Mark Shuttleworth, the founder of Ubuntu Linux, has explained why he thinks the DCC (Debian Common Core, but don’t let the Debian guys hear that) will fail. “The vision behind DCC, which is indeed compelling, is that it would provide a common platform for certification, and that the distros that make up the DCC would all ship exactly that same core. But it strikes me that this approach has never worked in the past. In fact, every distro ALWAYS modifies elements of the core, and with good reason. And while we would love that not to be the case, the truth is that the reasons to specialise outweigh the benefits of homogeneity.”
“Benefits of homogeneity” like important vendor/driver support, ease of use/administration, and binary applications that will run for ages and ages on the same kind of system?
What could be more important than that, unless you’re not actually interested in Linux adoption?
GNU/Linux is homogeneous over all the diferrent distibution and category of distributions.They all use the same software , kernel , basic tools , DE , WM , Xserver etc … Whats not homogeneous about that ?
You really think that because a platform exist , and its all the same that vendor/driver will pay to make there version run on any platform ? Why is there absolutely NO BSD’s support at all then , beeing the oldest and if we follow the lies of some who dont know what freedom means freest ?
The administration of a GNU/Linux system is easier then both APPLE and Microsoft , because everything is upgraded and secured , not just the base.
“and binary applications that will run for ages and ages on the same kind of system?”
Binary is the problem , it work on one single system , static and source are always better and the way to go it you whant your application to work.
“What could be more important than that”
Freedom , Right of ownership , Security , Access to all regardles of any racism or special criteria (economics) , Evolution , possibility to build on the shoulder of giants , etc …
“unless you’re not actually interested in Linux adoption?”
The kernel come with every GNU/Linux distribution.
Shuttelworth as shipped more GNU/Linux CD , DVD , computer and as done more for all of GNU/Linux adoption then the vaporware from the DCC and all the member who are in it.
GNU/Linux system and Real OEM vendor will come in time , Just as they did in every other category.
You say “The administration of a GNU/Linux system is easier then both APPLE and Microsoft , because everything is upgraded and secured , not just the base” and “Binary is the problem , it work on one single system , static and source are always better and the way to go it you whant your application to work.
Surely you jest. Administering a GNU/Linux system may be easier for an experienced admin, but for the average home user? Have you used an OS X machine for extended periods of time? Drag and drop install of applications. Drag your unwanted applications to the trash and voila it’s uninstalled. I use Linux as much as the next geek, and to say that OS X is harder to administer than Linux is misguided at best.
One of the major hurdles Linux is facing is the lack of the ability to run binaries on all systems. We’ve all experienced them, being unable to install a particular binary because it wasn’t packaged for our system. Diss Windows and OS X all you want, but with Windows, most programs work with Win9x all the way to XP in a single binary. Same goes for many programs in OS X. No, open sourcing everything is not the answer, as there are many commercial software companies who are jittery about releasing all code.
So no, as Linux is Poo mentioned, homogeinity is key if you want desktop Linux to succeed. Anything that simplifies things for the user is going to help.
Define “desktop Linux”. If you’re talking about the corporate desktop, then the things you mention are really not that important. Instead, what is important is the ability to maintain a large number of machines with minimal effort from an experienced admin. Linux excels in this regard much more so than does Windows or OS X.
Attacking the home desktop market with Linux is, at this point, a useless endeavor. The home desktop market sucks. The margins are small, the customers are demanding, and the required software base is enormous. The business desktop market, in contrast, is much less varied, with bigger support contracts and knowledgeable customers (professional IT folks). Why try to sell 100 copies of Linux to 100 different home users, all of whom need different software, and none of whom will pay for technical support, when you can sell 100 copies to a single company to install on 100 PCs all running the same software, along with a nice fat support contract?
Edited 2006-01-04 21:11
No , the major GNU/Linux problem is driver and hardware maker support , I know that instalation “Drag and drop” style of software is not what anyone would consider the only thing to do when administering a system ( Klik exist ).
