Are there too many Linux distributions currently available? Can there be too many? This article explores the effect of the large number of distros out right now and suggests that progress could possibly be made through a consolidation.
Are there too many Linux distributions currently available? Can there be too many? This article explores the effect of the large number of distros out right now and suggests that progress could possibly be made through a consolidation.
“Divide to reign”
There are too many distros and people are lost.
Also there is no unification for software.
.ebuild, .rpm, .deb. .blablabla…
None of them are compatible. I don’t know any Linux web site to download software where the packages are compatible for all distros. Unless source code! Not good, really…
That would only be a problem if you switch distro’s every week, I guess.
I mean, why do you want to use a debian package on a gentoo system and vice versa?
You are thinking about this from your perspective. But what about the company that wants to release software but has to be put thought the hassel of supporting 4 different install systems. Not only that there is no unified way of making them so each one has to be made differently and supported differently.
“I mean, why do you want to use a debian package on a gentoo system and vice versa?”
Because the author/maintainer has only released the .deb?
There aren’t to many distro’s, but there are to many different ways to install Software.
The debian way, RPM (SuSE, Mandrake, RedHat/Fedora etc), install from source (basically all distro’s), tgz (slackware), the gentoo way etc.
What is more important is that all these different distributions learn to cooperate more. but this is difficult. Why are all the installation methods so different? Because the vision of the developers is so different.
But, there is also the LSB (Linux Standard Base), and more and more distributions are following this standard, meaning that 1 howto should suffice for most of the distro’s following the LSB.
Another improvement made in the last few years: The dependency-hell, almost every distro has the dependency-hell-problem solved with some sort of installation tool.
Creating a package (RPM, TGZ, DEB or whatever) should also be made easier for the starting linux user. Not every software package has a binary readily made for every distro. A bit more standard (or an easy way to create all) packages would be the most ideal.
Walter said:
Creating a package (RPM, TGZ, DEB or whatever) should
also be made easier for the starting linux user. Not
every software package has a binary readily made for
every distro. A bit more standard (or an easy way to
create all) packages would be the most ideal.
I cant speak for RPM and DEB, but making a Slackware package is trivially easy.
But that’s besides the point. Everything I could possibly want is already compiled for Vector (or available on linuxpackages.net; not a big diff), which is a minority distro. Considering that most needs are met out of the box I highly doubt this is a serious problem.
“Are there too many Linux distributions currently available? Can there be too many?”
From a consuer desktop user yes there is. From a computer engineer/scientist no there isn’t. There needs to be some standards set. Things are getting messy between distros. Nothing wrong with having many distros but they need to follow the same standards. This is the single biggest issue with linux today and is keeping it from going to more desktops. In many ways Linux distros are so unique in some ways that you might as well use Microsoft Windows(go ahead bite and flame me).
There can never be enough Linux distributions since most people only use the most popular or most functional. The plethora of distributions not only helps satisfy those ultra picky people but also acts as competition for other distros.
thanks to my 1 year old doberman pinscher, an interesting thought comes to mind. if i want a dog, what size do i want ? what color do i want the dog to be and what is the purpose for getting the dog ? do i want the dog to be male or female ? thanks to the open source community of dog breeders, i can easily make my choice once i’ve considered all my options. linux is the same concept for me.
With regards to the “too many ways to install software” issue above, I don’t thing it’s really an issue of too many packaging standards.
After all, deb and RPM files basically cover the vast majority of distro’s out there, yet RPM files from say FC4, often won’t work with other RPM-based distros like Novell, Mandriva, or even earlier versions of FC.
Or the fact that Ubuntu packages don’t work for Debian, and vice versa.
Hence, even if all distro’s used the same packaging format, you’d still have the same problem of packaing for so many distros anyway.
I mean, personally I think it’s just a free market, and over time, a equilibirum point will be reached, as the weaker distros die off, and the old stalwarts (eg debian, slackware, redhat, gentoo etc.) remain.
So if you let things be, maybe they’ll sort themselves out?
But then again, you could argue that the proliferation of distros is harming the adoption of Linux, and that if this situtation isn’t rectified, Linux will die off before market forces are allowed to act properly.
As much as the freedom crowd will shout this town, too much choice can be a bad thing. It results in inefficient use of resources, doubling up and redundancy of efforts, and difficulties for the enduser.
But then again, look at the American car market. Eighty years ago, how many car brands were there around? Hundreds…
Look around again – how many are there now? (hint: it’s between one and three)
[…] and that if this situtation isn’t rectified, Linux will die off before market forces are allowed to act properly.
On the contrary: Not that I want this to happen, but If Linux dies because of the “too many distros” problem, this would actually be a very good example of market forces acting properly. It all depends on whether you look at the “Linux” market or the “OS” market, the former being a part of the latter.
Linux is the same thing as “UNIX” used to be — many different flavors, each available in its own proprietary ice cream cone.
The funny thing is that there is not even *source* compatibility across the various flavors as the tools you need to build the code have to be installed in different ways on different flavors of *ix.
Linux is perhaps the geek’s greatest testament to his inability to work and cooperate with another human being.
Linux is perhaps the geek’s greatest testament to his inability to work and cooperate with another human being
I really wonder how you got modded up when cooperation between developers is one of the key strengths that made Linux possible in the first place.
I second pravda‘s opinion.
…cooperation between developers is one of the key strengths that made Linux possible…
first, the issue was “cooperation” between nerds and humans, not between nerds and nerds.
second, what “cooperation between developers” are you talking about? – they cannot agree on important things – packaging, gui…
People are not lost, if you can’t look down the list of linux distros and pick out what looks the best, you’re not ready for this consumer world we live in.
Also you can install package managers for handle the many different package formats out there. Although each format is well provided for, so one only need to use their distros format.
I think I’ve said this before, but what the hey.
The best way to challenge MS and Apple is surely to make a product that aims to do the same sort of things. I think the best way to do this is with a new distro that deliberately breaks all compatibility.
1. Throw away all unneeded lagacy, and make a fresh start to be made. If I was to do this, I would in particular get rid of the UNIX file tree.
2. Think up a new name, and sell it as <Name> ™ based on GNU/Linux technology.
3. Keep a stable structure that third parties can create software for easily.
Phil,
That’s great thinking and I completely agree with you. However the catch 22 is that there _will_ be people who think that it should be X feature instead of Y and soon enough you’ll have another fork!
To critique your idea I think it’d be best if a for-profit (*GASP*) OSS company released a distro like you mentioned above. Company’s want to spend money with other company’s and not with OSS “Projects”.
Your thoughts?
– j
That was the thinking behind the name part. While MS can talk about Windows software; Ubuntu, Suse, whoever, always talk about Linux software. That makes people think that they’ll be able to run any software with Linux in its description, which will always annoy if they find they can’t, because it isn’t packaged for their distro for example.
If an OS was sold on the strength of its own name, X, rather than just the Linux name, then software for it will be released for X, rather than for Linux as a whole. Any derived distro, Y, would not have the same name, and therefore would not have the association with software produced for X. All X should need to do is prevent other distros calling themselves “compatible with X.”
That does of course expose the greater problem, in my opinion, that horrible damage may be done to many other distributions if they become too marginalised, as software originally written for any Linux system may disappear.
“as software originally written for any Linux system may disappear”
A simple matter of porting…
I think “I ported MPlayer to my Operating System” sounds much much better than “I packaged Mplayer for my Distribution”
I must say that when I first looked at the title of the article I was intrigued. But by the time I finished reading it I thought that it was an awful article. Not only did it not provide any solutions, it didn’t even state why the problem existed.
The LSB cannot be truly taken seriously as it requires rpm support. This means that truly great indipendant distributions like debian, gentoo, ubuntu, slackware all cannot be complient unless they remove what makes them unique.
