It’s been two weeks since Rick Spencer made the case for a rolling release approach in Ubuntu. Having a rolling release is one of the very top suggestions from the hardcore Ubuntu user community, and after years of it being mooted by all and sundry I thought it deserved the deep consideration that Rick and his team, who represent most of Canonical’s direct contributions to Ubuntu, brought to the analysis.

It’s obviously not helpful to have mass hysteria break out when ideas like this get floated, so I would like to thank everyone who calmly provided feedback on the proposal, and blow a fat raspberry at those of you who felt obliged to mount soapboxes and opine on The End Of the World As We Know It. Sensible people the world over will appreciate the dilemma at being asked to take user feedback seriously, and being accused of unilateralism when exploring options.

Change is warranted. If we want to deliver on our mission, we have to be willing to stare controversy in the face and do the right thing anyway, recognising that we won’t know if it’s the right thing until much later, and for most of the intervening time, friends and enemies alike will go various degrees of apoplectic. Our best defense against getting it wrong is to have a strong meritocracy, which I think we do. That means letting people like Rick, who have earned their leadership roles, explore controversial territory.

So, where do we stand? And where do I stand? What’s the next step?

What makes this conversation hard is the sheer scale of the Ubuntu ecosystem, all of which is profoundly affected by any change. Here are the things I think we need to optimise for, and the observations that I think we should structure our thinking around:

Releases are good discipline, cadence is valuable.

Releases, even interim releases, create value for parts of the Ubuntu ecosystem that are important. They allow us to get more widespread feedback on decisions made in that cycle – what’s working, what’s not working. Interestingly, in the analysis that played into Rick’s proposal, we found that very few institutional users depend on extended support of the interim releases. Those who care about support tend to use the LTS releases and LTS point releases.

Release management detracts from development time, and should be balanced against the amount of use that release gets.

While reaffirming our interest in releases, I think we established that the amount of time spend developing in a cycle versus spent doing release management is currently out of whack with the amount to which people actually DEPEND on that release management, for interim releases, on the desktop. On the server, we found that the interim releases are quite heavily used in the cloud, less so on physical metal.

Daily quality has raised the game dramatically for tip / trunk / devel users, and addresses the Rolling Release need.

There’s widespread support for the statement that ‘developers can and should use the daily development release’. The processes that have been put in place make it much more reliable for folks who want to track development, either as a contributor to Ubuntu or as someone who ships software for Ubuntu and wants to know what’s happening on the latest release, to use Ubuntu throughout the development cycle. For those of you not aware, uploads to the edge get published in a special ‘pocket’, and only moved into the edge if they don’t generate any alarms from people who are on the VERY BLEEDING EDGE. So you can use Raring (without that bleeding edge pocket) and get daily updates that are almost certain not to bork you.  There is a real community that WANTS a rolling release, and the daily development release of Ubuntu satisfies this need already.

LTS point releases are a great new enhancement to the LTS concept.

On a regular basis, the LTS release gets a point update which includes access to a new, current kernel (supporting new hardware without regressing the old hardware on the previous kernel, which remains supported), new OpenStack (via the Cloud Archive), and various other elements. I think we could build on this to enhance the LTS with newer and better versions of the core UX (Unity) as long as we don’t push those users through a major transition in the process (Unity/Qt, anybody? ;-)).

Separating platform from apps would enhance agility.

Currently, we make one giant release of the platform and ALL APPS. That means an enormous amount of interdependence, and an enormous bottleneck that depends largely on a single community to line everything up at once. If we narrowed the scope of the platform, we would raise the quality of the platform. Quite possibly, we could place the responsibility for apps on the developers that love them, giving users access to newer versions of those apps if (and only if) the development communities behind them want to do that and believe it is supportable.

Phew.

That’s what I observed from all the discussion that ensued from Rick’s proposal.

Here’s a new straw man proposal. Note – this is still just a proposal. I will ask the TB to respond to this one, since it incorporates both elements of Rick’s team’s analysis and feedback from wider circles.

Updated Ubuntu Release Management proposal

In order to go even faster as the leading free software platform, meet the needs of both our external users and internal communities (Unity, Canonical, Kubuntu, Xubuntu and many many others) and prepare for a wider role in personal computing, Ubuntu is considering:

1. Strengthening the LTS point releases.

Our end-user community will be better served by higher-quality LTS releases that get additional, contained update during the first two years of their existence (i.e. as long as they are the latest LTS). Updates to the LTS in each point release might include:

  • addition of newer kernels as options (not invalidating prior kernels). The original LTS kernel would be supported for the full duration of the LTS, interim kernels would be supported until the subsequent LTS, and the next LTS kernel would be supported on the prior LTS for teh length of that LTS too. The kernel team should provide a more detailed updated straw man proposal to the TB along these lines.
  • optional newer versions of major, fast-moving and important platform components. For example, during the life of 12.04 LTS we are providing as optional updates newer versions of OpenStack, so it is always possible to deploy 12.04 LTS with the latest OpenStack in a supported configuration, and upgrade to newer versions of OpenStack in existing clouds without upgrading from 12.04 LTS itself.
  • required upgrades to newer versions of platform components, as long as those do not break key APIs. For example, we know that the 13.04 Unity is much faster than the 12.04 Unity, and it might be possible and valuable to backport it as an update.

2. Reducing the amount of release management, and duration of support, for interim releases.

Very few end users depend on 18 months support for interim releases. The proposal is to reduce the support for interim releases to 7 months, thereby providing constant support for those who stay on the latest interim release, or any supported LTS releases. Our working assumption is that the latest interim release is used by folks who will be involved, even if tangentially, in the making of Ubuntu, and LTS releases will be used by those who purely consume it.

3. Designating the tip of development as a Rolling Release.

Building on current Daily Quality practices, to make the tip of the development release generally useful as a ‘daily driver’ for developers who want to track Ubuntu progress without taking significant risk with their primary laptop. We would ask the TB to evaluate whether it’s worth changing our archive naming and management conventions so that one release, say ‘raring’, stays the tip release so that there is no need to ‘upgrade’ when releases are actually published. We would encourage PPA developers to target the edge release, so that we don’t fragment the ‘extras’ collection across interim releases.

 

That is all.

93 comments:

  1. Dennis Martin Herbers says: (permalink)
    March 12th, 2013 at 2:09 am

    “Our working assumption is that the latest interim release is used by folks who will be involved, even if tangentially, in the making of Ubuntu, and LTS releases will be used by those who purely consume it.” -> Most regular consumers are using Ubuntu 12.10 Non-LTS because that is advertised as the “last Ubuntu release” with the “newest features”. Your assumption seems weird.

  2. mark says: (permalink)
    March 12th, 2013 at 2:18 am

    @Dennis – we have data to support the LTS assertion, from web browser fingerprints.

  3. manny says: (permalink)
    March 12th, 2013 at 2:22 am

    >””Separating platform from apps would enhance agility.
    >Currently, we make one giant release of the platform and ALL APPS. That means an enormous amount of interdependence, and an enormous bottleneck that depends largely on a single community to line everything up at once. If we narrowed the scope of the platform, we would raise the quality of the platform.””

    This will have a big positive impact. It has been long over due.

    Separating of platform from apps is similar to the “layers separation” that some “half/semi-rolling” models are doing. i.e.: http://chakra-linux.org/wiki/index.php?title=Half-Rolling_Release_Model

    >””LTS point releases are a great new enhancement to the LTS concept.
    >I think we could build on this to enhance the LTS with newer and better versions of the core UX (Unity) as long as we don’t push those users through a major transition in the process (Unity/Qt, anybody?)””

    Woohoo ! :)

  4. Andre says: (permalink)
    March 12th, 2013 at 3:10 am

    I seriously think anything but a semi-rolling/rolling release with LTS in the middle, is pretty pointless.I know chance is scary, but what about the benefits?

