smaller set of Linux versions. Perhaps my thinking may be influenced by the flavours of Windows, but it makes sense (to me) to have a more focused, unified perspective.

I'd have an Enterprise/Server distro, a Home/Entertainment distro and possibly a Productivity/Office distro. Now pull in the people from each current distro, whose work in those areas have impressed people in the community and at a design/decision level (Linus and a select group of peers) and have them co-operating to improve their areas of expertise. Each of those could still be freely customised by users, but at least the developers wouldn't be wasting time replicating effort.

[aysiu]

There isn't that much duplicating efforts, since most work is upstream and not distro-specific.

There are two things I wish Linux devs could agree on: .deb v. .rpm and QT v. GTK.

I don't think, however, such an agreement can be come to de jure. It'll have to happen de facto with one distro becoming the mainstream standard and all the others basing themselves off that distro.

It's okay to have choice and freedom, but there should also be standards.

[Changturkey]

There isn't that much duplicating efforts, since most work is upstream and not distro-specific.

There are two things I wish Linux devs could agree on: .deb v. .rpm and QT v. GTK.

I don't think, however, such an agreement can be come to de jure. It'll have to happen de facto with one distro becoming the mainstream standard and all the others basing themselves off that distro.

It's okay to have choice and freedom, but there should also be standards.

Same, I just don't like running QT apps in a GTK environment, and vice versa. I wish the two groups would make a unified toolkit, with the benefits of both.

[Yfrwlf]

It's okay to have choice and freedom, but there should also be standards.

I think that's the main thing a lot of users just don't get, and it's the cause of flame wars. Said this before but I'll say it again. ;)

"Unifying Linux" can mean two things, forcing everyone to use one piece of software (bad), and having interoperability, common frameworks, APIs, and otherwise a system of communication (good). The first is impossible and stupid except in some rare cases where you have to provide some kind of support, etc, the second means Linux users have more choice. It means programs get along with one another. If we have to have some separate GUI languages because there are different benefits or certain ones are easier to develop than others, like GTK+ and QT, fine. But, let the communication occur between the environments built on them, so that the two can be inter-operable with each other to give users more freedom. In the future, who knows, perhaps one library will die, or perhaps there will be a way to modularize things more so that each one matters less and the "cores" can be easily swapped, for example what wxwidgets is trying to do.

I want there to be more than one browser, IM client, etc, and standardization gives us more choice so you can have that. Take Empathy for example, it uses the telepathy framework which is completely separate and which any app can use, meaning there can more easily be different variations of IM programs using it, so there will be more of them, each with different layouts, each with different features. Even better would be a program that allows you to have whatever layout you want by selecting an option, that'd cut down on duplication even further.

Linux really has the potential to be, and in many ways it already is, the "best" OS, which can meet the needs of any and every user out there, but it needs more modularity/flexibility, interoperability, and programs, and more programs will come with the success of the first two things. That's the most important part of Linux in the long run, what's under the hood: X, gcc, GTK, QT, Linux kernel, basically anything to make it easier to develop apps for Linux, etc.

[rucadulu]

I like the variety.

[Ericyzfr1]

While I believe that competition is healthy, too much competition can also have negative impact and tend to slow down progress as too much resources are spent working on adapting identical features. The ideal situation may be 5 major distributions, allowing to intensify the development and integration of new features....Just my opinion.