Slashdot Mirror
Dell Opens Up About Desktop Linux
An anonymous reader writes "Michael Dell explains his company's Linux desktop strategy in an interview at DesktopLinux.com. He says that it's not practical for Dell (the company) to support numerous distributions due to their incompatibilities, but that he doesn't want alienate large segements of the Linux community by selecting a favorite Linux distro to standardize on (Ubuntu appears to be his favorite, at the moment, by the way.) What he'd really like to see, is for the popular Linux distros to converge on a common core platform, according to the article."
Funny, thats what most haven't-quite-switched-yet Linux users want too...
Han shot first.
he is right - Too many incompatable distros are hurting the advancement of linux in the corp marketplace. In a way having just one overweling popular distro making up 80% of the Linux marketplace would actually help with Linux's more wide acceptance.
http://www.leadmagnet.50megs.com
The guy makes sense, IMHO. After all, how many times have you seen slashdotters whining about various installers and packages, etc? As far as standardizing the core system, that's what the LSB is for, and POSIX to some extent.
C|N>K
I don't need Dell to support linux in the traditional sense. And I don't even need them to sell me a PC that doesn't have windows.
All I want from Dell is a commmittment to ship hardware for which open source drivers are available -- for them to say, for example, we need open source audio drivers or we won't ue your soundcard/integrated chipset, or your graphics chipset, or whatever. If Dell leaned on vendors, they'd give open source developers the info they need to support their products.
The not having to pay for windows thing is tricky, and I know it bugs a lot of people. I understand why. But for me the bottom line is that I just want stuff to work, and a Dell with a windows license is still a good machine at a good price, even if you don't use the license.
It would be cool if Dell could make sure that dual boot people could reinstall windows in a differently sized partition, though -- if they could make sure that you get the installation CDs or whatever else you need to do that. I haven't really been following things, but I hear that some people get machines with ghost backups of windows instead of a real install CD. That sort of thing is a problem from a practical point of view for a linux guy who wants the ability to dual boot.
Why don't they just make their own distro using something such as Linux From Scratch? THey have enough money to create an entire Linux division in their corporation.
Ryan - http://www.thecosmotron.com/
Just test your hardware on 1 or 2 distro's, make mention of that, and give people the option to buy a dell pc without an OS. Now you not only alienate large sections that don't want that linux distro you sell on your pc... but the whole group that doesn't want windows
Thats good for you. But for the other 98% of the population want a PC that is already up and running with all the apps, drivers, and configuration set. So it's the much larger market that Dell will chase to sell too. The real money is in companies that buy 20,000 identical systems with a huge service contract. Not us computer geeks that tend to build our own anyways.
http://www.leadmagnet.50megs.com
I thought part of doing business is about making the right (highly subjective) decisions at the right time. This is just another decision that they will have to make. They could have chosen to install a selected few distros depending of the type of computers. Debian on a server, or maybe strike a deal with Red Hat or Novell and install RHEL or SuSE. On the desktop, they could have installed Ubuntu or Mandriva. But hey, if they like to install something else instead, that's a good thing! It will still a start for wider support and acceptance of Linux distros. Or they could have decided to not install any Linux distros by default and maybe miss a great chance. Who knows?
For the lack of a better sig.
Ultimately, all mainstream Linux distributions could derive from the same basic base (with the exception of those which try to fit Linux in tight places, for example). There is no reason that RedHat, SuSE, Debian, et al have to have so many differences beneath user-space software. (Consider the wildly different boot-time initialization scripts in each of those distributions. Ironically, there is a modular system in place.) Consolidate the similarities and expand by extensions which do not eliminate cross “distro” compatibility. There are already efforts to this effect. This is no magic bullet for any particular problem, but it will help eliminate the throat-cutting within the community and encourage computer manufacturers like Dell to offer Linux solutions.
Join Tor today!
From the article: "People are always asking us to support Linux on the desktop, but the question is: 'Which Linux are you talking about?'"