  5. killerkiwi says: (permalink)
    March 12th, 2013 at 3:14 am

    +1 for the app separation like Chakra

  6. Lammert Nijhof says: (permalink)
    March 12th, 2013 at 3:15 am

    I have 12.04 LTS, 12.10 and 13.04 installed. And I am currently mainly an interested consumer. For my daily hobby I am almost completely working with 13.04 and I am impressed by the quality it brings. I never use 12.10 anymore, but I still use 12.04 LTS from time to time and so does my wife. The only time I had some problems with 13.04 was in early December as far as I remembered caused by the fact that the daily builds did not install very well. In my case I would welcome a rolling development release continuing without interruption from the same repositories. I think I would expect to upgrade, if I went from intermediate release to the next release. A 7 month maintenance period seems quite logical. For the intermediate releases I would only expect bug and security fixes, probably for OS and some core applications. For the LTS release you could make a distinction between between core applications and all the others in the software center. I have noticed that Firefox and Thunderbird are updated also to the latest release and the same could be done for e.g. the new office release, development software like Eclipse, Virtualbox or the new releases of the multimedia applications. If you consider to add that to the LTS point releases, I think you should offer a possibility to enable/disable it for each core application using e.g the “software sources”.

  7. Fabio Rafael Rosa says: (permalink)
    March 12th, 2013 at 3:28 am

    I think you should advertise LTS better, and direct most users there.

  8. Andrea Grandi says: (permalink)
    March 12th, 2013 at 3:35 am

    Dear Mark, for me it was clear even a couple of years ago, so I can’t do anything else than agree with you ;)

    6-months release never made sense for me. LTS did instead. Server users and hardware vendor (targeting desktops) want very stable releases. They surely don’t want a release every 6 months. On the opposite side, developers and power users don’t want to be left behind and they always want updated software.

    Rolling release doesn’t mean unstable. If you release often and you release few things, the product can be even more stable than a rushing 6 months release that risk to be unstable just because “you” (meaning the product leader) want to put a certain feature within time. Releasing often only makes the continuous integration process easier. So, to be short: keep going with LTS every 2 years, drop the usual 6-months release and keep rolling from LTS to LTS. After all it’s the concept that Scrum development is based on.

    This is just my advice, I only count as 1. But if you remember my past advices, you adopted them all with time :P (I tried to convince everyone 2 years ago to drop the actual Unity and keep going with Unity-2D that was made in Qt. I also proposed remote UDS 1 year ago, etc….) so I’m just happy to see that now you’re closer to the vision I had earlier :)

    My best regards.

  9. Tylor Steinberger says: (permalink)
    March 12th, 2013 at 3:38 am

    I definitely have to agree with Dennis Martin Herberts regarding interim releases. I have a fair amount of friends who use Ubuntu, and they always upgrade to the newest releases because it is just simply the latest and greatest. Most of them just being regular end users. I could be wrong, but I definitely think that most end users would typically upgrade to keep up-to-date.

    I think that ‘Separating platform from apps’ would be hugely beneficial for Ubuntu and it’s users. I found it very irritating when I first began using Ubuntu when I’d upgrade and an application that I used all of the time wasn’t in the new repository yet or was just an older version thats dependencies were with an older version of package x and no longer worked.

    Overall, I think this could definitely work.

  10. Chris Fisher says: (permalink)
    March 12th, 2013 at 3:38 am

    I like this, seems like a really good middle ground, and it’s clear enough that I think it’s fair to expect end users to understand it.

    It could be a bit short for cloud hosted instances (7 months vs 18 months) so I think some messaging needs to be done to train those users that the LTS releases will be seeing more feature improvements, and they should be using those.

  11. Bill Wood says: (permalink)
    March 12th, 2013 at 4:25 am

    This post is for myself and NOT on behalf any customer / client / or other person or companny who I, or my company, may do business with.

    ==============

    I am currently at a large technology products company who has a number of significant commercial offerings. Recently an internal question was asked about the future of their products in the mobile space over the next 12-18 months. My internal response raised the Ubuntu project on mobile as a possible disrupter to the marketplace.

    The ability to have a multi-purpose platform (a FULL OS), delivered in a cell phone sized form factor, with a possible hot dock station for real computer processing presents a huge disruption to the entire tech marketspace. IF Ubuntu can pull this off it represents a gigantic move that makes single use devices like PCs, cell phones, and even pads obsolete or semi-obsolete. As cell phone processing power improves over the next couple of years this represents a HUGE shift in the technology landscape.

    ==============

    With that background, and as an outsider, here is my take on this issue of “rolling releases” vs. separating applications from the OS (the “semi-rolling” release).

    The idea of separating the applications is appealing to me personally, and I’m sure it would be to business. This provides a much more limited development surface for regular updates, or in the case of any problems or security issues, a simpler avenue to provide quick fixes. While there seems to be quite a bit of MS hate in the Linux world, MS does have their “Tuesday patch” process pretty well ironed out. And it seems as though it is a good approach as well. They have learned their lessons from paying customers. As a technology consultant in a NON-OS area (I do SAP work) I believe the “semi-rolling” release in some form would have merit and make for a higher quality OS where the applications can “catch up” as needed.

  12. Benjamin Kerensa says: (permalink)
    March 12th, 2013 at 6:41 am

    I like the idea of splitting apps from the foundation.

  13. niagr says: (permalink)
    March 12th, 2013 at 8:35 am

    Thank you for making the right decisions despite pressure from a regressive community.

  14. M1AU says: (permalink)
    March 12th, 2013 at 8:56 am

    As a user, I find the most annoying thing about Ubuntu LTS releases that the apps are most likely always out of date, so after some time the system is full with dozens of PPAs which then prevent me from upgrading to the new LTS release. Please consider separating the platform from the apps, so that app developers can push out updates as soon as they’re available and tested.

  15. Fred R says: (permalink)
    March 12th, 2013 at 9:57 am

    I think “populated” backports may represent a sort of Rolling for LTS releases. Unfortunately, except for some multiverse packages, backports are really empty… (same for Debian indeed).
    Moreover, “suggested updates” and “backports” may appear confusing as well…

    As release cycle is currently discussed, does Ubuntu really need an intermediate version every 6 months ? 1 intermediate release between each LTS should be enough, with support for 13-14 months (1-2 months after each annual release). Open source software are now quiet mature, and should upgrade pretty well on top of an intermediate release for a maximum of 1 year.
    Example: 12.04 LTS, 13.04, 14.04 LTS, 15.04, 16.04 LTS

    Of course, it would break the 6-month commitment to some open-source projects such as KDE and Gnome.
    But missing 1 out of 2 releases may not hurt (and PPA can provides upgrade for those interested by current version)
    Furthermore, these projects may realize that their 6-months releases doesn’t provide as many improvements as they think, and that releasing every year may be really comfortable and may help to provide better products !

    I would like to conclude with a simple statement: My mother is able to upgrade Libreoffice to version 4.0 on windows XP, and I can’t do the same on Ubuntu LTS by simply updating the system.
    Regardless what this debate may bring, we may think, 1 day or another, to be able to provide this kind of behaviour to the Linux world:
    to be able to upgrade 1 particular version of a software, to any version of Ubuntu, on top of which this particular software is able to compile !
    Sound so obvious, and unfortunately, there is no distribution providing this behaviour…

  16. RG says: (permalink)
    March 12th, 2013 at 10:10 am

    I like the splitting apps (developer driven updates) idea, but doesn’t carry security issues within? This is why, I suppose, Apple added a sandbox to their Mac’s app store or Google requires a permissions’ manifest in Android.

  17. Justyn says: (permalink)
    March 12th, 2013 at 10:15 am

    Splitting apps from the foundation sounds a very worthwhile goal.