Ultimately I see that as the creators of free software (not the distributors) improve their products, then the use of distributions will start to disappear. Once a base system + Xorg + GNOME/KDE/XFCE/E etc. become good enough then nobody will care what modifications have been made by the distribution.
Anyway, shitty article.
We should spread the word about Autopackage ( http://autopackage.org/ ). You can install an Autopackage-file on every distribution !
@Phil: “2. Think up a new name, and sell it as <Name> ™ based on GNU/Linux technology.”
Ubuntu, Mandriva, SuSE, etc., they are all trying to do this, and probably they won’t collaborate (Mandriva and SuSE, for them money is more important..).
Autopackage isn’t a good way to do it. It’s dependent on C. I mean your program must compile a part of C source in order for it to be “autopackagable”. Try it with some other language. You’d need to manualy translate that C code they provide and compile into your app, if that’s even possible as that HACK is really low-level kernel shit stuff.
That’s rubbish, there are autopackages of programs written in C++ and Python too. There isn’t anything C specific about any of it.
I found the article to be an interesting read – but consolidation is a bad term to use – standardization is more appropriate. Observe:
1. “the company’s [microsoft’s] focus on a limited number of operating systems has allowed it to successfully market and develop these programs”
Microsoft only has one desktop/server OS: ‘Windows” which it customizes and then renames – there is no difference between this and linux distros.
2. “The goal of a consolidation should be to “trim the fat” and make funds and brain power available to those distributions which need them the most”.
With the exception of major companies that fund linux development most individuals that work on linux and linux apps are doing it as a hobby. No one should tell them what to do with their free time. If they choose to work on Fedora-OK, Ubuntu? OK! their own distro? sure! your time – you choose what you want to do.
3. “Severely limiting the number of distributions available could be catastrophic to Linux’s growth in the future, but if this is done reasonably the distributions which are currently the most fit for survival will only get stronger.”
Apparently this author does not frequent distrowatch. This is already happening by a darwinian (pun intended) process of natural selection. linux distros that don’t have a following get dropped by their developers. 100+ distros are abandoned already since the inception of linux as an OS (and there are probably more that are unaccounted for)
4. ” While this is the strongest argument for the preservation of the current number of distros it is also proof of the limited scope of many of the varieties.”
You really don’t need a one distro fits all mentality – are they limited some of them? yes – so what? If the linux is “StrongARM linux” – will it work on an itanium? – duh:)
5. “…. in the same way that FireFix is a threat to Internet Explorer”
Just pointing out that a spell checker needs to be used before you submit an article
6. “…the best attributes of each can be brought forward and used in the successive versions.”
Remember “united linux” ? They just had version 1.0 and the it crashed and burned
Now, some further comments:
What linux needs is not consolidation. They need a standards body (and if they already have one, they need to get off their arses and do something). If you have standards, developers can develop to those standards. For starters you need the ability to use any package with any package handler – this means that a DPKG can be read by the RPM package manager and vice versa so that things are easy to install. (dare I say a “FAT package” ?)
Additionally – there should be a clear cut way of fixing what is affectionately known as “dependency hell” Each app should be able to dynamically determine what it needs, if it is installed, and if not installed ask the user if s/he would like to have things installed for them (and of course optimized and configured).
Developers, especially the ones that do it for free, should not be told what to do (unless the willingly are part of a team of developers) and they should never lose their creativity – this is what drives linux – BUT there should be standards.
Now as I said there is no need for a “one linux distro does it all” but potential users need to have a local “market” where every active linux distro is listed, with all of its information so that users can make a comprehensive decision – something like distrowatch but more in depth (much more) – with more search options to determine architecture type – strengths and weaknesses of the distro and so on.
just my 2 cents
“For starters you need the ability to use any package with any package handler”
Do you have any ideas for how this is to be done?
Take cron. Let’s say that I want to install a crond package into my system.
I want this daemon to be handled by my init system.
Should my system figure out what scripts to generate on installation or should this FAT-package have scripts for every conceivable init systems?
yet another talk about why Linux isn’t this and that and why and what and what not.
*gah*
For some reason I find that typo funny.
I see some people are wanting standards across the board for all distros which I find unrealistic. There should be different standards for the type of mission the distro is shooting for.
Those distros wanting to be the next desktop OS should strive for more compatibility between each other. Each distro could have their own package format but must also accept other formats as well.
I would not place the same desktop OS standards on rescue CDs or Live CDs. Nor would I place those standards on enterprise server distros.
This article and the articles about linux on the desktop seem to think that all distros are trying to be the next desktop OS. But they are not. May be distros should add some labels to the names of their product so people would know what the product is built for.
For example:
Server for servers.
Desktop for desktop PCs.
Rescue for system recovery and healing.
Live for trial and mobililty.
Vintage for older hardware.
The industry could work out the best label names.
Yes, there are too many distros. I think one of Linux’s greatest strengths can also be it’s greatest weakness. Too many people off doing thier own thing and not enough of them consolidating thier efforts into one project. They would be able to accomplish much more if they get all get together and focus and one distro.
> Too many people off doing thier own thing and not
> enough of them consolidating thier efforts into one
> project. They would be able to accomplish much more if
> they get all get together and focus and one distro.
Not necessarily true. Look at the (American) space program: 30 years ago we sent people to the moon, and nowadays we can’t even get them into space. But a few people off doing their own thing are having some luck creating alternative space vehicles; reassigning these people to do shuttle work probably wouldn’t accomplish much at all.
Analogies on the internet are like comparing apples to oranges. I’ll give you credit for not using a car one though.
> Analogies on the internet are like comparing apples to oranges.
They’re both fruit, they’re both approximately the same size and shape, they both have a protective skin, they both have seeds inside and are mostly water. It’s overextending analogies which is really the problem.
But in a sense, I think this is a problem with the article, too. There’s a definite sense of wanting to “win” the desktop by having the most of a specific distribution, the same way as Microsoft does. But they don’t really allow for winning by having the most usage split between multiple distributions; (though it’s mentioned) they don’t really take into account that the desktop may not be what any particular distribution is aiming for; they say consolidation might solve the problem of package installation, without seriously weighing it against the lack of choice, the possible lack of innovation, and the elimination of valuable niche distributions, some of the major differences over using Microsoft’s products.
Windows and Linux(-distributions) are both OSes, but that doesn’t make them the same any more than apples and oranges both being fruits does.
There, I think I’ve successfully overextended that.
I rather agree with the author but the problem isn’t “consolidation”. Like mini-me mentioned above, it’s a matter of standardization and unification across these various distributions.
There must be some sort of body out there to monitor and regulate standards through the OS. It seems the main focus is on “packages”. I personally feel that all the available packages for the various linux are shot. Of all the package systems, portage is my favorite (I’m not here to argue this or promote Portage/Gentoo) but what good is a great package system like portage if it only runs on Gentoo? What if I like Slackware and Portage and want to run them? I’ve personally tried portage on Slackware (a few years back) and it wasn’t fun (because of differences in where Slack and Portage expected files to live).
As such I got aggravated with linux and have moved to FreeBSD (again I’m not starting a battle here) because of the more standardized layout. If I need to know how to edit something in my httpd.conf file, I know it’s in /usr/local/etc/ (because that’s where 3rd party conf files go!) with the rest of my third party conf files and not undecidely placed in /etc or somewhere else by a renegade package or ego-tripped developer.
Although the BSDs now have a similar problem with various products available, it is not as rampant as linux is (yet!). I wish Linux the best of luck and feel that this problem will eventually clean itself up as the more potent distros will float to the top and others will sink.
– j
Yes if you’re some kind of totalitarian who thinks operating systems should be doled out only by the os police. I’ll make a distro any way I like and anyone who doesn’t like it can kiss my freedom loving ass.
too many automobiles.. They are all different requiring different shop manuals, different mechanics and different styles. When I buy a car there is an incredible diversity of features… However I somehow cope with it. Although the basic designs are the sme, most parts are not interchangeable. Perhap the government should require they all be the same.. Perhaps Mitsibishi should buy them all out and just sell one model-fits-all cars. Is this analogous to linux?