Dell does a pretty good job of supporting different versions of Windows (at least 98, NT, ME, 2K, XP). "Support" really means "drivers that work with our hardware" -- they could easily sell Linux without providing software support. I'm sure one of the bigger Linux distros (Red Hat if nobody else) would be happy to team up with them for a co-branded/co-marketed PC.
$nice = $webHosting + $domainNames + $sslCerts
Dell is a clever guy. In this article, I think he is just giving his best response to keep the Linux guys happy. Does anyone really believe him when he says ""Microsoft has not talked to us about Linux. If they did, I wouldn't care. It's none of their business."? Sorry, I don't believe that for a second. Companies have to go through rounds of negociation with Microsoft, re who much they are going to pay, joint marketing etc. He's saying Linux never comes up in these negociations?
The fact is, Dell is the one company that could make Linux on the desktop happen, if they wanted to.
You hit it right on the head. Having Dell push hardware manufacturers to support FOSS would be a great boost for Linux. Having them try to "standardize" the Linux world would be a complete failure, and worthless to boot.
And you have heard correctly - most new systems don't come with a full install CD anymore (I buy for a medical practice). Now, you get either a "recovery" CD (most of which wipe your partitioning) or the aforementioned ghost partition (usually with an option to burn a CD backup).
It was one of the things which helped me sell Linux to the practice, when we had to buy an off-the-shelf copy of Win XP for a machine (which came with Win XP) that took an unexplained OS crap and couldn't be retored from the partition.
Using plain ol' text since 1968
After buying a Dell Inspiron 600m sometime in the summer I figured what the heck and installed Fedora Core on it. A few months down the road I've managed to get everything (including SPDIF out, TV-out, WLAN, suspend to ram) working. The only thing I haven't had a chance to play with is hot dock and undocking. If I want to either dock or undock, everything must be shut off and rebooted...anything else ends up freezing the system.
Having said that it seems perfectly Dell compatable... would just be nice if tech support would accept my linux-based diagnostic info when contacting them for tech support. I've had one harddrive completely die (replaced next day), but now I have bad sectors and htey won't help me because I'm running an unsupported OS.
Why do people keep saying this? The reason that there isn't a large price difference in a Dell system loaded with Linux vs. a Dell system loaded with Windows is that Dell uses commercial Red Hat Linux.
You will save money if you order a Dell with no OS (well, FreeDOS is usually shipped with the system) versus one shipped with XP. You just can't order every system that way.
Why hasn't anyone thought of that before?
We could call it the Linux Standards Base or something like that.
I also immediately thought of the Linux Standard Base. Unfortunately, that relies on rpm (which Ubuntu (and others) don't use by default, but which can be supported if certain packages are installed).
I don't see what would be so bad if Dell started doing what a lot of software companies do--support the biggest few (Red Hat and/or SUSE). Hobbyists will be happy knowing the hardware works with SOME distro. If Dell finds it economically feasible, they can add support for other distros (possibly even as some pay-extra-for-support). Monarch computers and others do exactly this--installation costs for various distros depend on the cost of a license & time and difficulty of install. The support for some of these is provided through the O.S. vendor. Or you can purchase extended support at a fee (which can also relate to the time and difficulty of support).
just provide the drivers... the community will deal with the rest...
Which drivers? I've seen WiFi drivers work on one distro and not another. I've seen them work in one version of say SuSE only to fail to work on upgraded versions. The kernel has a lot to do with what works. I'm all for Dell supporting the hardware but they would have to provide several versions of the same driver to make this happen. Like or not, he has a point. He might be using it as an excuse, which is another matter, but he makes a valid point. You can't bitch about Linux not being on the desktop when there are such varying varieties. I'm a huge Linux fan and have used it since about '99 or so. Yet, the Gnome/KDE wars along with the "this distro does X and this does Y" is both a great feature and a sticking point.
This question of "which distro" is a misleading one. Pick one that you think will meet the needs of your customers. Ubuntu is a nice fit for home machines and laptops. Dell already has some enterprise Linux machines out there so they could easily offer a choice of Ubuntu or Enterprise on workstations and servers. Once one distro works on a Dell machine, the likelihood is that any other distro of choice will also work. All this talk of fragmentation in the Linux distros misses the important point that open source is more about source-level compatibility than binary compatibility. As long as software can be compiled successfully on a Linux distro, it can be used.