    The current process makes running the latest version of an open-source app harder on Ubuntu than it is on Windows/OS X or even Android, and it usually results in having a mess of PPAs installed.

    I deal with various personal (that is non-enterprise) users who would be much happier using LTS releases but get too frustrated with so many of their applications creeping out of date, sometimes by years.

    I’m looking forward to hearing more about this.

  18. hitaisin says: (permalink)
    March 12th, 2013 at 10:49 am

    Hello, I’m ubuntu user from 4.10 :), I always use latest version, in last two years I started using ubuntu beta versions, now I’m using 13.04 from january. So it’s really stable, only problem that I often see those “bug happened, send report” windows.

    As user I want: latest firefox, libreoffice, unity (with every release it becomes better and better, so I want latest version) and latest video drivers to play games on steam. I don’t know much about “core” system, I just want latest version of apps, like in android. You can have old android 2.3, but you can download latest versions of all apps.

    As user I don’t like how PPA system work, it’s not really convenient to install PPAs, half of them are broken, I always see “can’t download repositories information”.

    Software Center is not as good, as it should be. Almost no apps from ubuntu app showdown appeared there, just can’t install them. USC must look like chrome web store or google play. I want it to look like modern website, with comments, ratings, reviews, app news, big screenshot, videos, always updated, integrated donation feature for free apps, etc…

    Many people still use ancient windows xp, I know many people who just don’t understand why he/she needs to upgrade system if everything is working. I think only geek users, and people how has interest (their job or just hobby) in IT and computers want “latest versions”. LTS version with ability to update firefox, libreoffice, unity and maybe some more apps (devs must have ability to update their apps in LTS repositories) are what is most people need. This is my understanding of question :).

    I wish you good luck! I hope ubuntu will succeed, I’am already put money to shelf for ubuntu phone, just waiting when it appears.

    P.S. It is really sad that after so many years still there is no default calender app for ubuntu. I hope ubuntu core apps project will fix it and they make desktop version too.

  19. Dennis says: (permalink)
    March 12th, 2013 at 12:22 pm

    It is nice to see what looks like thoughts (including comments) swinging towards reasonable conversation about a coherent pragmatic strategy. The team that studied the original concept and Rick Spencer who seemed to be the face of it did some great work and approached it with passion. In the end what results may not be what they originally envisioned but I for one am thankful that there is a “think big” attitude that is balanced with concern for the end user. I am just an every day user without the technical ability or time to contribute meaningfully (other than switching friends and family) but in the end I feel good about where Ubuntu is headed.

  20. Mike Frett says: (permalink)
    March 12th, 2013 at 12:23 pm

    Hi Mark, I’d like to speak to you as a user who converted fully from that ‘other’ OS. I was going to use XFCE Debian because of the 1-3 years in between releases but; I chose Xubuntu 12.04 because of the 5 year cycle. I LIKE things when they stay the same, I dislike being ‘forced’ to upgrade when the upgrade is something that looks different than what I’m use to or breaks my Hardware or Software.

    I don’t mind the point releases that add a new Kernel or Hardware support, I don’t mind daily updates to squash bugs either. I DO dislike that a lot of Apps in the Software Center are very out of date, usually causing me to go to the App website to Compile something myself or pickup a premade Binary. Basically I like being up to date, but I don’t like things (the GUI) functioning or looking different then what I like.

    I am the kind of people that are migrating from ‘other’ OS’s to something that doesn’t force me or change what I’m use to. I dislike what I’m about to say, but if something happened that started forcing me to use something I didn’t like or changed the way it functioned; that would force me to find a different Platform that offered an LTS type of Cycle =(

    I really dislike being forced, Mark. It’s not the Linux way, it’s the way the ‘other’ guys do it.

  21. Franck says: (permalink)
    March 12th, 2013 at 12:42 pm

    What about delta updates? That Ubuntu takes 100x the bandwidth of Windows for security updates is such a major embarrassment, due to stone age apt-get that redownloads half the distribution every time (that is about every week) there is a one-line security code fix in the major monolithic packages (kernel, browsers, office) and all their dependencies given that granular packaging is broken. For people in expensive bandwidth environments a Windows license is cheaper that paying the apt-get tax to their telco… and even when bandwidth is cheap it’s such a waste of resources!

  22. Pjotr says: (permalink)
    March 12th, 2013 at 1:40 pm

    Excellent proposal, Mark! I’m all for it.

  23. Henry Gernhardt says: (permalink)
    March 12th, 2013 at 1:42 pm

    As I’ve been watching this entire hullabaloo in progress, I’ve reminded myself a million times that I don’t view Ubuntu as a Linux distribution.

    I view Ubuntu as an */Operating System/*.

    This means that I /like/ the idea of splitting the apps from the platform, along with shifting maintenance to the app developers. I /like/ the focus Ubuntu has made on creating a very appealing and usable User interface and experience. I /like/ the convergence of Ubuntu across platforms. I /like/ the synergy I’m seeing in the desktop experience, and look forward to the time when my Thing That Tells Me Stuff (a first generation Verizon Samsung Galaxy Tab 7) can run a usable, stable, and nicely app-populated (apopulated?) Ubuntu operating system.

    This /also/ means that I view each Ubuntu release as */a new operating system/*. I don’t look at moving from 12.10 to 13.04 as an upgrade. I look at it as an */install/*. From scratch. Bare bones. Factory reset. Back up ~/ and pray. Start-at-the-beginning kinduvathing. My work environment requires stability, so I look to the LTS. If the LTS won’t work on my hardware, I use the latest release and stop at the next LTS (the last time this occurred was with a Compaq laptop; I tried 10.04 and had to quickly move to 11.10). I won’t move from the LTS unless I absolutely must, and then I prefer to move to a new LTS.

    Although my use case doesn’t support rolling releases, I */like/* the idea you’ve put forward. You’ve kept the LTS release, kept the 6-month release cycle, and added rolling release goodness all at the same time. Sure, interim releases don’t have the support length, but that’s perfectly fine. It pushes folks towards one side or the other. Rolling releases for folks who like latest and greatest, and LTS for those who demand stability and longevity. The 6-month release can serve as an intro point for those who aren’t sure which way to go. Of course, there will always be those who want to stick with the 6-month release. It’s good to know that, according to your proposal, those releases will still be there.

    Mark, I have to applaud you in standing before the onslaught which has come your way. I applaud your goals. Personally, I /like/ using Ubuntu and will continue to do so for as long as it is supported and meets my needs. I earnestly hope that, as this entire thing settles into normality, the Ubuntu community will rally around what you and Canonical are attempting. I truly see all of this as a great move towards crushing Bug #1.

  24. Ubuntu - pomoc | Mark Shuttleworth przedstawił swoją wizję przyszłości Ubuntu says: (permalink)
    March 12th, 2013 at 2:12 pm

    [...] swoim blogu przedstawił dwa punkty, które aktualnie są rozważane przez [...]

  25. Thomas Kluyver says: (permalink)
    March 12th, 2013 at 2:23 pm

    I agree with everyone else who’s been saying that separating apps from the platform makes sense, both from a user and a developer perspective. I often see third parties advising users to avoid distro repositories because they’ll be out of date – that reduces security and makes updates more awkward (because not everything is managed by apt).

    @RG: There’s already been some thought about how to handle security for applications uploaded through developer.ubuntu.com. See https://wiki.ubuntu.com/AppDevUploadProcess

  26. Dale Beaudoin says: (permalink)
    March 12th, 2013 at 2:26 pm

    Hello Mark,

    “There’s widespread support for the statement that ‘developers can and should use the daily development release’. The processes that have been put in place make it much more reliable for folks who want to track development, either as a contributor to Ubuntu or as someone who ships software for Ubuntu and wants to know what’s happening on the latest release, to use Ubuntu throughout the development cycle. For those of you not aware, uploads to the edge get published in a special ‘pocket’, and only moved into the edge if they don’t generate any alarms from people who are on the VERY BLEEDING EDGE. So you can use Raring (without that bleeding edge pocket) and get daily updates that are almost certain not to bork you. There is a real community that WANTS a rolling release, and the daily development release of Ubuntu satisfies this need already.”