The problem isn’t in the number of distros. However many distros there are, the only problem is duplicated effort. The real issue is compatibility, which can be helped by creating good standards.
Granted, that’s it’s own can of worms. Who makes the standards? How do you convince people to follow them?
But the idea of getting everyone together to make one distro certainly won’t work.
I was installing some autopackage things yesterday (from their own site) and:
– 2 of them were corrupt (dunno why. tried downloading again, and same md5, so I downloaded well)
– another one of them said I didn’t have something and bailed out, instead of installing it
– and another one installed normally, but when I tried to run it, one of the dependencies wasn’t present (gosh i love those missing lib xpto.o.1 errors), and so it didn’t work
Those who did work installed two menu items on my kde menu for each app (on the same place).
Ah! And i’ve seen a demo of autopackage and it’s gtk gui, but why did I get autopackage with a mix of gtk and shell gui?
Autopackage seems very interesting, but it seems to me it needs a bit of work. Or maybe I was just unlucky.
Can you please post a more precise bug report to the forums? In particular:
* What packages you tried
* What system/distro you are on
If we don’t have this information, we can’t fix anything.
thanks -mike
There is a fundamental reason for the large number of Linux distributions that people like the author just don’t get: It’s more fun to start your own distro than to work on an existing distro.
People who start a new distro have an idea for how they want their distro to work. And it actually doesn’t matter if this idea is new or not. The point is that they want to try to build a distro their way. The fun is to have complete control and do all or most of the big jobs yourself.
Using your ‘brain power’ to work on an existing distro results in work that is just not fun. And the bigger the distro, the less fun the work is. That’s because all that’s left to do is fix bugs. All the major parts of the distro are set and will not be changed much. Sure you can try to get them to change how a major feature is done, but the hassle and politics involved to get it accepted is probably not worth it. It’s easier to fork the distro and put in your ideas in your own distro.
All these little personal distros are not a waste of effort. If the developers of these small distros were to stop working on them, they would not start working on the large distros. They wouldn’t be interested. So nothing is lost by them spending their time on their own distro. It’s that or nothing.
There is, of course, a place for the large distros. They have the money to pay developers to work on the boring bits. These are the areas that nobody would choose to work on, unless they’re paid for it. These boring bits are what gives the distro polish and stability. For some customers, that’s worth paying for.
The incompatibilities between all these hundreds of distros is also not a big deal. That’s because most Linux users do use 1 of the top 5 or so distros. Any new user of Linux would be steered to one of these large, stable and polished distros. These distros have well built packaging systems to show you the packages available for your distro, and will help you install software built especially for your distro.
For the much smaller group of Linux users that use the smaller distros, usually they know what their getting into, and have chosen to deal with the hassles involved.
Very well said, you totally, “get it”, everyone should be saving your comment as the standard form letter response to the stupid question, “Why another distro?”. Thanks.
Mostly agree with you, except for one thing.
People who start a new distro maybe didn’t have an idea for how they’d like their distro to work but rather their OS.
A distribution is about QA, bughunting, maintaining packages resolving dependencies and that kind of stuff.
The creation of new distros is probably more the effect of no other distro beeing capable of supplying the intended OS. So the poor OS creator is forced to create a distribution for this OS.
Debian-Ubuntu is an example of when an OS actually can depend on a distribution.
I’m a Linux user for many years, 3 years as my main desktop. There are three things keeping Linux back:
1 – software installation
2 – 3rd party applications
3 – GUI for all aspects of the system
1 – easily solved by dropping rpms and using either Portage or Apt-Get
2 – not so easy, mostly up to 3rd parties, we could make development easier (LSB, LSB, LSB !!!!)
3 – fundamental difference between users and developers (Gnome preferred by developers, KDE prefefred by users), not so hard to solve, have a QT and GTK frontend to ALL system admin functions
1 – easily solved by dropping rpms and using either Portage or Apt-Get
Common mistake to associate rpms with Apt-Get and Portage. rpm is to dpkg while yum is to urpmi, apt-get, apt4rpm(depreciated in favor of smartpm) and Portage.
2 – not so easy, mostly up to 3rd parties, we could make development easier (LSB, LSB, LSB !!!!)
In Fedora environment, third party adapt to the policies.
3 – fundamental difference between users and developers (Gnome preferred by developers, KDE prefefred by users)
No true. DE is a matter of preference.
While I used to believe there are too many, I’m not quite sure today. A lot of distros have done an excellent job at targetting particular niche markets a lot of other distros were ignoring. Are there a lot of distros? Sure, but that’s just what happens when you have permutations of the “main” distros into each category and niche.
As the Linux cause grows many distributions are becoming better marketed and more user friendly, but each has a long way to go before that can match the marketing power of Microsoft or the appeal of OS X (which happens to be UNIX based)
Marketing power doesn’t necessarily result in a good OS.
One of the less widely recognized reasons why Linux has not yet toppled Windows, despite it many advantages, is how divided the resources available to Linux are.
Not really, the (F)(L)OSS world has a eco system just as life itself.Why is it bad when there are more Linux flavors (ideas,implementations) avaible? I don’t hear people say there are to many different cars so the automotive industry is pretty much diffuse.Analogue is life itself where you ideally have the freedom to choose whatever to wear,drive,eat,date,OS(flavor) to use within the referential fit of the law of course.
After having used a lot of Linux distros (and other OSś) i think i can say all flavors are pretty consistent within there own eco-system.The problem is some people can’t handle to much information and (mandatory) availabilty of choices in their life.
it is difficult for Linux to provide a unified front.
Why does there have to be one unified front?I hate to be pressed inside a cabinet with a label.Furthermore it’s a great strength of Linux that diversity.Some Linux distroś are specialised in rtos, some are specialised in size (Damn Small Linux),others are very good security wise (Adamantix,OpenBSD),some distros such as knoppix are the king of compact power distro on a Live CD.S you see for allmost any task/target there’s a Linux (solution).One company can go bankrupt or face a hostile takeover, a user community of more than hundreds of distros not.
the fact remains that its mass recognition and acceptance happen to be unparalleled
As MS wants us to believe:There’s only one OS.And they have faced a lot of trials based on the suspicion of monopoly practices.
the fact remains that the splintered existence of Linux distributions has definite effects on its impact
Not on the server market.
their is no sort of governing body
What about the company you bought the OS from? Ever heard of support contracts?
The problem is there’s no problem and Linux is still being developed on and improves by the day.Whereas windows only patches vulnerabillities and ships new features where nobody is waiting on and the rest (especially security) remains nearly unaffected.This has been consequent the case between ME and XP.Let’s wait and see if Longhorn can make command.
for linux users. Hundreds of distros, dozens of windows managers/desktop environments, dozens of different software installers. It would be almost impossible. It would certainly be impractical, considering minute market share for desktop linux. That is why ISPs don’t provide technical support for linux; not like they do for windows.
This is the kind of thing that most linux advocates never think about.
And it’s the same case for games. That’s why even though OSX and Linux might have desktop market share partiy you’ll see something like WoW for OSX because its a known quantity.
Doesn’t OS X also have a directX port from Microsoft? Not that I could find that on Microsoft’s website, man that’s an awful search engine. Apple’s site has nothing too; maybe I was wrong.
Someone tell me!
Imagine providing a competent example of this actually happening.
On my current ISP, I’ve run OS 9 through Tiger, Win XP Home, and Debian Sarge (unstable and stable) over half a decade (last 2 OSes for 2 years) and the only technical issues I’ve had with my ISP have been their accounting software thinking my credit card’s expired every time there’s a Federal holiday.
As far as networking problems are concerned, I’ll leave it to you to guess which OS is the problem child on wireless networking, continually borking its own connection every other month for no reason.