Its not because you allow your vendor to sell you binary software only that the system is the problem , if it run on one GNU/Linux it can be made to work on all GNU/Linux system. Your vendor just choose to remove your ability to do so , not GNU/Linux. BTW I whant to see you install *ALL* windows 98 software on XP and *All* Mac OS 9 software on Mac OS X , yet binary compatibility there is never a problem , its the software fault in those cases , yes right …
I Agree Open Source is shit alone , because of some of the traitor license who allow closing of the source. Noble principle who cant work in reality in human society as show by 35 years of existance ( Open Soure created everything in IT ).
Linux is Poo is wrong as usual and so are you , people dont say my Debian is not blue Like Mandriva or Brown Like Ubuntu so its not working , they say it dont work because the hardware they bought whas not made for GNU/Linux and the community as not yet made driver for it. Go look at the desktop of your friend , unless its in a closed corporate environment not one is similar to the other after a couple of days of use.
Before you continue Crying and ranting and foaming at the mouth :
– Most people dont install OS
– Most people dont install software
– Most people dont administer there System
– Most people go out and buy turn-Key
For the momment GNU/Linux dont have its own Dell like turn-key vendor. For the rest people think there is no support vendor for GNU/Linux.
BTW why do you need to install software , GNU/Linux as Live-CD and DV this days ?
Proprietary drivers from nvidia
nvidia-glx
http://packages.debian.org/unstable/x11/nvidia-glx
and from ati
fglrx-driver
http://packages.debian.org/unstable/x11/fglrx-driver
If you don’t want them, you can use opensource version.
In case of radeon, driver version from xorg 6.9 really double the speed (AGP 8X instead of 4x).
Try Chromium at 1280×960 or Blender.
Because you really think that they are up to what windows get ?
go read :
http://www.phoronix.com/scan.php?page=article&item=357&num=1
http://www.phoronix.com/scan.php?page=article&item=359&num=1
Also there is more to computer system hardware then just 2 type of graphic card.
BTW I whant to see you install *ALL* windows 98 software on XP and *All* Mac OS 9 software on Mac OS X , yet binary compatibility there is never a problem , its the software fault in those cases , yes right …
Before I say anything, dude you need to chill out. We’re only talking about operating systems.
You do realize that OS 9 and OS X are practically different operating systems? Nevertheless, on OS X there is Classic mode that allows you to run most OS 9 software. Most software that will work on Windows 9x will work on XP. Most software that will work on OS X.1 will with X.4. Most Linux binaries don’t even work across versions of the particular distro, much less cross distro. This is something that desperately needs to be worked on, if you want Linux to work well as a home users desktop.
Most people dont install OS
Agreed, and it is unnecessary. That isn’t something that the average computer user is expected to do.
Most people dont install software
Er… yes they do. My mom who is mid-50s and is someone I would call technophobic/computerphobic installs software. From simple games, to work related stuff. Since I moved out years ago, she has no one to do tech support and she copes well with installing all the necessary software.
On a related note, there are occassions when they need to remove software too, e.g. to make space on the hard drive for more pictures/documenets/etc.
Most people dont administer there System
They damn well should know the basics. Add/remove software, changing various preferences, nothing too taxing. We are talking about home users here, not your typical systems administrator.
Most people go out and buy turn-Key
I take it you mean generic devices? Don’t quite know what this is.
BTW why do you need to install software , GNU/Linux as Live-CD and DV this days ?
I was taking you seriously until I read this. Why do I need to install software? Perhaps it’s because Linux out of the box doesn’t have all the necessary tools that I need (e.g. Eclipse, Java, MATLAB, etc)? LiveCDs do _not_ cut it for average desktop use. Not only does it slow down the system as everything needs to be read off a CD/DVD which is much slower than a hard disk, you lose the functionality of your CD/DVD drive.
I’m sorr to say this, but he last sentence I quoted and the fact that you think Open Source created everything makes it very hard to take you seriously.
Perhaps it’s because Linux out of the box doesn’t have all the necessary tools that I need (e.g. Eclipse, Java, MATLAB, etc)?
By default OpenSource MATLAB on (Knoppix) Linux: Octave or Scilab. Or buy it from mathworks
LiveCDs do _not_ cut it for average desktop use. Not only does it slow down the system as everything needs to be read off a CD/DVD which is much slower than a hard disk, you lose the functionality of your CD/DVD drive.