    I think you said it all here. I am not against progress and new ideas but when something works as good as the current release system is then why change it? A great knowledge base is generated from those end_users who test the daily .isos and the interim 6 month releases leading up to LTS. I would just consider if it would cause any extra burden many current beta testers and their testing techniques of all Ubuntu flavours?

    “3. Designating the tip of development as a Rolling Release.

    Building on current Daily Quality practices, to make the tip of the development release generally useful as a ‘daily driver’ for developers who want to track Ubuntu progress without taking significant risk with their primary laptop.”

    Many credible testers definitely have more than one machine working . That’s just conventional wisdom to assume so (and practice in practice). I think it unreasonable to make such a dramatic shift to accomodate developers when all they need do is carry an LTS on a USB stick for their primaries. I really do not see the logic in incorporating this concept. From a beta testers point of view it is always ‘if it ain’t broke
    don’t fix it” and I think this fairly applies to the way the release system is set up currently.

    There certainly may be benefits in the long run to incorporate these concepts but all I ask to consider is ask the proposers if the downtime involved in brining such a concept to fruition is worth enduring all of the down-time deployed along with it?

    Thanks for Unity ! Thank you .. and my eyes thank you :)

    Kind Regards,
    Ventrical

  27. k1l says: (permalink)
    March 12th, 2013 at 2:57 pm

    I think there is alot of truth in this new proposal.

    - The LTS is the way to go for most “users”. But the problem so far is, that hardwaresupport (with kernel etc) is best in the latest interim release and the apps get outdated very fast these days which results in a billion PPAs and causing alot of trouble. So most “users” go to the interim releases when they should (systemwise) stay with a LTS. So making a LTS with better “point-upgradaes” and seperate the apps layer would be a real advantage.

    - 18month support window is way to long for interim releases. if you want long support you should stay with LTS. So 7 or 8 month support would fit since you have to upgrade anyway and if you want to stay: use a LTS.

    - i dont think a rolling release would bring any advantage. of course its attractive from the technical side of view but the benefits dont suit into the targets of ubuntu for the whole community and the whole userbase. the development cycle offers a way to get the latest updates if you are interested in these.

  28. manny says: (permalink)
    March 12th, 2013 at 3:13 pm

    @Fred R

    Actually it has been done. Mandriva moved to 1 interim release (once per year) in between LTSs and Opensuse moved to 2 interim releases from 3 (every 9 months).

    I think the interim actually fragment and also add too many freeze periods that halt development and of course consume maintenance resources after, so the less the better.

    Since we want to avoid the “one giant release” as stated in the post: “Currently, we make one giant release of the platform and ALL APPS. That means an enormous amount of interdependence, and an enormous bottleneck that depends largely on a single community to line everything up at once.”

    And also the support for these will be just 7 months, so users need to upgrade right away (but why wait 6 or 7 months? the type of user that will go for these, which are developers, testers and power users should be able to keep rolling even prior if they want.) then IMO the interim should be more of a “snapshots with support” than an early “full blown giant” release.

    And to me snapshots can be just as useful as the full releases, look at the LTS point releases are just snapshots, but on some hardware (EFI) you can not install 12.04.1, but 12.04.2 should work.

    The good thing about the hybrid rolling model is that each project (or layer), like the phablet, QT-unity, flavors, core stuff, Apps, etc. can advance as fast as they need without becoming a bottleneck for the others then each roll out their respective snapshots.

  29. Anzan says: (permalink)
    March 12th, 2013 at 3:19 pm

    Mark, this proposal sounds very reasonable and, likely, doable.

  30. Arioch says: (permalink)
    March 12th, 2013 at 3:59 pm

    The great Carl Sagan, in his novel ‘Contact’ wrote: If every person could go out to space and look upon our planet from outside, then realize how insignificant we are, then wonderful things would happen.
    Mark went out to space, he came back and changed the world with Ubuntu. At least it changed my world.

    Keep up the amazing work, +1 to separating apps from core.

  31. Bernhard Schaefer says: (permalink)
    March 12th, 2013 at 4:02 pm

    I like this proposal. I’d appreciate it. It’s a good combination of LTS and enabling the people who want the brand new stuff to use it. Thank you, Mark!

  32. Lagos says: (permalink)
    March 12th, 2013 at 4:32 pm

    I have voiced my share of opinions about the rolling release proposal, but this time instead of criticizing I will applaud you. This makes a lot of sense. Good work.

  33. Dylan McCall says: (permalink)
    March 12th, 2013 at 4:35 pm

    That looks lovely! Since there’ll only really be one interim release active at a time, I wonder if it makes sense to name them? It might be tidier if we just call them Ubuntu+1 and leave the fun names for the current and next LTS.

    I’m glad lots of people seem to be on the same page about splitting applications away from the base platform. What do you think about using a different format from Debian packages for applications, too? This seems like a good opportunity to collaborate with the broader ecosystem on a modern application package format that is aimed at the single job of packaging applications (thus much simpler to work with), that supports application permissions (for sandboxing!), and has straight-forward compatibility across distributions ;)

  34. El 'maremágnum Ubuntu' que no cesa says: (permalink)
    March 12th, 2013 at 4:36 pm

    [...] con el artículo que Mark Shuttleworth publicó ayer en su blog bien entrada la madrugada, en el que volvía una vez más a hablar del tema rolling release. A [...]

  35. Mahesh says: (permalink)
    March 12th, 2013 at 5:05 pm

    I don’t understand why most of the linux users assume that nobody needs LTS. I mean many of us are on slow Internet connection and can’t afford even monthly updates of firefox. I do however see that by separating the kernel and supported apps makes things better here. I like this new proposal and I think it is a good balance of both sides of arguments.

  36. ag.restringere says: (permalink)
    March 12th, 2013 at 5:14 pm

    The best part is separating applications from the distribution release cadence. Applications should evolve and be released on a cadence determined by the application developers and should under their control. This is how things work in the wider Windows, Android and Apple world so Linux developers who make cross-platform applications could sync development and release across all of their channels. This will allow them to deliver current software to all users in the LTS, Interim and “bleeding edge” Ubuntu versions and consolidate support at a single version instead of supporting multiple versions. This should also include things like GPU Drivers as well. Nvidia for example has it’s own release cadence and they certify drivers at their own pace. Often Ubuntu nvidia-drivers are behind the officially certified stable Nvidia driver versions and this leads to redundancy in Ubuntu support channels. With GPU drivers it is key for users at all Ubuntu versions to have the latest officially certified GPU drivers to ensure the best performance and stability for their systems.

  37. Jef Spaleta says: (permalink)
    March 12th, 2013 at 6:08 pm

    @Mark,

    “we have data to support the LTS assertion, from web browser fingerprints.”

    I few questions about that comment.
    Is this Canonical internal web traffic, or some external web traffic stats like wikimedia’s?

    Which specific useragents strings are you tracking? I thought both firefox and chrome at some point on Ubuntu stopped providing vendor specific release info. I’ve talked to the wikimedia people exactly about this issue with user agent strings. So I’m very unclear about how you track this.

    If you can explain the methodology, I might be able to reach out to the wikimedia stats guy and see if he can do a special squid reporting run and correlate from their browser client data with what you are saying. As of now the public data analysis from wikimedia doesn’t give the LTS/non-LTS breakdown. But again, I’m not even sure how to approach it because the User agent strings for the most commonly used browser just say “linux” now when running on Ubuntu.