It is also important to track the latest stable release. If Dell produced Ubuntu-configured machines, it should attempt to make sure that the version is current with the latest stable release. This would also encourage hardware manufacturers to provide Dell with Linux-supported hardware and that might in turn help increase the number of devices that have linux support. Wireless networking is a key area where support is tricky.
Cheers,
Toby Haynes
Anything I post is strictly my own thoughts and doesn't necessarily have anything to do with the opinions of IBM.
Common core platform for Linux will never happen. Developers and the core user community are too afraid of standardization - just see what's holding Linux desktop GUI back: there is no standard GUI (at least when it comes to widgets, menus, style and configuration) in the same way as on Mac. Why is there no standardized desktop? Because developers and the core user community abhor any idea of such a lockdown that limits their ability to tweak the system. Imagine a situation like this: "That's a fine application there, Mr. Developer, but its user interface doesn't conform to the distribution regulations and hence we cannot include it in the distro". It's exactly the same thing with distros.
The owls are not what they seem
I think Michael Dell's, as well as other wannabe-Linux users desires to see Linux converge into one single platform are not going to happen anytime soon. In fact, I don't think it even matters at all too.
Linux and most open-source software are by nature a federated, bottom-up form of software development where multiple versions abound. This is because there is no one single entity(person or corporation) who knows which features are best for users, *and* the best way for those to be implemented. Hence forks abound which allows users, aka the free market, to decide which versions/software suit their own requirement. Compare this to proprietary software where the corporation decides which features you want, when you want them and in what form you want them.
Waiting for Linux to converge into a single platform with a market share >80% would imply that other versions have failed to see what users desire, and one company(or group of individuals) has been able to capitalize on that and advance its market share.
Now Ubuntu(I use that myself) has to an extent been reading what lay users desire from a distro and implementing many of their needs well. But as Ubuntu becomes more popular are other distros going to sit still watching it reap all the laurels? I don't think so. They will evolve too. If you think that isn't possible then ask yourself how the hell Ubuntu managed to gain so much in the last couple of years? Do you think such innovation will stop after Ubuntu?
Finally, imho lay-users are not going to want to switch to Linux in the near-term. Because switching an OS for them represents a huge task which they will undertake only if:
1. They are thouroughly dissatisfied with Windows, or
2. They are thouroughly enamoured by the benefits that Linux offers
Unlike what we may all think, on the whole most people are not thouroughly dissatisfied with Windows. Sure they may have to deal with patchy security and those occasional crashes but hey, who says Linux doesn't have issues? I've had Ubuntu lock up on me more than a few times. I've spent a better part of the first month trying to get streaming videos to play on Firefox. Did I quit? No...so why would a Windows user?
To sum up, expecting Linux to converge into a super-distro isn't going to happen. Simply because open-source by nature is designed against the formation of monopolies. Since code upto a certain point is freely available to all, a new fork can be established by a brighter, more innovative, more responsive group in much lesser time than in the prop. s/w market. So an 80%+ distro would mean that nobody else read the market and changed course.
What's so difficult to understand about that?
"If the Linux desktops could converge at their cores, such a common platform would make it easier to support. Or, if there was a leading or highly preferred version that a majority of users would want, we'd preload it."
For a company which has been supplying $300 low end machines with scrap hardware and shady driver, this doesn't make much sense to me. Even with failed venture in Linux market with Red Hat back in 2001, I don't ever recall Dell ever putting any effort in supporting customers half way decent.
Sure, they had "support Red Hat and SuSE or United Linux" logo. And because of that, Dell's association with so called "leading or highly preferred version", it treated Linux as an OS, not a kernel. and when someone states "I support Linux" normally you don't convertly support only "some portion" of GNU/Linux distro, but work with Linux kernel developers with half way decent driver support so that EVERY distro can benefit from it.