Simplest installation method of an OS: knoppix-installer
Faster than CD/DVD version, Knoppix on USB key
http://www.knoppix.net/wiki/Knoppix_Customisations
Interesting , It dont mather and I need to chill out , yet your still the one crying and complaining and making false observation.
“You do realize that OS 9 and OS X are practically different operating systems?”
You do realize that Different distribution are different OS shipped by different group/company , I think that the name changes would be enough of a clue … Do you ask your mazda dealer to fix all the default in your Ford even do there all basically *just* cars.
“on OS X there is Classic mode that allows you to run most OS 9 software.”
Which dont work with all files and software.
“Most software that will work on Windows 9x will work on XP.”
No , but then again thats why companie’s invested to upgrade there software to XP.
“Most software that will work on OS X.1 will with X.4.”
Its upgrade version of the same OS …
“Most Linux binaries don’t even work across versions of the particular distro”
The problem again is binary , there made for one versiona and are optimized for one Distribution , the solution is to use the source and the same software , often newer one , that come with the distribution. Thats the problem with people like you , your pointing at what dont work. Let me repeat Binary are the problem of the software vendor , not the distributions.
“This is something that desperately needs to be worked on”
No , because its meaningless , Hardware support and driver support is what’s missing , its not because some neophyte cant install a software that a software cant be installed.
“if you want Linux to work well as a home users desktop.”
75% of the client I have that use GNU/Linux on the desktop are home users …
“Agreed, and it is unnecessary.”
No , an OS already installed for someone who cant do it for themself or by themself is necessary …
“Er… yes they do”
No they dont , they ask or pay an expert to do it for themself.
“My mom who is mid-50s and is someone I would call technophobic/computerphobic installs software. ”
I guess you need to reevaluate what your mother is , its not because you insult her and downgrade her that she fit in the category that you whant to put her in.
“They damn well should know the … too taxing.”
Interesting view of what administering a system is …
“I take it you mean generic devices? Don’t quite know what this is.”
I know , thats the problem you dont know what your talking about at all , Dell , HP , Gaetway are all turn-key , you buy them , plug them in the wall , press power button , there ready to use.
“I was taking you seriously until I read this”
I aint taking you seriously at all.
“Perhaps it’s because Linux out of the box doesn’t have all the necessary tools that I need ”
You do the instalation yet it dont have what you whant …
“LiveCDs do _not_ cut it for average desktop use.”
Yes it does , you just have ” special DUDE chill out take pot all day ” requirements that your unable to meet on your own.
“Not only does it slow down … is much slower than a hard disk”
Get a newer system.
I’m not sorry to say this , but you clearly dont know what Open Source means it means that the source are open and availaible ( How do you end up with so many UNIX if the source are not open ? ). I dont take you seriously at all.
but that long term binary compatibility is why windows is so hole-ly! programers have to always choose to cut off customers or to write secure programs for windows… in addition the binary “push forward” of releases sees versions constantly requireing “compatibility” modes that emulate the bugs of old versions.
The advantage of the OSS way is that when you recompile and repackage a program it gets made up-to-date… using all the latest patchs for all the libraries it needs. If something breaks, it breaks because it should be fixed and it’s not done like that anymore. It makes the pain of security spread out over time rather than big huge bites of incompatibility at once.
Of couse, that’s why commercial companies like binaries… they can require you to purchase new versions to fix “bugs” that shouldn’t have been there anyway. Also, the binary thing is a bit of a sore spot with most of the core OSS development teams. They opened their work up for everybody, and it’s core operating system work… it’s rather silly for driver and utility makers to complain about opening their work for just one part in a whole system. Also, there’s the fear that if the core OSS parts get to have too stable a binary base they can be co-opted by sheer mass of programs that “require” each other with hiden binary only stuff. We already have binary only nVidia drivers, what if a proprietary program for DRM media required them? video editing? playing music? every cool little web app your kids want?? sooner or later the core binary players have cornered the market away from the very people that created it… sure you’d have “linux” but it wouldn’t work with “real programs” without paying “tax” to a cabal of driver writers.