  38. Andrew says: (permalink)
    March 12th, 2013 at 7:56 pm

    Thanks Mark, first RickRoll’D in years!

  39. Anzan says: (permalink)
    March 12th, 2013 at 9:33 pm

    Mark, please no new packaging format for applications. Canonical is in enough trouble.

  40. Otus says: (permalink)
    March 12th, 2013 at 9:34 pm

    “For those of you not aware, uploads to the edge get published in a special ‘pocket’, and only moved into the edge if they don’t generate any alarms from people who are on the VERY BLEEDING EDGE.”

    Is this a new proposal or something that is in the process of being implemented? From the discussion on ubuntu-devel, I understand that the pocket is currently only used to check that the packages compile and automatic tests (where available) run. No one is supposed to be using it for manual testing.

  41. La propuesta de Mark Shuttleworth para los lanzamientos de Ubuntu says: (permalink)
    March 12th, 2013 at 10:14 pm

    [...] – Mark Shuttleworth: «El Rolling Release no me convence» Fuente – Blog de Mark Shuttleworth GA_googleFillSlot("468x60_ubunlog"); [...]

  42. Drew Johnson says: (permalink)
    March 12th, 2013 at 11:30 pm

    I totally agree with the part about LTS NEEDS to be strengthened. I’ve been saying that for years (and getting criticized for it by fellow Ubuntians who must have thought it was somehow “disloyal” to suggest alternatives). The in between releases are inconsequential for the vast majority of the members of your stated audience — less technical computer users. They shouldn’t have to deal with what are essentially beta releases. Let people like me who are avid Ubunu community members work through all the issues that the non-LTS releases will bring. That’s what the Ubuntu community is for. That and for you to listen to our feedback on important issues.

    “we have to be willing to stare controversy”. Really Mark… at some point I hope you’ll own up to the fact that your condescending attitude (that you’ve displayed even here) towards any alternative feedback actually creates the backlash that you profess to dislike so much. It’s almost like you want to create controversy in order to get more press and more attention. The faux pas list is getting very long indeed, but IMO I don’t see real apologies from you.

    Living with this drama and bad blood in the GNU/Linux community is not a requirement. If you’d simply apply the Code of Conduct to your own behaviour and how you treat the wider GNU/Linux community you wouldn’t have this phenomenon. But you take a position that you are above the CoC. SABDFL. Yeah right. Is there any such thing as a “Benevolent Dictator”? I don’t see how. Its an oxymoron in my opinion. Please start actually listening to the community Mark.

  43. mark says: (permalink)
    March 13th, 2013 at 12:20 am

    @Otus, yes, you’re right, I stand corrected, the bleeding edge pocket is for machines not humans :)

  44. Simon Strandman says: (permalink)
    March 13th, 2013 at 12:35 am

    This is awesome if you can pull it off! Linux will never be able to make a dent in Windows’ market share for desktop PCs if people need to upgrade their os just to get new software and drivers.

  45. Rob Boudreau says: (permalink)
    March 13th, 2013 at 2:13 am

    I’m not sure I understand the need for an interim release at all. If it’s decided to go with update “snapshots” for LTS, bringing newer kernels and applications to it, why not just use the snapshots as the interims? Why have two different releases on two different schedules? Here’s a scenario: LTS 14 is released. Three months later snapshot 14.1 is sent out. Three months (or whatever time-frame that’s decided) later the next snapshot 14.2 comes out. New users, who grab the iso of 14.2 have the same support channels and length of support as the original users of 14 at this time, which with six months gone by since 14 was released, it’s four-and-a-half years now. Each new user that grabs the latest LTS snapshot iso has support that lasts until the LTS end of life. When the next LTS comes out, users who started with snapshots will have a choice: stay with the current LTS and it’s support with reduced updates, or jump on to the next LTS (which should be easier with the current one closer in makeup from all the updates).

    I think if you’re going to try and keep the LTS’ fresher with updates, they could be the “interim” releases, there’d be no need for any other. I would think that at the least it would make support easier as there’s only the LTS’ to support. It should also make packaging and repository maintenance easier as everything is either for the LTS or the rolling.

    You’d have the rolling for those who just have the need to be on the cutting edge (if not actually bleeding), who would greatly enlarge the pool of people fire-testing the latest. You’d have a LTS that would be more attractive to users as well as vendors because it’s getting regular updates and staying pretty current. And you wouldn’t need to support an interim for any length of time.

    Just some thoughts…

  46. robin werkman says: (permalink)
    March 13th, 2013 at 3:16 am

    I think a rolling release would be great, I hate after having every thing setup like I want it, updating it,nthat then I have to reinstall from scratch again becuse theres a new release.

  47. Ante Bajto says: (permalink)
    March 13th, 2013 at 8:25 am

    Dear Mark,

    rolling release is good way of thinking, but you must be prepared on situations similar to one with 12.04 to 12.10 upgrade, when ATI graphics owners (some nVidia, too) lost graphics or Unity launchers, and could not work anymore without obscure workarounds, and rollback was not possible, so users had hard time recovering their systems and data. Since Ubuntu is aiming on wide population average users, I am really scared what will happen in situations like this one, if happen again. Catastrophy.
    Furthermore, I need to add another point. I changed my hardware since then (believing that new graphic card will be better supported), and, guess what? Only opensource graphic driver is working with my APU (AMD A10 with HD7660D). No more gaming, even on Linux native games! And ATI propriatery graphic drivers (I tried to install it both manually and automatic) provided no Unity launcher. Ok, you can blame ATI for not opening hardware. But on the other hand, it is logically to expect that hardware that worked well on older version of operating system MUST! work on newer version. Especially when rolling released.
    So, something is wrong in your approach.
    I mean, do you know how user must feel when you say to him that his hardware is not good for your operating system. In this situation, I think that your operating system is not good for user (and his hardware). And I pull 2 bottom line arguments:
    Windows is good for my hardware.
    Mint is good for my hardware.
    Ubuntu is not good for my hardware.
    Conclusion?
    I am not sure when, or will I ever return to operating system that failed on me.

    Trust me, I am Ubuntu fan, I installed it on many machines, my house is all Ubuntu. I appreciate what you are doing (from phones, tablets, laptops, PCs, especially on gaming – Steam – i think it is right way to make user base wider), but I think you went too far too fast. Smaller steps, like rolling releases are great, but please, make sure that it works on our machines.

    Final thoughts: I just need Mark to see this comment, please. If you find it not appropriate to be listed in your blog comments, feel free not to broadcast it.

    Regards,
    Ante Bajto
    Croatia
    bajto.ante@gmail.com

  48. Ampers Taylor says: (permalink)
    March 13th, 2013 at 10:26 am

    Overall I think the whole approach to this is very sensible. When people work for love it is so important not to make changes unless the majority is with you. The problem always being, you will lose people whatever you do. Perhaps losing implacable or excitable people who won’t work with consensus in the long run might not be so bad.

    The only really personal issue I have work be for eight months support as I only move up after a release has been out two months. But then, I use Ubuntu Forum for support, so that’s no real problem.

    Ampers.

  49. Ubuntu: Shuttleworth dà spunti sul ciclo di sviluppo | oneOpenSource says: (permalink)
    March 13th, 2013 at 10:31 am

    [...] di Shuttleworth, che potete leggere integralmente nel suo blog, sembra aprire qualche spiraglio e propone soluzioni interessanti per mantenere allo stesso tempo i [...]

  50. ed says: (permalink)
    March 13th, 2013 at 12:22 pm

    It’s all about applications, on my Nexus 7, phone, or (windows if I used them :-) ) the moment new application is released, I get an update or am able to update instantly without any problems. So, if new amarok, digikam, office, numpy or ipython are out I want them on my system. This should be the norm, I can’t wait for 6 months. I don’t care if I’m running the current kernel or 2 years old one, the same goes for X server, Mir or Wayland. But applications must be new.