Even today, Michael Dell either can't see it or is too naive. One would think, Dell had learned their lesson and support Linux kernel developement and community and not "leading or highly preferred version" distros. However this goes to show, Dell didn't.
"Don't let fools fool you. They are the clever ones."
Like it or not, the reason Microsoft has a foothold on the desktop market is because of its relative ease of use. A worker or home user can by taught the basics of checking their e-mail, writing documents, etc. in Windows and Office via memorization. They learned Office 95 back in the day, that training investment carries over to the latest version with just a few add-ons. If you really want to see how seriously important backward-compatible trainng is, turn on the "blue screen, white text" feature in Word as well as the WordPerfect compatible function key layouts. Or the "slash" menu in Excel for hardcore Lotus 1-2-3 users. Microsoft knows they have the lock because of this. Mac OS X, for example, is much easier to control centrally than Windows is, but no one switches to it because their staff is used to Windows. Even if Office si a work-alike, relearning keyboard shortcuts and other tricks is time-consuming.
Companies do not want to invest money retraining their staff. It was hard enough getting them to learn MS Office or WordPerfect the first time. There are a few things that need to happen before Linux makes a big push for the corporate desktop:
-- Make it "just work." Windows' big strength is that I can go to CompUSA, buy any old crappy piece of hardware, plug it in, and have it work without having to load kernel modules, edit config files, etc.
-- Standardize it. Pick an office suite. Pick a window manager. Pick _a few_ of the hundreds of obscure GNU applications and bundle them as a standard tool set. Wrap in some administration and deployment tools that are brain-dead simple to use. No normal user wants three office suites, four window managers, etc.
-- Completely hide the guts from the end user unless they want to see it. Mac OS does a great job of this. I have the command line and access to the config files if I want it, but the GUI is more than adequate to tweak most items.
Dell's other big market is home users. The same rules apply, just more so. Home users do not have the patience to learn Linux internals. My advice would be to start with an Ubuntu-like base, and go to work making the OS just work for normal users.
If Dell starts shipping every box with some Linux distro, that distro will immediately become the "common core platform".
Paid Q&A/Research
You know you need to re-evaluate your coding-to-social life ratio when you start making jokes in code
Nothing costs nothing
If we look back a few years, we used to have a larger number of vendors that made linux-compatible modems, sound cards, video cards, wireless network cards, scanners and so on. MS has been able to lean on these enough for them to make it hard or impossible for Linux (or BSD) support. Michael Dell could help counteract this method of MS, if he so chose.
As others have posted and will post some more, it doesn't matter which distro he chooses. Once he supports one, the others can take up the slack themselves. Debian stable, being a slow-moving target and completely unencumbered, would be my first choice, but Ubuntu or Kubuntu would give a predictable release schedule.
Beta is broken and the link to classic doesn't work. Stop wasting our time or there won't be anybody left here.
Which comes first? The popular distro or the market share?
Think if Dell offered Linux to the average consumer and worked with the Vendor to provide support it'd give them the market share? Of course. Dell would make colourful foldout instructions for whatever Distro they choose. Dell would make drivers specificly for the distro they choose. Just like they did with RedHat on the server OSs (try getting OpenManage to run on other distros... hell in a handbasket).
So I'd say this again- if Dell were to pick one, that'd see a big boom in popularity and familiarity with users.
-M
when you see the word 'Linux', drink!
Your service hasn't been worth much since about 2001, so it's no big loss for the user. Then you can stop making bad excuses for not wanting to offer Linux because MSFT will find a way to raise your OEM license costs if you do.
That's our life, the big wheel of shit. - The Fat Man, Blue Tango Salvage
Consider this. Linux has support for LOADS of filesystems. Windows? Just a few. Most Windows users would immediately see the many fs options they could choose from, and have a panic attack. I sure as hell did my first time with Linux. I didn't know whether or not to use ReiserFS or what! Even with a GUI installer, most Windows users, I think, would shake their heads and go back to simplicity.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
Dell is still clearly thinking in terms of large, impenetrable operating systems. Choosing which Windows operating system to install (ME? XP? 2003 Pro? etc,) is critical. The wrong choice will sink a line of computers.