“This is the kind of thing that most linux advocates never think about.” Hint: if they haven’t HAD to think about it for close to 10 years, try to infer something from that.
When was the last time you called one of these companys for “support” ? They have some poor indian reading off a card to solve the most common problems. They don’t pay enough to hire people who can honestly solve even the most basic problems.
When was the last time you tried to get support for Windows from an ISP? My poor boss did it once, I work as an IT assistant/consultant/etc for him, and he unfortunately knew more about networking and PC’s than they did (until they referred him to the support line for his DSU). I believe that was Qwest too, or maybe MSN; I forget.
…yet happy.Linux gets so much press these days. Given these,I am sure Linux will sit happy on many desktops. And that’s great.This is the advertising Linux need.People will soon realize that there is an alternative.Let it come.
Linux isn’t focused enough, if everyone worked together it could beat Windows. So, I propose that the community throw its identity in the garbage and focus on one distribution so that we can beat Microsoft by becoming more like them!
I for one like the fact that there are many distros to choose from. And I know I’m not alone in this. 😀
–Jed
UNIXPunx, The revolution will not be analog.
around the same time the various Unices or BSDs decide to merge into one cohesive OS. In the meantime the world would be a much saner place if people started treating the individual distros as OSs unto themselves.
***rant***
In my opinion, what we need is not some unified distribution, but some way to make people realise that whne you go buy a suse/redhat/etc distro, you are actually buying a suse/redhat/etc OS. It’s just some grand coincidence that they are all based on the linux kernel and simmilar software.
If we then want to simplify 3rd party application development for such a varied set of platforms based on the Linux kernel and surrounding software, then we, as a community should get together and try to perhaps draw up some “standards”. Then each such OS based on the Linux kernel could sport that they do comply to standard A, B, or C that each bear some meaning relative to such things you usually see on software you buy in stores. For example:
This game requires DirectX 9.1c – Why can we not come up with such a simple statement?
We have freedesktop.org and such “standards” bodies, why not either use these as a platform for such an “experiment”. This would of course require some cooperation with the major distros that are available today, as well as some cooperation with hardware vendors and 3rd party software developers, free and non-free such developers.
There might even be such attempts underway already, and then we ought to support them all we can.
Personally i love the way anyone can put together his or her own personal distro if they feel like it. No one forces you to use it. It’s basically a game of “survival of the fittest”. Each distribution is a unique OS.
As i mentioned, these are just my thoughts on how we could alleviate some of the percieved preassure on “Linux” (rembember, linux by it self, is not an OS)
***/rant***
Nice to know that someone actually gets it.
Exactly!
Hundreds of distros, dozens of windows managers/desktop environments, dozens of different software installers. It would be almost impossible
They all use TCP don’t they?Do they all have a different tcp/ip stack?
Besides how hard is it to configure networking on most modern distro’s.For starters you could program a database who could be used by the agents and which would provide the basics of getting the box online.Few rules:most modern distros supported fresh installed only and as bonus support if the agent knows the not so common distro.I think the knowledge in generall of most helpdesks is marginal.
> Few rules:most modern distros supported fresh installed
> only and as bonus support if the agent knows the not so
> common distro.I think the knowledge in generall of most
> helpdesks is marginal.
You do realize that people usually build on top of a completely corrupted system that anyone with half a clue about computers would clean and reinstall first, do you? Yet these people don’t have the slightest clue.
Let’s take the example of connecting to the net. I was asked once by a neighbour of mine to connect her computer to the net. When I inspected the computer, I found
1) several error messages to pop up on reboot
2) many drivers not properly installed, especially not cdrom and not modem
3) three different versions of the ISP software installed
When I asked her for the driver cd of the modem (in words: “some cd that was bundled with the modem”) she handed me several cds of which one could possibly contain the driver. I figured out that it was the wrong one, but the Windows cd contained the correct one, so I went ahead. I guessed which version of the ISP software was the correct one. I tried to get it work, but in the end I gave up because of cryptic error codes that didn’t tell me the slightest bit about the problem.
My only idea how to fix it: clean the whole disk and reinstall everything. Would she do that? Of course not. She has better things to do.
What does this story tell us? You cannot expect a “fresh installation” unless it either happens fully automatic or if it never deviates from that state after installation.
So, I propose that the community throw its identity in the garbage and focus on one distribution so that we can beat Microsoft by becoming more like them!
Beating MS is just a side-effect.What’s more important i think is happy and trustworthy coding practices so people all around the world can really enjoy from day by day improved software and a well thought concept.
I doubt wether the amount of distro’s is the issue.Healthy competition is good, also amongst the distro’s.While there are a lot of distro’s to choose from there is a lot of cohesion call it consistancy within the distro’s own worlds.
And once again misses the point that Linux will always be miniscule on the desktop without some kind of desktop standard. Hell, even proprietary Unix had agreed on Motif/CDE. And nobody wants to touch a desktop standard with a 10 foot pole because of political issues.
And then you’ve got the fact that in the real world having an ideology about operating systems is like having an ideology towards a particular brand of toaster.
At this time, the best chance for Unix on the desktop is OSX. Linux is way too fragmented to ever make a real dent in the desktop market.
The largest error of this article and some of the replies is the idea that Linux is an operating system. Linux is a kernel, not an operating system. It isn’t particularily useful until until you add a lot of other stuff to it (mainly GNU pieces). Mr. Torvalds does exercise QA over the kernel but after that pretty much anything goes.
We are not going to see small details worked out between distros while large issues like BSD vs.SysV style init differ between them.
All this said I would like to see mainstream Linux to evolve closer to each other. There is still plenty of room for bleeding edge distros for hackers and enthusiasts but I’d like to see all the mainstream distros work out their differences.
For the very reason cited; it’s to hard to target the spectrum of package mangers.
Maybe the FHS standard can be extended in the direction of a packaging standard. Just minimum data requirements. XML, anyone? Then the distros can just align themselves to that.
I have been a fully time Linux user since 1994, back when I used to download my Linux from usenet messages.
I think that the only way something like this will happen is if Linus steps up and says “we should concentrate on x” where x = package format, desktop envrionment, etc.
I DO believe that eventually there will be concensus around these things.
Linux acutally has a great idea for a standards body that created specs for a common Linux that guarenteed if developers coded for that standard then their programs would run on any distro that was compatible with that standard. This is what is really needed the same idea as that but stronger. Given a standard like that there is no reason that any program needs to worry about which distro their program is running on, they just need to ask “does your distro support the standard”. The problem came in when commerical companeis like Oracle etc were too whimpy to attempt support like they should have so they teamed up with other commerical Linux companies like Redhat, SuSE etc. and ignored the “standard”. If companies like Oracle had just stuck to the standards and not code their apps for particular distros we would be further along than we are today.
The only other way for any sort of consolidation to occur is if a particular free distros becomes the “defacto” standard and has a large majority of desktop users using it. If ubantu say or fedora become THE DESKTOP distro of choice then eventually they could push some standards. Right now things are equally divided according to the amount of people using the popluar distros, until one becomes more popular than the others this will not probably happen.
BTW, all of these issues really only affect desktops…Linux will still grow to be one of the most popular server OS’s (maybe tied with windows in the future) in the world no matter if people are still using both KDE and GNOME, or some use RPM, some don’t or not.
I think that the only way something like this will happen is if Linus steps up and says “we should concentrate on x” where x = package format, desktop envrionment, etc.
He’s probably the only one that has enough pull to get something like that done. He always says that all the interesting stuff is happening on the desktop anyway. And didn’t OSDL have some kind of desktop initiative at one time?
Yes, feed the trolls. Linux is ready for the desktop and choice is good. Package management is different on Linux, packages don’t have to be compatible across distros because we have apt/yum/apt4rpm/urpmi/portage all resolve dependancies and all have GUIs that sit on top of each system. And for those who have to have the latest and greatest on their specific distro, there is autopackage. No we don’t have .exe files that we can download from download.com or tucows but that is a good thing.