  51. Ubuntu, Shuttleworth & rolling releases says: (permalink)
    March 13th, 2013 at 3:14 pm

    [...] anonical, Ubuntu Linux’s parent company, has an ambitious plan with a short time-frame: One operating system for computers, smartphones, tablets and TVs by early 2014. One problem with this is how do you get there fast enough and one answer, rolling releases, has got developers upset. Now, Canonical and Ubuntu founder Mark Shuttleworth has a new proposal on how to handle rolling releases. [...]

  52. fhf says: (permalink)
    March 13th, 2013 at 4:55 pm

    In some ways its good. Drop interim releases or treat them like BETA or something. My previous school introduced Ubuntu but not stable LTS 12.04 but 12.10 and they will get no updates for 13.04′s Unity (as you said in yours post). LTS should be standard and interim releases only BETAs for developers etc.
    Rolling from LTS to LTS! Ubuntu forever! :)
    Bests,
    fhf

  53. Jan Goulding says: (permalink)
    March 13th, 2013 at 8:58 pm

    I’m agnostic about rolling releases – I just want me system to do what I want it to do when I want it to do it. If this means that I have to wait a few months for an upgrade then, okay. If I want an app that’s at the raw bleeding edge I can get the code for the developers site (or sourceforge) and install it before the update is in the distribution repositories. Sadly since the release of Unity this has not been the case. The fact that since upgrading to unity my PC has done an impersonation of a snail on valium to the point that I can go away and make a cup of tea whilst waiting for the browser to open (on a machine that is an 64 bit quad core with 1Gig of ram), due to badly integrated drivers tells me that Canonical have lost the plot. That the work round of launching Unity into 2D mode to speed things up was completely removed before the driver issues were resolved, leads me to believe that Canonical are entering a Microsoft frame of mind: “it’s not a bug, it’s a feature” and that is my issue with Canonical right now – not how they issue releases.

  54. John Mills says: (permalink)
    March 13th, 2013 at 10:06 pm

    Mark,

    If you read this think about the utility for an end user who just wants a desktop that works and is up to date with the applications they use. The current point releases are in essence just point releases (beta) for the LTS anyway. By using a point release for the LTS something akin to a service pack in Windows you add value to that release over a long period of time. If you want a stable release with new versions of your applications in Ubuntu you have very little choice but to resort to using PPAs that are potentially unstable and break your system. There is no guarantee either that upgrading from release to release will result in a system that is not broken.

    As others have said using a semi rolling release is much more sensible and valuable for a typical desktop user. The separation of applications and the rest of the system should have been implemented years ago. Microsoft nor Apple support third party applications they have no control over and I don’t see why canonical must either. The software centre would work in a much better way if application developers uploaded the latest version of their software and made it available to all through this mechanism. PPAs in my experience have caused me significant issues and I would much prefer to see my updates come through a single interface in the same way as the Play store or Itunes. I would dearly like to see the Software Centre in Ubuntu work in a similar way. After all this is what the majority of users would expect having used a modern mobile phone or operating system. For a good few years now the subject of application updates in Ubuntu has been discussed. Infact there is a critical bug report that is assigned to MPT regarding this very point.

    https://bugs.launchpad.net/ubuntu/+source/software-center/+bug/578045

    Please Mark consider the semi rolling release model and put some ‘wait’ behind the idea that applications be separated from the core Ubuntu desktop. If this is done there will unquestionably be a big uptake in the number of desktop users for Ubuntu and potential revenue stream for Canonical.

    Best regards,

    John

  55. Donald says: (permalink)
    March 14th, 2013 at 1:57 am

    Cool, maybe you will manage to reclaim some power users from the clutches of Arch Linux.

  56. John Mills says: (permalink)
    March 14th, 2013 at 6:19 am

    Mark,

    If you read this think about the utility for an end user who just wants a desktop that works and is up to date with the applications they use. The current point releases are in essence just point releases (beta) for the LTS anyway. By using a point release for the LTS something akin to a service pack in Windows you add value to that release over a long period of time. If you want a stable release with new versions of your applications in Ubuntu you have very little choice but to resort to using PPAs that are potentially unstable and break your system. There is no guarantee either that upgrading from release to release will result in a system that is not broken.

    As others have said using a semi rolling release is much more sensible and valuable for a typical desktop user. The separation of applications and the rest of the system should have been implemented years ago. Microsoft nor Apple support third party applications they have no control over and I don’t see why canonical must either. The software centre would work in a much better way if application developers uploaded the latest version of their software and made it available to all through this mechanism. PPAs in my experience have caused me significant issues and I would much prefer to see my updates come through a single interface in the same way as the Play store or Itunes. I would dearly like to see the Software Centre in Ubuntu work in a similar way. After all this is what the majority of users would expect having used a modern mobile phone or operating system. For a good few years now the subject of application updates in Ubuntu has been discussed. Infact there is a critical bug report that is assigned to MPT regarding this very point.

    See launchpad bug : 578045

    Please Mark consider the semi rolling release model and put some ‘wait’ behind the idea that applications be separated from the core Ubuntu desktop. If this is done there will unquestionably be a big uptake in the number of desktop users for Ubuntu and potential revenue stream for Canonical.

    Best regards,

    John

  57. Andy S says: (permalink)
    March 14th, 2013 at 9:21 am

    Mark, very good proposals. I have one worry which I don’t think other commenters have addressed – experienced end users installing Ubuntu on inexperienced end-users’ computers (for example me, in Northern Ireland, installing on my mother’s computer, in Scotland). Currently the Ubuntu website points straight to the interim release for downloads, unless you go looking for the LTS. So my mum ends up with 12.10.

    Now it’s one thing to expect the inexperienced user to run updates, but an upgrade is a different process which I wouldn’t expect that user to do. However 7 months support is not enough for this situation to work.

    The main reason my mum (and presumably many other people like her) use and like Ubuntu is the feeling of safety online. Realising they are no longer receiving updates and are out of support would probably make her lose much of that confidence.

    I’m not suggesting that more than 7 months is required (cutting it back seems like a very fair move), but the website should be changed to point primarily to the most recent LTS, specifically so that inexperienced users don’t end up routinely unsupported.

  58. Fred R. says: (permalink)
    March 14th, 2013 at 9:59 am

    “separating apps from core.”
    That’s what is happening in the BSD world from ages, and applications are called ports…
    The tricky point would be to define what is called “core”, as kernel dev is separate from basic utils (sorry, no troll here :-))
    core = kernel+gnu utils ? Everything from grub/kernel to Unity ? Adding only “official apps” ?

    In the past months/years, Canonical took some strong/courageous/controversial (pick what you think…) decisions.
    Maybe the future of Ubuntu is grub/kernel/upstart/Mir/Unity/selected GTK apps and leave all the rest in separate repositories ???
    For sure, maintaining a rolling process, with such a reduced amount of packages, should be much mode doable !

  59. Link: Mark Shuttleworth über die Zukunft der Updates in Ubuntu. :: 1337core says: (permalink)
    March 14th, 2013 at 12:13 pm

    [...] Mark Shuttleworth über die Zukunft der Updates in Ubuntu. [...]

  60. André says: (permalink)
    March 14th, 2013 at 12:48 pm

    I agree, been waiting 6months for a fix to this problem, what happened about preserving working samba ? https://bugs.launchpad.net/bugs/1075923

  61. LinuxLife Blog » News: Shuttleworth: LTS-Versionen von Ubuntu mit mehr Updates says: (permalink)
    March 14th, 2013 at 4:16 pm

    [...] dass er ein Rolling-Release-Modell ablehne, und deutete einige mögliche Änderungen an. In seinem aktuellen Blog-Eintrag wird er jetzt konkreter und stellt einen neuen Vorschlag [...]