If your goal is Linux adoption, homogeneity is only useful to the extent that it helps adoption (duh). The reason Linux generally gets accepted in many markets is because of specialization — it can be specialized to do a particular job better/more cheaply/more easily than can Windows, Solaris, etc. Homogeneity might help adoption if you’re trying to sell a desktop OS, but that’s not the main focus of most of those who are selling Linux. Certainly, its one of the more high-risk, low potential for return markets that you could target with an alternate OS.
The lessons of the past cannot be forgotten. The things that people often clamor for, pretty interfaces, ease of use, homogenity, even technical elegance, are not necessarily the things that make for a successful product. Lot’s of products have had these things, (BeOS, OS/2, etc), that have not been able to make the kind of impact on the market that Linux is making. More specifically, simply making a “better Windows” is a futile exercise. Competitors must offer something that Windows cannot in order to be successful. Specialization is a very powerful “something” that offers Linux products an edge in the marketplace.
Thank you, rayiner! I’ll never understand why so many people are so insistent that the only form of success that really matters is success on the consumer’s home desktop. It’s a peculiar form of blindness at work which would cause someone to miss the phenomenal successes Linux has experienced over the last decade.
I’ll never understand why so many people are so insistent that the only form of success that really matters is success on the consumer’s home desktop.
Well, there’s no money in it to start off with compared to business desktops as he’s pointed out. However, it’s all Microsoft have to look forward to with Vista……..
Some people always knew that Linux would never be mainstream on the home desktop, and don’t consider it a failure.
And there are some that are convinced it is best thing since sliced bread on the desktop.
Those people generally should get mental help.
Its not because you keep repeating yourself that GNU/Linux desktop is not mainstream and put your hand on your hears and shout “LA LA LA LA dont EXIST!” , and say it dont compare to slice bread because you can afford 3000$ in software , that everyone is like you and that they think its not a god send.
You’re missing the point completely. You can’t talk about “the desktop”. It’s a nebulous market. It makes much more sense to talk about specific use scenarios. Linux can make an excellent professional desktop, for example. It can make a great corporate desktop. It can also make a very good educational desktop, or even a home user desktop in the right circumstances (ie: for users, like my mom, that cannot admin their own machines anyway). Does it make a good replacement for your average gamer? No. Your average Windows “power user”? Probably not. Your general home user? Not until Dell starts shipping machines with Linux preloaded! Does that mean its a failure? Only by some weird definition where healthy market share growth and nice profits for several large companies doesn’t count as a success.
Your complaints about Linux on the desktop seem to revolve around beating up the straw-man. Who exactly should get mental help? Who claimed Linux was a good fit for all market segments? Where are these phantoms you are chasing?
I agree with you rayiner, but the fact is, Ubuntu is targetting the average desktop user, and as such homogeinity is something they are rightly striving for.
Thank you rayiner for that incredibly clairvoyant insight.
See he stole everything from Debian (e.g look at Kubuntu installer)
He never participated in DCC (dreaming that walking alone will bring him big fortunes like Bill Gates)
The core principle of Open Source Community is to share ideas and improve the softwares. He should know the term Collective Responsibility.
If he doesn’t like DCC he can shut up and do his own rocket science.
If he has millions to invest why can’t he develop his own OS rather than building Ubuntu air castle on Debian land?????
Before you open your mouth to speak again, try this… Download and install debian sid. Then grep through all of the changelogs with grep -i “Ubuntu” and see how many results you get.
While not all of them are practical, *many* of the things that Ubuntu changes go back into core Debian. Take for example the entire suite of x.org packages. Ubuntu packaged and qa’d x.org and then submitted the packages to Debian. They were high quality and are now in Debian. What is this air castle that you speak of? No really, enlighten me. As many of the developers that Mark Shuttleworth employs were and still are DDs, much of their work through Canonical goes directly into
Debian.
Seriously people, understand that Ubuntu is a friendly
fork and is not evil. They submit changes back to
Debian when appropriate.
Obviously you just realized that Ubuntu is a better Debian.