I think Linux doesn’t have to change that much and new users don’t have to learn a lot anymore, what has to change is the people who write these articles have to become more informed about what they are talking about. They are spreading missinformation that hurts Gnu/Linux much like M$’s Get the Fiction.
No we don’t have .exe files that we can download from download.com or tucows but that is a good thing.
Are you creazy? Download.com, and C|Net in general are one of the best resources on the Net.
Wealth of easy-to-install software is what makes Windows so popular… And you don’t want that for Linux? You’re creazy. No one wants to compile software, only you.
Download.com, and C|Net in general are one of the best resources on the Net
I haven’t looked in a while: do they finally have an offline cache and offline search capability or do they still rely on retro technology like online searching through web forms?
No one wants to compile software, only you.
As a rule of thumb, only developers, packagers or bleeding egde “alpha is way to old” power users compile software.
Everyone else uses precompiled stuff.
Separately downloading .exe installers and running them is a maintenance nightmare!. Let me give you an example.
I run Debian @home, W2K and Debian @work. Consider this 3 installations of an OS. On all of them I have Firefox, Thunderbird, Gaim, Gimp, Vim (only picking that SW, which is present on all). If I want to update my Debian machines, I only need to run two commands to update all the software installed, including the 5 packages mentioned previously. On Windows, I need to find 5 files on the web, download them, unzip some of them, start 5 executables, 5x click through some dialog windows.
Now tell me, which of the two ways is faster and more convenient. Hint: if you don’t want to type “aptitude update” and “aptitude upgrade”, you can start Synaptic from menu and click a small number of times. I guess it would take max. 5 clicks. That’s much less clicking than the W2K way. Hint 2: It stays two commands/five clicks on Debian for any number of packages. For W2K, the number of clicks grows linearly with the number of packages. Same applies for newly installed software. So much for “easy-to-install software“.
Standalone installers are remains of an off-line world, where distributing files on CD was the primary method.
If somebody wants to bring commercial SW into the discussion, consider this: Unless you specifically ask, you are expected to download the Oracle database from the web. And though Oracle ignores the system-specific package database, it has it’s own. Not an ideal solution, but it shows that package databases are used by commercial software as well.
P.S. Guess which of the installations mentioned do already have the latest patched Firefox, Thunderbird and Gaim and which do not?
Linux is 10000 savages, each building his or her own crappy mud hut.
It will never compete with 10000 savages building one giant mud mansion.
Perhaps not, but I’d wager at least some of those “crappy mud huts” will still be around long after the the “giant mud mansion” is nothing more than a fading memory.
You do realize that people usually build on top of a completely corrupted system that anyone with half a clue about computers would clean and reinstall first, do you?
That’s exactly why i said:> |Few rules:most modern distros supported fresh installed only and as bonus support if the agent knows the not so common distro.I think the knowledge in generall of most helpdesks is marginal.|
You have to have a working policy in order to give commercially support.
Phone support has to be as quick as possible.A lot of users have through (incompetent) usage often the strangest system inconsistencies which can produce dito error messages.Sometimes even after an fresh (default) install those error messages you described can occur.At least you can then eliminate the OS/software parameters.Great change it’s the modem itself is scr..
Sometimes if you are somewhat lucky all you have to do is alter the modem string.
ISP software
Well if you have the number to call and you know the primairy (secundairy) dnd,default gateway and your ip-address (if a static one is allowed) you don’t necessarily need the ISP software in order to go online other than for additional software such as Compuserve.
What does this story tell us? You cannot expect a “fresh installation” unless it either happens fully automatic or if it never deviates from that state after installation.
That’s true.You never know for sure if the customer on the line has in fact a freshly installed system.You still have to give support and solve the issue no doubt about it.Have done it for multiple projects.Although it was easier to go along with Borland customers and the Compuserve clientele where poor clueless the job sucks.
The whole point, I thought, of the GPL was to ensure that software/Operating systems could be improved/altered at any stage.
If somebody likes distro BAR-Linux, but wants Foo adding to it, thats cool. It can be built, and even distributed as BAR(foo)-Linux.
By saying “one Linux to rule them all”, surely this defeats the purpose?
Yes, it takes more understanding to support a variety of flavours of Linux than it does to support a single distro. Everything has its swings’n’roundabouts.. Not always “better” and “worse”. Just different.
Diversity is a blessing and should be celebrated.
Last I checked this isn’t a race. And winning the desktop isn’t going to ensure we all go to OS heaven. So what’s the point of all this prosylytizing? Will I recommend Linux to friends if they’re fed up with their OS? Certainly. Am I going to fanatically harangue them and the OSS community to join my crusade and convert en masse? Most certainly not.
Here’s the point:
– Choice is good.
– New distros/programs/projects crop up because there’s always a Better Way ™ to solve a problem.
– Package management is so advanced and integrated on most Linux-based OSes that it’s a rare event a user will find a program they can’t install natively. This makes the “interoperability” whining nearly a non-issue. And when they eventually do find said package, what’s the difference between a Windows user realizing the program they want only runs on a Mac? Or XP? Or NT for Alpha?
– People should use what they want and the best of breed (or most hyped) will eventually become most popular.
I’m not dismissing a standard base or interoperability, but one can’t force a community to dramatically change direction, especially not a community so opposed to monolithic entities as the OSS crowd.
For the most part, I think those that recommend seeing distros as independent OSes are correct. Nobody expects the same configuration options on XP, NT, ME, Win98, etc. Nobody expects Maytag’s parts to be interchangeable with Sears parts. And certainly nobody expects Maytag washing machine parts to be interchangeable with their dryer parts.
</ramble>
Not saying that a One True Linux is the solution but you forgot the bad things.
The bad:
Not beeing able to sync my phone with my computer because the vendors “Install-CD” assumes PC == Win32
Trying to figure out what takes the most time. Install and confugure samba OR convincing my friends to install and use WinSCP
Trying to convince my friends that UT2k4 and ET is all the games they’ll ever need (don’t bother responding, I know there’s more)
I didn’t “forget the bad things”. If those items are important to you, then you should get a Windows box or learn more about Samba or write a letter to your favorite game company, etc. I thought I just said “use what you like”. What is so hard about this?
Your friends use Windows and refuse to switch. You want to play graphically intensive online games and swap files with them on their native filesystem. What do you do?
Hint: Nobody’s forcing you to use Linux.
Seriously though, people are working very hard to overcome the issues you discussed. I’m not minimizing their importance. However, One True Linux will not solve these issues any quicker than several Linux-based OSes can. These issues aren’t really related to this conversation, however pertinent they are to the adoption of Linux in general. Everything you pointed out relates to the chicken-and-egg problem regarding widespread adoption and support of Linux by other entities.
For the time being, since you are so dependent on vendor support and Windows tools, I suggest you use Windows.
P.S. That wasn’t meant to be condescending or dismissive, but I don’t think you understood my post very well. No OS can win all battles.
I’m just saying that a bigger cut of the platform market is in our interest, for the single reason that more mind share will make some things easier.
I agree whole hartedly that the cost can be too high though…
More mindshare always makes things easier
If we had it, we wouldn’t have any of the issues you speak of. Linux is growing, however, and soon manufacturers and users will have to face the music and expand to *our* platform if they wish to increase their revenue and user base.
I see no benefit for the people who would do the work. What do all of these desktop/install/UI/blah article authors suggest as payment to the developers to give up their freedom to innovate and slavishly work on yet another mundane word processor or email program for grandma? I see none of the writers offering to donate the blood, sweat, skills, time, or money required to make yet another boring Windows replacement. The articles all mention “them” doing the work, never “us.”
Ishmal
Even if I think it is a disheartening state of affair that so much of the Linux resources are spent re-packaging the same software base, the fragmentation of linux distros has a more immediate effect on Linux’s mindshare: it is invisible !