  62. MattiK says: (permalink)
    March 14th, 2013 at 7:44 pm

    Great! This is right way to go.

    Do not have to reinvent the wheel again, see: http://en.wikipedia.org/wiki/Open-system_environment_reference_model

    I see Ubuntu here as “Application Platform”, and this is simple for end users and developers.

    What is necessary, is to define those interfaces and backward compatibility/support policy. Two release backward compatibility should be enough. Next step is to get certification.

    And what are those interfaces?

    My opinion is that for compatibility reasons there should be full support for LSB interfaces, Java and PHP. And then you need some preferred native API. My opinion, it should be Qt 5 based on platform scope (touchscreen, sensors etc.).

    If you are supporting LSB, prepare to support X Window system next 10 years or something, even if X is deprecated in standard.

    LSB may be not necessary to support any more in future if the standard doesn’t updated. If the standard is updated and X is deprecated, it need to be supported next ten years. If more drastic approach are taken andLSB support is dropped in Ubuntu, then in future it is nice if the platfrom is “cleaned” to support only Qt5, Java EE and Python 3 top of *nix. Of course, default applications and most used “killer apps” should be rewritten then.

  63. Dragos says: (permalink)
    March 15th, 2013 at 6:35 am

    I am only interested into LTS. I still have a 8.04 LTS production box running :))

    You should only produce LTS every 2 years and support them for 7 years: 3 years + 2 of extended support1 + 2 of extended support2.

    This will help Ubuntu into business customers.

  64. Alex says: (permalink)
    March 17th, 2013 at 1:01 am

    Hi Mark,

    Thanks for everything you have done to make Ubuntu so great. I know there’s been a lot of drama and controversy lately but I’d like to let you know that I highly appreciate what you are doing and I’m sure there are many others who feel like me. Keep up the awesome work.

  65. Clive says: (permalink)
    March 17th, 2013 at 2:53 pm

    Mark,

    This strategy makes a lot of sense – I applaud you for it. However, one request of you please… Would you be kind enough to institute a new “roll-up” process when a release [let's start with LTS] comes to the end of it’s support window. Basically, take the original installation ISO and update all the packages with the “current” editions that you’ve patched over the lifespan of that release. The idea would be, for those of us who have to continue to support a release, that you’ll be saving us excessive downloads and updates [and taking load off your servers in the process].

    If you feel like a stretch goal, how about an ISO-building utility that would let me “roll my own ubuntu” DVD ISO, adding those packages I use most frequently and creating a new ISO image. This would allow me to make a customised ubuntu that contains what I need, and again, takes load off your servers.

    Nice idea thought…

    Thanks

    Clive

  66. sujal says: (permalink)
    March 18th, 2013 at 11:54 am

    Hi Mark,

    I am an end user. If you wish to have total control, then you will have to make everything in-house. This will help you to get a clear picture, decide future direction and increase confidence when making deals ;)

    Ubuntu is good and canonical taking good direction. For people who do not like Unity, there is no-nonsense traditional Xubuntu. Mir is good initiative and will surely help canonical. I think cutting down support period will help a lot. I also suggest

    1) Have only one release per year. So next version has a lot of new stuff, to upgrade. Also devs will have enough time to think, implement and innovate.

    2) No rolling release please. Nothing happens if you are not using latest software. So I do not like rolling release. I do not like daily updates. I want my PC to be stable and forget about base system so that I can work properly.

    3) Have a dev release after RC and then wait for 1 month for bug fixing. Then release it for common public – end users. An extra layer will give more confidence to non-technical end users. So End User version may be 12.0.1

    4) Yes to LTS and Point Releases, ISO respins.

    Thanks and Regards

  67. Momist says: (permalink)
    March 18th, 2013 at 11:22 pm

    As an ordinary “consumer”, I would welcome better quality in the LTS releases. I moved away from plain Ubuntu on the upgrade from 10.04 to 12.04 as one of those who found the jump into unity, together with loss of support for older hardware, a step too far. I understand that the updated versions now available will have resolved those issues with a quicker unity and better hardware support, but those features should have been in the 12.04 LTS at the outset.
    I would also welcome the chance of updating apps within a supported LTS.
    I also agree with those who suggest that the LTS should be better promoted, with the latest/greatest being offered with the caveat that it will only be there for a limited time.

    Keep up the good work, whichever direction it goes!
    Momist

  68. historyb says: (permalink)
    March 19th, 2013 at 12:26 am

    I like the idea of a rolling release.

  69. Ubuntu: Kürzerer Support für Zwischen-Releases | virtualfiles.net says: (permalink)
    March 19th, 2013 at 3:09 pm

    [...] der Diskussion ist ein Vorschlag von Mark Shuttleworth für ein überarbeitetes Release Management. LTS-Releases sollen in den ersten zwei Jahren ihres [...]

  70. Dave says: (permalink)
    March 20th, 2013 at 8:30 am

    I would appreciate Ubuntu to go back on focussing, that things work, rather than making too big changes. A stable Ubuntu with the newest programs in the repo is always better than totally renewed one with a lot of unfinished or badly supported features

    -the ATI fglrx drivers are causing problems since the switch to unity… there are tutorials in the web on how to fix this, why not let it work out of the box
    -Samsung printer drivers still don’t seem to work
    -Performance has drastically dropped in 12.10

    More than that tho, a how about a better upgrading system. Generally I would love to always have the newest distro, but it is well known that the upgrades don’t work as well as clean installs. I would love to have an upgrade option, where Ubuntu would backup the user’s Home directory (along with mailboxes and programm settings) and makes a list of installed programms. Ubuntu should then wipe the old installation, restore the home directory and reinstall the programms on the list (if found in the new repos).

  71. Ubuntu to halve support length for non-LTS releases says: (permalink)
    March 20th, 2013 at 12:42 pm

    [...] Developer Summit (UDS) in March; project leader Mark Shuttleworth commented on the suggestion and put forward a “strawman proposal” on his blog. The decision by the distribution’s Technical [...]

  72. bsniadajewski says: (permalink)
    March 21st, 2013 at 11:29 pm

    Hello!

    I have been using Ubuntu variants (mainly Kubuntu) since 9.04 and Linux in general since 2005 (started with Suse and Mandrake). I really like your proposal here. I have a couple ideas though:

    - Create a “pointer” in the repositories that always points to whatever the current interim and lts releases are at the moment. Opensuse has something like this, called “OpenSuse_Current” I believe.

    - make the universe (at least) repo “rolling” as in the way you’d have the applications be updated in your proposal. This would untie the universe repo from the release scheduling and allow those who are, say, using Kubuntu 12.04 LTS to have access to the latest KDE without having to add PPA’s, which could cause problems when it comes time to upgrade.

    Thanks,

    Brandon

  73. Cosas que hacer ANTES de instalar Ubuntu 13.04 #1: Copia de seguridad de las PPA | Linux-OS.net says: (permalink)
    March 24th, 2013 at 12:19 am

    [...] que las versiones LTS tendrán cada vez más protagonismo en el futuro. El mismo Mark Shuttleworth ha confirmado esta nueva tendencia, que supone el refuerzo de las actualizaciones parciales de las LTS.  Y [...]

  74. kikl says: (permalink)
    March 24th, 2013 at 4:27 am

    “Separating platform from apps would enhance agility.

    Currently, we make one giant release of the platform and ALL APPS. That means an enormous amount of interdependence, and an enormous bottleneck that depends largely on a single community to line everything up at once. If we narrowed the scope of the platform, we would raise the quality of the platform. Quite possibly, we could place the responsibility for apps on the developers that love them, giving users access to newer versions of those apps if (and only if) the development communities behind them want to do that and believe it is supportable.”