So, what Linux operating system to pick? It doesn't matter! Choose whatever distro you think you can support the best. Preinstall Weezix (distro maintained by George Jefferson's wife) for all I care. If you can show me Weezix running, drivers and all, that means that I can copy the config to my distro of choice. Yes, that takes some expertise. But there are tons of people with that sort of expertise nowadays.
And here's the kicker: within two months, step-by-step instructions will appear on the forums and wikis of the major distros. Within six months, most distros will automatically support that machine out-of-the-box.
It doesn't matter which one you choose, it only matters that you choose! Though you can make everybody's lives a lot easier by selecting hardware with open source drivers. Too bad about the graphics card situation...
Oh come off this nonsense. This is not about elimating choice, this is about practical reality. It does not make sense to try to support a product across so many distributions that are fundamentally the same operating system. All features that users are interested in exist in, quite intuitively, user space. They do not care what the init scripts are doing or what kernel they are running. Not a single user noticed Apple's transition from BSDi to FreeBSD 5 when they released Panther and that is precisely how computers should operate. Why cannot we, in the open source community, offer the same thing? The answer is: we can, we just don't want to. What users do care about is their own experience, not that of the developer. If we are ever to eliminate the mind-numbing status quo, we have to be pragmatic and offer solutions to problems without introducing a swarm of new problems.
Join Tor today!
Linux had a movement towards a single distribution--it was called UnitedLinux. It died due to lack of interest.
The problem is, when you put companies in the driving seat for a push to a single Linux distribution, you get crap like RPM being made part of the standard. Personally, I'm glad UnitedLinux failed to gain overwhelming momentum, because life's too short to have to deal with RPM.
GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
Nobody who takes more than a second to think about this is going to believe that partisans of a particular Linux distro are going to be more offended by Dell chosing the wrong distro than Dell not offering Linux support at all. For one thing Linux support for one distro would be driver support.
In fact, it's hard to believe that Dell cares what Linux partisans of various stripes think at all. While they probably do sell the odd machine to evangelists and hobbyists, those people can be counted on to customize the machine beyond the pale of supportability anyway.
No, for whatever reason, Dell doesn't see profit maximization in that direction at this time. <portentious music> We may not, probably are not privvy to all the reasons why this is so. </portentious music>
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
"If you look at DistroWatch, you'll see zillions of these distributions. Which one should we do? And, everyone keeps telling us that they want different distributions. So, our conclusion is to do them all and let the customer decide."
We love Linux, and we're doing our best to support the Linux community.
Mr. Dell,
Please put your products where your mouth is. If you are so supportive of Linux, please put your suppliers on notice that you will not buy from them without Linux drivers, please design and promote your full line of PCs as "Linux ready", provide strong customer support for 3 or 4 distros (they really aren't that different under the hood) and please, oh please, sell them at a price that doesn't include an M$ tax.
Still waiting after all these years.
Granted, the FOSS would do well to court more hardware manufacturers and vendors, but the "We won't support linux because there are too many distros" line is old, tired FUD. First step, (as many have said here) is to not use hardware that doesn't have open source drivers. Second step, pick a distro, then support it. If you expect the FOSS community to come to you and address some of your concerns, you'll have to meet them halfway. You like Ubuntu? Good! Pour some money and people into the project so you can have an easily supportable distro.
I know, I know, to start such a thing is easy, but to keep it updated is going to cost _real_ money. But they can outsource that to their folks in India, or wherever.
Browsers shouldn't have a back button!! It's all about going forward...
People who stand by commercial operating systems seem to have a hard time grasping the idea that open source projects are based on a principle known as freedom.
I'm starting to think that most people are into bondage.
Freedom is a messy, messy thing. It's not clean, or easy, or cheap. It's not a delicate giver-goddess fashion-plate saint; it's a rough player, dirty and sometimes mean. It takes dedication to maintain, intelligence to master, and a willingness to honor others' freedoms.