Most linux companies (talking about commercial distros here) have very little money to spend on advertising the many merits of Linux and trying to attract OEMs to Linux. MS and Apple have loads of money and are very aware that people don’t perceive well the technical advantages of any OS. So they communicate massively.
Now you’d think that the Linspire, Xandros, Mandriva, and other RedHat and Novell would pool resources to ensure a long term promotion of the platform to counterbalance their individual insignificance in terms of communication. But no. There is NO communication on Linux as a product and a platform. The result is that the general public has no problem with fragmented package management and the variety of distros because they have no awareness that Linux even exists!
Then it is compounded by the fact that all linux distros use different solution to the same problems. There are good technical reasons for that but distro makers have to wake-up : these reasons don’t transpire beyond development mailing lists. So the only result is a waste of resources and a confusing picture for the would be adopter. And confusion = time (to figure it out) + risk (of getting it wrong) for most people. Not a very attractive proposition.
But asking distros to team up to come up with a common platform is asking them to agree to something that would make them mostly irrelevant beyond providing support. That’s wishful thinking a best.
So let’s hope Novell (or anyone else with deep pockets, maybe not MS) decides it’s show time and buys out Mandriva, Xandros, Linspire and TurboLinux and why not, RedHat and imposes itself as the de-facto standard Linux and starts rationalising and offering tools for third party to target that distro in terms of packages, drivers and certifications.
Your “solution” is not only unlikely, it’s not going to solve anything. The nature of OSS is that it can’t be controlled or dominated by a single entity. Novell or Microsoft can buy out whomsoever they desire, but that won’t buy them a stitch more power in the OSS world. People will still do their own thing if met with resistence to new/better ideas.
Don’t cry. This is actually good.
A little analogy:
Bacteria sit below most people’s radar. They may be small, but they’re powerful, efficient and highly adaptable. This is why they have staying power and have remained successful for eons.
Dinosaurs and megamammalia (word?) were huge. They were slow to adapt and heavily dependent on current conditions and infrastructure for their survival. When the environment changed, they were wiped out; victims of their own temporary success.
Dominating the marketplace need not be the goal. Staying power is what proves the effectiveness of a platform (or anything of worth, to be honest). I like my OSS nimble. So what if Grandma hasn’t heard of it?
Like it or not, if any, commercial or not, distro manages to emmerge as the BIGONE (didn’t say it will happen), it will be the de facto standard. Most open source projects don’t like packaging, let alone for 5 distros.
But as you say, since it is open source, other distros will still be around and be used as before. But whenever someone wants exposure for his app in the linux world, it will be a lot more simple : just do a package for the BIGONE distro.
Right now, there are plenty of apps on gnomefiles and kde-apps that I can’t install, and I run debian unstable.
And for as much as I care for Linux survival, I’d like it to evolve towards something rather more complex and full featured than bacteria 😉
Perhaps I should have used rats, instead. The problem is that anything that’s extremely adaptable (biologically speaking) is innately hard to get rid of. We tend to classify these as pests. So my choices were viruses, bacteria, rats, etc… not exactly a dream list
With regards to GnomeFiles and KDE-Apps, I do understand the frustration, but I see those sites more as window-shopping than buying. When I see an app I like, I fire up the ol’ package manager and install it natively. Repositories exist for damn near anything available in the online universe. However, it would be nice to have an integrated function where, when you right-click on a link in your browser, a context menu pops up with an entry to “Install link as package”. This sends the link string to your package manager, which searches for the best match and automatically begins the install. Again, all native. One less step. Wouldn’t that be heaven?
As far as de facto standards go, I wholeheartedly agree. But that’s, again, not something that can be forced. Whichever distro has the broadest appeal and best capitalization will become, as you say, the BIGONE. The ecosystem of distros will still exist, with each one filling its own specific niche or dying of competition. Isn’t biology fun!?!
However, it would be nice to have an integrated function where, when you right-click on a link in your browser, a context menu pops up with an entry to “Install link as package”.
Beats me that it doesn’t exist yet except for Linspire.
What drives me to these considerations is the realisation that a lot of opensource software is easier to get and better integrated for windows and OSX than for Linux : mplayer, vlc, vnc viewer with nice configuration guis, mozilla(s), rssowl, limewire, win2vnc
a lot of opensource software is easier to get and better integrated for windows and OSX than for Linux
Besides possible Windows or OS X only open source software, I highly doubt that.
Practically all software for those platforms has to be downloaded somewhere and installed separately while you usually have them in your distributors repositories where there are – assuming a proper GUI front end – just one click away, including dowloading and installing.
But I guess if one is used to having to do all by your own, an automated and integrated solution like on Linux is quite a challenge
Don’t forget: it’s the specialised animals that tended to get wiped out in changing ecologies.
>Don’t forget: it’s the specialised animals that tended to get wiped out in changing ecologies.
That’s absolutely right. Linux distros, however, are incredibly diverse and address a host of different niches. Thus, if one segment is wiped out by a sudden technology shift or lack of popularity you don’t lose all of “Linux”. Similarly, due to the nature of OSS, there is no such thing as EOL for these distros. As long as there’s interest and a copy floating around on the ‘net, Linux distros and OSS will always be available.
So there really is no reason to worry about Linux distros getting “wiped out”.
Standards are what the Linux community needs not consolidation. As long as developers are following the LSB model then consumers will see standards and more simplification for the user with out sacrificing security. Though I admit it can be confusing for consumers with all the distributions out there most of which seem to be just a clone of another distro with out adding anything new which doesn’t benefit the community as a whole. Creating a new software application or distribution can be good as long as it’s inovative, not a copy.
Anyway, there are lot of people disputing whether Debian based (ie: Debian, Linspire, etc) distributions are better than RPM based (Red Hat Linux, Fedora Linux, SuSE Linux, Mandriva Linux, etc) with out considering a few factors to make their case. From what I can see at least in the entertainment industry RPM distributions mainly Red Hat and Novell distributions play a major role both for commercial software (ie: Maya, XSI, Houdini, Shake, Smoke, etc) and open source. For corporations using Linux again Red Hat and Novell are more dominant. Both Novell and Mandriva have been working towards simplifying the Linux experience for home users. Debian based developer Linspire is also targeting the home user but fails at targeting companies and has no where near the software availability for all sectors that RPM does. So these factors should be taken into consideration when developers are packaging software whether open source or commercial. My comments may upset some Debian users but again look at the facts. Developers need to look at what consumers are using prior to developing/packaging software so as to meet their needs, not just the developers or a few geeks.
Linux lacks of many important features
self contained application
resource manager
small icons embedded inside exe
and so on
mime/type is a bullshit so all the stupid Desktop Entries….
Linux lacks of many important features
Does it?
self contained application
Can be done, but isn’t as nice maintainance wise
resource manager
What kind of resource is this about?
small icons embedded inside exe
You can embed any kind of data in your executable, even movies. Not a very sensible thing to do, but nevertheless technically possible.
mime/type is a bullshit
Is it?
Try removing all MIME type entries from your web browser configuration and check how well you can browse when all you get is application/octet-stream.
Good luck.
For some people the generic home desktop serves as their primary reference point for computers. Other people, and I’m in this group, have other frames of reference for what computers are, how they are used, and how they should operate, which is more attuned to how computers are used in corporate/research/university enviroments consisting of servers, workstations, and general office desktops. Perhaps I’m too dismissive of the home user market, but I honestly couldn’t care if Linux is ever an appropriate choice grandma. From my perspective, the home user market is uninteresting, fraught with high support costs, and plauged with with diminishingly small profit margins. As far as I’m concerned, MS and Apple can duke out for the home users. On the other hand, Linux has distinct advantages in many usage scenarios (server, workstation, and desktop) within managed environments.