-nX
Amen, least original project EVER.
At the most we can say Ubuntu and Debian are room partners. The Have to share utilities. Ubuntu can’t say to other partner like “I have old ford taurus and partner(debian) have lexus so I will not pay my share of utilities. Painting your outside home Dark Brown will make you distinct, but will decrease value of entire neighbourhood.
Seriously, what is your point? Debian and Ubuntu do share utilities and Ubuntu developers consistently submit patches upstream to Debian. HTe most notable of these being xorg, which has been mentioned previously. What more do you want? What the hell does Ubuntu making their default theme brown have to do with anything?
There is just no pleasing some people.
I dont see why they would need Ubuntu , if as you put it they dont pay and contribute there share ? The reality is its the opposite , Ubuntu as brought back life into Debian , the right way , by contributing , distributing and paying for developper and for free CD shipped worldwide. Making all its change availaible for no cost and GPL.
The DCC Allaince product is vaporware , its proprietary and its only adopted by those who are in it.
The DCC Allaince product is vaporware , its proprietary and its only adopted by those who are in it.
You know squat.
Any distro can be built on top of DCC and advertise itself as “DCC Powered” or DCC-derrived.
Educate yourself… http://dccalliance.org/
“The DCC was released A MONTH AGO”
Feel free to offer a link to a ***product*** using it.
“Any distro can be built on top of DCC”
No , thats why its not included in Debian core and why most Debian distribution stay away from it …
“and advertise itself as “DCC Powered” or DCC-derrived.”
Feel free to show one currently shipping or using it or describbing itself as such.
“Educate yourself… http://dccalliance.org/“
Sorry , I already got an education , something that dont exist , yet is advertised as ready is *Vaporware*.
“You know squat.”
I know , its you , who dont know squat
The DCC kernel and Ubuntu kernel will be very
similar if not identical in future DCC releases, and I expect that collaboration will spread to other parts of the system such as X, ACPI etc.
Looks like DCC is becoming UCC (Ubuntu Common Core).
“Benefits of homogeneity” like important vendor/driver support,
All properly written apps and drivers can be installed on nearly every distro. I still play some old loki games on my debian unstable, and the success rate is about the same as with the migration from Windows 98 to XP.
ease of use/administration,
Depends on the targeted users of the distro. Skilled admins loathe GUIs as much as casual users fear CLIs. That’s why one rigid standard would be awful here.
and binary applications that will run for ages and ages on the same kind of system?
See above. If the vendors bothered to support their apps as well as volunteers support OSS apps then neither Windows nor Linux would have to suffer from backwards-compatibility problems, since most of them would be trivial to fix if one had the source.
What could be more important than that, unless you’re not actually interested in Linux adoption?
Believe it or not, most of us aren’t interested in bowing down to vendors of proprietary crapware that view supporting non-Windows platforms as a nuisance.
Dreaming of a unified package format is hopeless and IMHO pointless. The base libraries are stable enough as they are and old replacements are available when needed. Vendors should provide generic packages (like vmware and NVIDIA do) and let the distro-specific tweaking be done by the distro makers themselves if necessary, just like is done with open source packages.
Edited 2006-01-04 21:12
I’m sentimental to both homogeneity and specialization, but the fact of the matter is that there are tons of different distros, yet there are only a limited number of specializations. The _vast_ majority of people that want to run linux are of two different specializations — desktop and server. The wheel of Desktop Linux is being built by so many different distros and countless hours are being wasted. Additionally, application developers are reticent to target anything but ubiquitous platforms in fear of wasting time and money.
Plus, why would having a common core negate specialization? If specialized distro AllAboutPhizix happens to have the same core as the DCC, it’s still specialized, but now I know I can run application X on it too, because it’s DCC certified.
I’m not saying I favor the DCC’s implementation of a common core — I have really looked into it — but some implementation is needed.
It’s hard not to agree with Mark Shuttleworth’s elegant appraisal of the DCC. It has come over as a pretty fluffed and amateur effort from the start, certainly in presentational terms.