    One of the major reasons for dumping an LTS for an interim release is the fact that the apps on the LTS are not up to date. People would stick to long term releases much longer if they could easily keep the software up to date. I think at least well maintained key applications like libre office should be upgraded on long term releases – say for 2 years until the next LTS is published.

    “Very few end users depend on 18 months support for interim releases. The proposal is to reduce the support for interim releases to 7 months…” I think the direction is good, but 7 months appear to be too short. Most people want to wait 1 or 2 months until the release stabilizes. Consequently, I would favor a support cycle of 9 months.

    I think you should keep the 6 month cycle. It is dependable and creates reasonable dead lines for developers.

  75. wulan says: (permalink)
    March 24th, 2013 at 7:17 am

    Hi,Mark, I have heard that ubuntu has been choosen to be China’s standard OS recently.
    In fact,though I like ubuntu for such these years,but I really disappointed this time.
    Open source has dark future in china I think, besides, to speak straightly,I really don’t want to see spywares in ubuntu, if there is.
    Ubuntu is for human beings, I understand the dificulties in developing and spreading it, however, I prefer to make my donation to ubuntu instead of seeing it donated by some “devil” goverments…

    that’s all ,thank you.

  76. Gjermund G Thorsen says: (permalink)
    March 24th, 2013 at 8:41 am

    How can I buy shares in Canonical Ltd?

  77. Duncan Murray says: (permalink)
    March 25th, 2013 at 11:59 pm

    I agree with most of those premises Mark. As an Ubuntu user since 2009, can I just say a big thank you for all the work you, Canonical and the community has put into making a terrific distribution. I think the last idea of separating platform from apps would be great. I don’t understand the workings of linux/ubuntu (as it is nothing to do with my day job), but I always wondered why you couldn’t give the keys to the repos to trusted people in Libreoffice/GIMP etc… But however you do it, updating the LTS applications would be really superb. Keep up the brilliant work please!

  78. Bhaskar says: (permalink)
    March 27th, 2013 at 3:33 pm

    How about abandoning non-LTS release altogether and focus on LTS releases? I think that will be brilliant. You guys will be able to concentrate your effort within a longer span of time. Bug fixes, software and security updates will be provided. Are the non-lts releases necessary?

  79. Davor says: (permalink)
    March 28th, 2013 at 7:55 am

    Update apps and kernels for LTS – i was always wondering why people must stuck for 3 years with old repos.
    Excellent idea.

  80. foo says: (permalink)
    March 28th, 2013 at 1:10 pm

    “Change is warranted. If we want to deliver on our mission, we have to be willing to stare controversy in the face and do the right thing anyway, recognising that we won’t know if it’s the right thing until much later, and for most of the intervening time, friends and enemies alike will go various degrees of apoplectic.”

    As a long-time Linux user I think I am on the friendly side.

    I agree that “we have to be willing to stare controversy”, but if you don’t recognize when your friends go apoplectic, you may end up with a situation similar to Windows 8, where 90% of the users dislike the new interface.

    The only thing is that Microsoft is an stablished player with 90% of market share, and can afford to make mistakes. This is a privilege that new entrants, like Linux, don’t have.

    And here’s what I think are the two big problems with Ubuntu right now:

    1) Radical changes that alienate old users

    2) A general bad taste in design

    I think both items could be solved if Canonical acquired Linux Mint, and gave them the status of designers of Ubuntu.

    The people behind Linux Mint demonstrated, year after year, that they can excel in design.

    With their talent, plus Canonical resources, we could have an operating system that would rival Mac OS X — which I also happen to use and think is overrated.

  81. foo says: (permalink)
    March 28th, 2013 at 1:21 pm

    @Franck “What about delta updates? That Ubuntu takes 100x the bandwidth of Windows for security updates is such a major embarrassment”

    I think you are comparing apples to oranges.

    Windows is a barebones operating system, with no decent drawing programs (GIMP) or office suite (LibreOffice).

    If you add MS Office to Windows, you’ll probably make the updates 10 times bigger.

    “due to stone age apt-get that redownloads half the distribution every time”

    I love apt-get, because it allows most system updates to be performed without a single reboot. Compare that to Windows, which requires several reboots per update.

    As you can see, there are pros and cons, and while you don’t like apt-get, many people enjoy its power and flexibility.

  82. Allan McRae » Blog Archive » Interesting Links – March 2013 - One day this will feature a witty tagline… says: (permalink)
    March 31st, 2013 at 3:58 am

    [...] also looked at rolling releases, but decided not [...]

  83. Noah says: (permalink)
    March 31st, 2013 at 9:42 pm

    Mark, thank you for Ubuntu!
    I love it! :P

  84. Kenny Strawn says: (permalink)
    April 1st, 2013 at 10:35 pm

    So I take it the “rolling release” branch will in fact be Ubuntu’s Rawhide?! Yeah, it makes ALL the sense in the world…

  85. Anoop says: (permalink)
    April 2nd, 2013 at 9:23 pm

    I agree with making updated applications available for LTS releases and advertising LTS as the recommended Ubuntu system for new users.

  86. nemekan says: (permalink)
    April 3rd, 2013 at 1:58 pm

    This strategy makes a lot of sense – I applaud you for it. However, one request of you please… Would you be kind enough to institute a new “roll-up” process when a release [let's start with LTS] comes to the end of it’s support window. Basically, take the original installation ISO and update all the packages with the “current” editions that you’ve patched over the lifespan of that release. The idea would be, for those of us who have to continue to support a release, that you’ll be saving us excessive downloads and updates [and taking load off your servers in the process].

    If you feel like a stretch goal, how about an ISO-building utility that would let me “roll my own ubuntu” DVD ISO, adding those packages I use most frequently and creating a new ISO image. This would allow me to make a customised ubuntu that contains what I need, and again, takes load off your servers.

    Nice idea thought…

  87. Dave Anderson says: (permalink)
    April 3rd, 2013 at 9:26 pm

    I too like the idea of splitting the apps out. My Ubuntu updates very regularly and I don’t have any problems with it, but as things get larger, let people have more choice about updating important core stuff alone. And thanks for the excellent product

  88. Ike ahloe says: (permalink)
    April 8th, 2013 at 10:04 pm

    Ubuntu has always needed to separated applications form the release cycle. I come from the art-user area, which is an area that is not interested in being a power user beyond making sure they have the newest programs and features. I could never get any of my artist friends to switch because having to deal with PPAs just to use a a new stable release of something is totally a deal breaker (not for me but I’m a FOSS enthusiast). People want to be able to enjoy the release of the new stable product the day (or at least week) it comes out hassle free. a dual system of official repo with shipped software, plus application bundles for third-party programs would fix this.

  89. SonikkuAmerica says: (permalink)
    April 12th, 2013 at 2:13 am

    I would agree with kikl (above) on the idea of supporting 9 months’ worth of interim releases, primarily because Ubuntu is the 3rd most popular OS in the world, and every 6 months I would imagine there would be a 30-day window in which the Ubuntu distro servers would be overrun by users trying to upgrade (not to mention getting punted from the server in the process because of the amount of hits). Extending the proposed release support from 7 months to 90 days seems to be the way to go.

  90. SonikkuAmerica says: (permalink)
    April 12th, 2013 at 2:15 am

    And, by extending from 7 months to 90 days, I meant to say extending the 7 months to 9 months.

  91. Craig says: (permalink)
    April 15th, 2013 at 4:00 pm

    Rick Spencer? Sorry, never heard of him. He’s no Lennart Poettering.

  92. Ubuntu 13.04 (Raring Ringtail) says: (permalink)
    April 25th, 2013 at 1:23 pm

    [...] (in order to minimise critical pre-judgment of the release), suggestions about changing to a rolling release for the non-LTS desktop version, and hints from Mark Shuttleworth about exciting new features, [...]

  93. mark says: (permalink)
    April 25th, 2013 at 3:34 pm

    Too damn true.