Now, to be fair, even if I consider most articles of this sort to be ill considered due to their being concieved principally in relation to the model of home computers, there is a grain of truth to the problem. The problem is, of course, cross distro binary compatibility, which affects workstations somewhat, but is primarily an issue for desktops. Thankfully there is a rather simple answer to this problem which can be used today without any need for standards or, god forbid, limiting the diversity of Linux distros. Even better, it is a solution which is already rather familiar to most developers as it is exactly the same solution used on OSX and Windows: apps need to provide all, or nearly all, of their dependencies.
To create a cross distro package for applications you would compile your app relative to a directory which contains nearly all of its dependencies, and then package up not just your application but most of its dependencies up to and–for reasons of maximum portability–including the likes of glibc. When the app is installed (untarred + some config script if need be) it would go somewhere like /opt/packagename/ and beneath that directory would be /bin /lib /etc /share and /doc. Yes, it’s “inelegant”, yes, it’s wasteful of system resources, yes, it has scaling problems, but, yes, it works pretty well as distribution scheme for 3rd party applications targetted to desktop sytems.
Thankfully FOSS apps can included in a distros centralized repositories, which is a much more elegant solution. The only apps which should be packed in the way I’ve described are binary only ones from ISVs. Even better, you could easily break out your self contained application folder into a handful of dpkg and rpm packages–easily automated–which could then be made available for install via apt or yum or whatever as meta package (a task package for apt). That way the components of the package would then be updatable using the distro standard package management function. For thise without a standard package management utility, a tar.gz file would suffice with some some shell script provided to check for and apply any updates to any of the included components.
Linux is one of the most poorly engineered pieces of software ever written/assembled/glued/slapped together.
Reading the various responses with all the arcane instructions about how to get Linux to work, it just gives me that sad feeling of “special olympics”.
For example, I am technical person, but it just kills me everytime I have to install software on Linux. The program architecture is so fundamentally flawed that I cannot simply download a program and run it. I have to go through some damn package manager. Who am I? The CEO of UPS? Why do I have to manage “packages”???
I suppose on the plus side, the broken architecture of Linux does provide job security for the brain-locked idiot-savants that understand the Linux maze.
And the directory names… I think are the proof of the drug influence.
There are no human friendly directory names.
In this day and age, the Linux people cannot figure out that human beings read directory names. They are supposed to be English, not a bunch of abbreviations and obscure references to guru lore.
Pointing out the many deficiences of Linux is no hard task. The fun part is when you are done you get to multiply by 10000 as that is how many incompatible distros there are.
I have no idea why Microsoft worries about Linux. Maybe just to avoid antitrust. “Linux!!! boooga boooga!!!”
Oh yeah. I’m enjoying my drugs, thank you
. (Hello Echelon employee 
For example, I am technical person, but it just kills me everytime I have to install software on Linux
If typing apt-get install program-name kills you, you are not a technical person. At least not in the IT field. Simply, I don’t believe that. I do not expect my mum to be able to type it, but definitely I expect an IT person to be able to do that.
I have to go through some damn package manager
See, that is the problem. Mine is blessed and has super-cow powers. Maybe you should get an exorcist to take the damnation off your package manager.
Why do I have to manage “packages”
How about some definition?
From WordNet (r) 2.0 :
package
n 1: a collection of things wrapped or boxed together [syn: bundle,
packet, parcel]
2: a wrapped container [syn: parcel]
3: (computer science) written programs or procedures or rules
and associated documentation pertaining to the operation
of a computer system and that are stored in read/write
memory; “the market for software is expected to expand”
[syn: software, software system, software package]
[ant: hardware]
v : put into a box; “box the gift, please” [syn: box] [ant: unbox]
A debian package definitely fits the 3rd definition. With a little abstraction, it fits the first as well. Which is synonymous with “bundle”. That one is AFAIK used for OS X software and NeXT software. I don’t hear anyone complaining about that. I claim you are barking up the wrong tree.
And the directory names… I think are the proof of the drug influence.
Maybe they did some drugs, when the first version of Unix has been developed. If you did at least some research, you would have noticed that the directory names are historical.
If you use the command-line, typing less is better. If you click around in your favourite DE, directory names are not that important. Most of the “clicking” people don’t leave their home directories and removable-media mount-points. There, you can name your directories whaterver you want.
p.s. note to self: stop replying to trolls.
Mine is blessed and has super-cow powers
Hehe, exactly!
I don’t do drugs – I am a clean boy – although I do get boozed up from time to time – I guess that’s why I am a BSD person
ok – couldn’t resist the humor! 🙂
Now back to serious issues like RPM vs PKG and the future of linux
its what keeps a species and OS evolving!!
-nX
its what keeps a species and OS evolving!!
We are all compatible with the same food (at least in fuel terms, if not in taste terms). That doesn’t make us all the same !
If you are trying to use all of them, then yes there are far too many Linux distros available. If you’ve found one you like and are using it, I don’t see why it should be bothersome that there are other choices out there.
There are also lots of different automobiles available for purchase but I don’t ever hear anyone complaining about it.
Here is a scenario:
As a new user, if I want to test Linux, which Linux Os should I start with?
I think I will get separate distro recommendations from each person on this forum! To find out which one will suit my needs, I will need to install one by one. Let’s see… after 3 months, I find out that another distro now has an application that is really cool. If I install that, I will lose a lot of other applications I used to use with the current distro I am using.
See the problem from a non-geek point of view?
Linux needs to get its act together if it wants to reach the masses!
Linux needs to get its act together if it wants to reach the masses!
Let’s pretend there is such a thing as a one OS called “Linux”. Why do you believe its goal is “to reach the masses”. Some supportive arguments, please. AFAIK the goal of entities behind Linux is to make money (commercial companies) or to make better software (hobbyists). None of them need masses to succeed. You need big customers for the former, one satisfied person (often yourself) for the latter.
As a new user, if I want to test Linux, which Linux Os should I start with?
My advice – that version, which your friends, who are willing to help you, use. There’s no use to try all of them. Just pick one (one of major ones, if you really don’t know which one) and install it, modify according to your needs and abilities.
If I install that, I will lose a lot of other applications I used to use with the current distro I am using.
Not true. Most applications are present in all distributions. You don’t have to loose your home directory (which stores your data AND settings) if you install a new distribution.
As a new user, if I want to test Linux, which Linux Os should I start with?
Well…that depends on you.
This answer would apply to most things you can purchase. “I need to type a document, what word processor should I use?” The answer in that case is the best you can afford, which is the same answer you get when you ask what Linux to use.
As far as I can tell, the more expensive a Linux, the more “out of the box” it works. There are some obvious exceptions, but I noted that the last time I went to Fry’s it seemed that the price of a Linux was in direct correlation to how easy it is to use. For example, SUSE was the most expensive (I think SUSE is the easiest Linux of all) while the cheapest one on the shelf was the hardest one to use of all I’ve tried (Slackware).
Personally I’m a nerd, and I like Linux because I get to do nerdy hacking with it. I personally use Ubuntu because I am willing to pay nothing for an easier to use Linux. I like a challenge (to a point…).
See the problem from a non-geek point of view?
The fact that almost no Linuxes come preinstalled on machines.
Regular users don’t install OSes, so the “there are too many Linuxes for me to install and try” problem doesn’t matter to them.
Linux needs to get its act together if it wants to reach the masses!
Why would Linux be in a rush to do that? The masses seem to have an OS they like.
My point is that the OS/GUI/window manager-or whatever it is called, should be taken out of the equation. The end user should only worry about finding the right applications. When we talk about Windows or Mac Os, most people know what we are referring to. But in the case of Linux one is talking about the underlying OS without the graphical element-right? The end user, therefore, has to worry about too many things before getting it to work. It almost seems that one needs a Ph.D in Linux and its distros to use it-i.e takes a lot of research. The whole process is so confusing and bewildering for the uninitiated that it becomes a major hurdle for the adoption of Linux. What I notice is that for Linux geeks, just installing and trying a new distro or release every few weeks is what gives them immense satisfaction. The very fact they try new distros so frequently leads me to believe that they are not truly happy with the one they are using.