I don't think most people are up to the task. Or, at least, there's a large portion of society that isn't up to the task. Leading the way are corporations, as Freedom is anathema to business. Corporations work hard to limit choice, or better, to dictate choice ("You can have any color, as long as it's black"). Second is the government, as Freedom is difficult to govern. ("You can have any choice, as long as it's mine.")
And I think most people allow that to happen, because they don't want to have to excersize Freedom. (Citizens in the US are notoriously averse to excersize.) They would rather be stupid followers instead of intelligent independents.
My evidence?
McDonalds. Budweiser. Wal*Mart. MS-Windows. G. W. Bush. Ribbon magnets on vehicles. Star Wars I, II, III. Etc.
Each of those are demonstrably crappy products, yet each is a "leader," in some definition of the word. Each is patronized by more people than competing products (well, except G.W. Bush, but he's patrionized to by more people).
I don't know what this says about society, but it keeps me up at night sometimes. I hope someone figures out how to fix it.
Microsoft is to software what Budweiser is to beer.
So the GP's points are not moot whatsoever; they are spot-on.
The basic problem here is that much of modern Corporate America doesn't see their customers as anything but an ends to a means. Not a partner, not someone to serve; simply an ends to the next bonus. Dell is an excellent example of this; and if they truly wanted to serve their customers, they'd be providing what their customers wanted - not what Dell says they want.
The best way to predict the future is to create it. - Peter Drucker.
Last I checked, I still had to pay for a Dell box myself. That makes me their client; and them a vendor.
I could care less what Dell wanted. I know what I want, and Dell doesn't provide it.
The best way to predict the future is to create it. - Peter Drucker.
I couldn't care one whit for what Michael Dell does right at the moment.
They have had opportunities in the past to support Linux properly and they've discarded them for working with Microsoft- which is their right as a company. This whole sorta love it, sorta not affair with Dell has been ongoing for nearly 10 years now. I know about it because I was in the wings on parts of it all throughout.
The bulk of the argument Michael Dell's making is specious as it doesn't apply for Dell Computers as they're only really concerned about kernel support of the device buildup on a given machine- for all they care, they need only drop Debian, Ubuntu, or, god forbid, even Linspire on their product lines to "ensure that they work on delivery". If the kernel has support for the devices they ship on a given desktop and laptop, this will simply work and people can choose other distributions as they see fit for them- so long as any of their custom apps use something like the Loki Games installer or Autopackage (I'm for using Autopackage myself...).
This is all nice, but in the end, he's asking for Linux to be more like Windows (which it's not...) when he really ought to be less concerned about all of that and pick a default distribution they can comfortably support and support the devices in the Kernel however they can. It's not at all hard Michael- happens every day of the week. I've got a laptop from one of your competitors, any distribution will install on it, and the bulk of the devices (with the notable exception of the Broadcom WiFi (which there's a usable workaround, though I'd rather they didn't use that chipset...) and the silly on-board flash reader (which TI's preventing a version to be made- nifty device really, too bad TI's being stupid about it...), it all just went on and worked- with each and every distribution I put on it in 32-bit mode (64 bit modes work, but since the ATI chipset's...twitchy...it is more difficult to get 64-bit modes going. And it's nothing to do with the distributions per se, it's ATI's doing...).
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
People, when they talk about support, on slashdot, seem to think that support simply means having full driver support for all the hardware, released as open source so that it can be added to the kernel, and viola, all Linux distros are supported. While this would be a great step in the right direction, and would be sufficient for most current Linux users, it's not sufficient for Dell, who would potentially be introducing many more users to Linux, and then need to support them.
/etc/httpd, /etc/apache, /etc/apache2, etc; and when you find the right directory, some distros have 'traditional' httpd.conf file, while others [Debian as one example, if you have apache2 installed] have an httpd.conf file that is included by another file, like apache2.conf, but the other file is the 'real' conf file for that build of Apache, etc]).