However, there is a niggling problem with his wider analysis, namely that what is good for Ubuntu is good for Debian and the whole Debian community. Imho, that is not true. Ubuntu’s success comes at a cost to Debian, which is slowly being pushed into the role of surly-staffed DIY warehouse for those come to build their own distros.
I have a great deal of affection for Debian and I am sorry to see this process of competition draining its energies. And it is competition, no matter what Ubuntu and various “useful fools” may think. This article, after all, slyly pitches Ubuntu as the Debian-based distro corporations should consider, even mentioning their compliant kernel. The DCC is a competitor(s) in this respect, especially since Ubuntu are almost certainly planning a very serious push into the corporate side over the next few years and right in the firing line is Debian Sarge.
Sooner or later, a lot of folks with a foot in both camps will find that riding two horses doesn’t work and it will be a painful discovery. Come the time, I expect Ubuntu will issue a very elegant appraisal of that situation too, quite possibly in the form of a eulogy over the grave of Debian Stable.
Edited 2006-01-05 00:23
There is a lot of activity in GNU/Linux, some will fly & some will not. Everyone should step back and take a breath. I use Debian Sarge because it just “works.” I like Ubuntu, Mepis, Suse………., but Debian is quality, and in my opinion will thrive. That they are using innovation from “other distros” is a testament to their quality code. Debian is the foundation, and I feel that they actually wind up with more users because of the Ubuntus, Mepis(which moved me to Debian), new users who like the benefits of GNU/Linux will often start with a “user friendly” distro. Debian is very user friendly, however, my first try with Debian was with Woody, and I could not figure out how to install it! I felt that with all my experiance with “recovery disk” on WinXP, it would be easy. Mepis introduced me to Debian. The freedom of GNU/Linux is what is pushing innovation, but also hindering it’s wider awareness to the public & enterprise. The few people (and I mean few) I have converted to GNU/Linux are very enthusiastic. Sorry to venture from the subject, but Debian and GNU/Linux will continue to grow. Any alliance or other forms of cooperation will increase the usage of this fantastic OS.
Paul Sams
…if you’ve just started your way from Debian, youhave to say that diversity is good. Which can be true, on many levels, but not on what grounds the DCC was formed and as the name suggests, to have a common core, on which one can build their exotic diversities as they see fit.
Anyway, Ubuntu is good, even me as a stubborn longmany-years-old debianer and a new gentooer order many of the newer Ubuntu release disks to spread among people. Why, because they will get to know a debian-like system easily. Still, installing debian is very very far from being hard and cold as many people suggest. Installing debian unstable on a new laptop last week, with network install was a breeze, with everything working from pcmcia, bluetooth, net, wifi, you name it.
Anyway, if Shuttleworth sees the DCC’s future being void, I certainly hope he is wrong. I’d really like to see DCC succeed and provide a very nice foundation for many future distros to be based on.
It was clear from the outset that the DCC was simply another UserLinux (bizarrely, now a part of the DCC!), and an attempt by Ian Murdock to get his Debian (well, that’s a laugh!) seen as some sort of standard distribution.
Mark Shuttleworth is somewhat right about the differentiation though. If you see a screenshot of a DCC distribution then which one is it, and if these distributions are all the same what’s the point of having different companies and organisations involved? He is slightly wrong in that there are scenarios in which it might work (imagine an open sourced Windows packaged up with a Novell Client working properly), but the way groups like the DCC are going about it seems to be upside-down and with a lack of direction.
The alternative? Well, Mark Shuttleworth hinted at that as he sees Ubuntu as being seen as some sort of standard distribution! Game on!
> The alternative? Well, Mark Shuttleworth hinted at that
> as he sees Ubuntu as being seen as some sort of
> standard distribution! Game on!
Not quite. If you listen closely to this and other talks, he’s basically advocating source level compatibility through distributed source control sharing of various distributions. According to him, different distros will use different compile time options for different reasons to fill the needs from spartan distros that fit on cellphones to mainframes on different chipsets, from bleeding edge to stale mission critical conservative. Rather than create a lowest common denominator distro, each distro should focus on what it’s best at and share changes with other distros. It’s a pretty Gentoo-like idea that I don’t really agree with, but he does have a few valid points.