It is time to consolidate- if Linux is to adopted by average Joe and rule the desktop!
Cheers
It is time to consolidate- if Linux is to adopted by average Joe and rule the desktop!
Again, why do you believe that the goal of Linux is to rule the desktop? I could care less whether everyone uses Linux on desktop or not. It suits me on the desktop and that’s what I care about.
It might sound arrogant, but I’d rather prefer Linux being as powerful and flexible as it is now and adopted by a few percents of desktop users than being dumbed-down to a one-size-fits-all form and adopted by a majority of the desktop users.
I agree. There are superb competing Oses for the average Joe which also have the flexibility of running applications written for Linux
Cheers
Wasn’t there once going to be some standards body, but that failed, because SCO was in it?
Quote: “…if you can’t look down the list of linux distros and pick out what looks the best, you’re not ready for this…” End quote.
The above opinion the the biggest reason why Linux will never be a major threat to Microsoft. Most people don’t care what OS runs their computer; they just want to bring it home, plug it in, and have it work out of the box for the least amount of money. And they don’t want to go to some computer show or hole-in-the-wall computer store to get it, especially when the uber-geek reseller has a snooty attitude such as the above. They want to get it at Best Buy or Sam’s Club or someplace like that.
Linux will never be a major threat to Microsoft because theres nothing stopping them from basing their products on Linux.
What you need to understand is that Linux is no End-User product it is a marketplace of components to build one.
I don’t think you understand the nature of the GPL, either. It explicitly states that all changes must be included and disseminated. It enforces openness. Microsoft would be forced to share any and all improvements they make to any code they “borrow”. So no, while there is nothing to stop them from basing their products on Linux, there’s also nothing stopping the Linux community from re-incorporating those changes into their OS.
>What you need to understand is that Linux is no End-User product it is a marketplace of components to build one.
As far as this goes, you’re finally right… kind of
Linux is, in fact, just a kernel. One component. What you meant to say was OSS. Distributions, though, are full OSes, just like Windows (actually moreso, as most include a busload of software in addition to basic operating components).
and a very good day to you sir 😉 (thoght this was a long dead thread…)
while Linux is a kernel(-source distribution) I think that it is safe to say that in the mind share meaning of the word Linux includes most OSS projects.
I use the term in all it’s meaning, gave up a long time ago trying to define it.
One way I use it the most is to compare UNIX with Linux a kind of OS standard.
With all the trolls long gone, this is the best action this thread’s seen!
Could proove to be interesting.
When I say that most distributions isn’t OS’es I really define booth words in the same meaing.
For me:
Distribution, a compilations of packages that a guaranteed to work together in a somewhat coherent way.
OS, the end product you get when picking a set of packages and integrating them further. An OS also has slightly higher QA demands on the ‘just-working’ stuff.
Then no one would buy Windows… They’d run Linux distributions with Microsoft’s development libraries ripped off someone elses computer: People would distribute this just to spite Microsoft like they do in many countries today with full Windows discs.
That’s if Microsoft managed to not lose most use of their development software. But given that the “Raymond Chen” camp has lost its hold on Microsoft ideology I doubt that will continue: Especially since it’d be really easy on a Linux distribution because there are so many free easily available libraries for them to link to!
http://www.theregister.co.uk/2005/07/15/sco_knew/
Time for Bad Linux Editorial Consolidation. I’m tired of hearing “choices are hard” or “Apple will kill Linux with their new levitating P4s”. Editorialize about something decent and well-founded, or you might find yourself consolidated.
and that is why I have switched to FreeBSD…To be honest now I don´t see a reason to run Linux anymore.
Hundreds of distros, dozens of windows managers/desktop environments, dozens of different software installers. It would be almost impossible
They all use TCP don’t they?Do they all have a different tcp/ip stack?
————————-
Have you ever worked with end users? You seem to have no idea of the kind of calls an ISP gets. There is way more to than the TCP stack.
And don’t forget that different versions of linux, have different ways of configuring basic network settings.
Those wanting a single, standard platform have many choices like FreeBSD, OpenBSD, NetBSD, Solaris etc. The reason I picked Linux was because I could mix and match to my liking. You can use a System V style distro (redhat etc) or a BSD init style one (slackware etc). You can further choose a style of package management, system config tools, and base software which suits your liking. You get the idea.
If things like the LSB are implemented successfully, commercial software would stop functioning on “non-standard” distributions like Debian, Slackware, Arch Linux etc, leaving those users out in the cold.
While I agree that standardisation to a realistic extent is important, it should not be so much that it forces one to conform to a standard which isn’t to his liking. Why not just allow Darwin to handle it all
LSB with its reliance on Red Hat’s busted RPM system will be the death of Linux.
Linux is going to end up the private OS of a few big server companies so they lower the costs of development and deployment for their databases and middleware.
Maybe Linux will also have some presence in mobile devices.
But as a mainstream OS, there are just too many braindead morons who control the development of Linux for it to have any sort of bright future.
This discussion thread has convinced me that Linux is for tinkerers and not for people who want to use the computer do useful stuff. It looks like all the fluff about Linux being ready for desktop is just that. Sorry to be blunt but that is the way the cookie crumbles.
Linux is about community. The easy-to-use-apliance is not the single goal. Linux has many meanings, kernel, community, platform, os, distribution, religion… you name it, and it is confusing for the newly initiated.
As mind share is starting to grow the Linux trade mark is getting more and more attractive to use on products adding even more to the confusion. What is really needed is a stronger trade mark. For joe user the term Linux should mean roughly the same thing as “.NET” or “HT-technology” means: “Some thingamajig that makes stuff better.”
For developers however Linux should mean our community, our culture, the vast pool or resources availible for OS development… you know that kind of stuff.
What all this means for distributions is that we don’t really need to make distributions easier for joe user to install. We need distributions to make development and innovation easier.
There are clusters of “packages” that is roughly the same in all distributions and really doesn’t need to be diffrent, these could be consolidated.
We should be very careful with regards to lock-in’s though. I see allready how Linux is getting less and less of an inovation fest and more about optimizing whats there for stability and ease of use. Just take a look at the Reiser4 fiasco, it has been around for years and I have yet to see anyone dare to really use it.
What is even more embarassing is how Apple such cool things as Uniform Type Identifiers in the mean time.
It’s not Microsoft we have to beat, it’s apple. We can’t afford to be locked in ad-hoc, defacto or any other standards. Still we are in despearate need of standards to coordinate our efforts more effectivley so not everyone has to re-invent the wheel when they feel the urge to innovate.
“It’s not Microsoft we have to beat, it’s apple.”
I think Linux and Apple are partners in this war against the MS monopoly. For many Linux geeks, Apple offers the ‘standardization/consolidation’ they are looking for. These Linux folks have seen the light and have transitioned to Apple (not only for the cool hardware but also do their geeky stuff). Now that Apple will be running on Intel, it is only time before it becomes a truly viable option for Linux and MS on X86.
For those who clamor ‘options’ there are so many different ways of doing any particular chore on Apple. So this argument does not fly. But on the other hand, it takes the ‘chore’ out of computing.
Cheers
Yes is treats for tinkerers and quite good for the rest to.
[<<@>> ~]$ glxgears
20596 frames in 5.0 seconds = 4119.200 FPS
24616 frames in 5.0 seconds = 4923.200 FPS
24633 frames in 5.0 seconds = 4926.600 FPS
24639 frames in 5.0 seconds = 4927.800 FPS
24505 frames in 5.0 seconds = 4901.000 FPS
20634 frames in 5.0 seconds = 4126.800 FPS
23881 frames in 5.0 seconds = 4776.200 FPS