The truth of the matter is, I think what Michael Dell is talking about is that, if they are going to have a "Linux Desktop", they presumably would be putting a distro on the desktop, at which point someone's gonna be ticked that they are 'choosing winners and losers'. Someone suggested offering 5 or 10 Linux distro images which customers can choose from. This is somewhat ridiculous too, because then their Dell call-center helpdesk agents have to be able to remember differences between 5 to 10 Linux distros when trying to assist customers with problems (e.g., if someone calls in with an Apache question, does this particular distro store the configuration in
This is what Michael Dell is talking about when he says all the distros need to converge on a common core. All the files and configuration for stuff like apache, samba, X, KDE or GNOME, etc, need to be exactly the same across all distros so that support people aren't kept guessing at where stuff is and how it's setup. Even Dell CANNOT probably really afford to support multiple distros across thousands of customers (maybe if they ever reached the point where they had millions of customers who'd bought Linux desktops from them, they could afford to support multiple distros, but not from the start).
Lets take an open source program, which we'll call program X. Program X's developers happily continue adding cool new features and releasing out new versions. The problem is, the only way for GNU/Linux users to get program X is 1) build it from source, or 2) hope someone has built a binary for their distribution. Option #1 is time consuming, and tracking down build dependencies is difficult and error prone. Which leaves us with option #2, the current state of GNU/Linux distributions.
/lib library naming scheme is for). But most distributions have package managers that enforce only one version per libary. Example: KDE in Debian testing/unstable breaks horribly when a new version of kdelibs comes out. Typically, half of the kde packages depend on the old version (because they haven't propegated to unstable yet or were kicked out of testing), so "apt-get update" will happily uninstall half of KDE! The package manager is enforcing an unnecessary constraint. Why can't I have multiple versions of a shared library package installed?
Binary compatibility, or lack thereof, is why the current fragmentation of desktop Linux distributions is so irritating. My #1 criteria for using a distribution is that it has best selection of software. Currently, I'm using Ubuntu because it's Debian with a commitment to a 6-month release cycle.
I've thought a lot about the distribution incompatiblity problem, because it really hurts my desktop GNU/Linux experience. It's not the Linux kernel, which has excellent binary compatibility with userspace. It all boils down to poor package and version management, especially with respect to shared libraries.
The first problem: there's no technical reason why you can't have multiple versions of the same shared library (in fact, that's what the
The other reason for distibution incompatibility is the packages themselves. There are a few different (incompatible) package formats and much worse, each distribution uses its own (incompatible) package naming scheme. Thus, I might have a package build for Fedora Core that depends on libsdl-ttf (Simple DirectMedia Layer True Type Font library), but who knows how it will be packaged or what it will be named in any given distribution?
So, it's the damned package format, package naming, and dumb handing of shared libraries that's to blame. What's the solution? Well, I think the answer is technically simple, but complex to implement. 1) package managers that are smarter about shared libraries, 2) standard package naming scheme, and 3) someone, either library developers or a standards body, needs to build "official" versions of each libary release. This would give GNU/Linux application developers a common "SDK" so they can build their application independently of their distribution.
I can tell you right now that the Debian folks aren't going to suddenly drop everything they're doing any time soon. Ubuntu might get folded back into Debian, but that's a long way away, and I wouldn't bet on it. The same thing goes for Knoppix. It's even less likely for Linspire, because it's sold by a for-profit company.
And those are just the Debian-based distros, for whom it would probably be technically easiest to merge. What about SuSE (Novell), Fedora/Red Hat, Gentoo? Do you think they will merge with each other?
News flash: If Michael Dell doesn't want to serve the *actual* market, instead of some fantasy market in his head, I'm sure his competitors will be glad!
http://outcampaign.org/
the comment about converging on a common distro seems kinda funny comin' from a guy who built his business by never selling two like motherboards.
I guess to be fair Dell's come a long way, but it still doesn't belong in the ranks of IBM, HP or even Sun.
But distros already have standardized on a unified core: Debian. Ubuntu, Knoppix, Debian itself, and many others use the same Debian core and can use each other's packages with far greater success than the RPM folks could imagine in their wildest wet dreams..
Help us build a better map!