- Site menu:
- Latest Discussions
- Latest Newswire
- Archives
- Submit Story
- Story HOWTO
- Syndicate
- E-mail Alerts
Shuttleworth wants to support Debian
|
|
|
Comments (71)
51 earlier comments in the full thread.
I agree with Sander that there is quite some value for Debian in this approach. In any case something that is good for Ubuntu users is also of value for Debian and linux in general. I don't think the situation with Mint is anywhere comparable. Overall if Mark Shuttleworth's suggestion of getting major distros and major upstream projects in some synchronization is realized then it will be good for linux.
Quoting: shouldnt and doesnt it make more sense for downsteam to change and adapt than upstream to be forced to change and accommodateActually Debian's unpredictable release cycle was one of the major motivations for the Ubuntu "fork" rather than building on top of it, so I do not see how Ubuntu can change to unpredictable releases. Many Debian users started looking at Ubuntu for precisely this reason.
I dont disagree that the varying release schedule was a contributnig factor to the fork, or that people started looking towards Ubuntu for precisely the same reason (among others). What I do disagree with is putting pressure on upstream to address these issues. They knew where they came from, and why the fork, if Debian were to take on Ubuntu's policies then the only thing that would be left for Ubuntu is .... nothing but a name.
Changing upstream when changes to upstream wont change downstream distribution one lick under the guise of cooperation is false logic. Shuttleworth is pushing his dream of lockstep standardization but phrasing it politely. The sad part is, he wants every to change to suit him but wont even consider making the same adjustments he is requesting of others.
Changing upstream when changes to upstream wont change downstream distribution one lick under the guise of cooperation is false logic. Shuttleworth is pushing his dream of lockstep standardization but phrasing it politely. The sad part is, he wants every to change to suit him but wont even consider making the same adjustments he is requesting of others.
I imagine that Ubuntu wants to draw on Debian Testing rather Unstable because that would result in the eventual Debian Stable and Ubuntu LTS having the same version of the kernel.
Shuttleworth's contention is that having releases across the distro spectrum (including Fedora, Suse, etc.) that offer the same version of the kernel, and perhaps other packages as well, would help the whole FOSS ecosystem by providing at least the option of uniformity for a given release from many distros.
I imagine that it's hard for both developers and hardware manufacturers to deal with the cacophany of distros in the Linux world, with the prospect of getting a given driver in a more-popular kernel a lot more enticing tnan the current state of affairs (many kernels being supported in many releases, all with potentially different versions of the same packages ...).
I don't know if this is ever going to work, but I think a common kernel (or at least a very close one) for Debian and Ubuntu could only help with hardware configuration.
Shuttleworth's contention is that having releases across the distro spectrum (including Fedora, Suse, etc.) that offer the same version of the kernel, and perhaps other packages as well, would help the whole FOSS ecosystem by providing at least the option of uniformity for a given release from many distros.
I imagine that it's hard for both developers and hardware manufacturers to deal with the cacophany of distros in the Linux world, with the prospect of getting a given driver in a more-popular kernel a lot more enticing tnan the current state of affairs (many kernels being supported in many releases, all with potentially different versions of the same packages ...).
I don't know if this is ever going to work, but I think a common kernel (or at least a very close one) for Debian and Ubuntu could only help with hardware configuration.
I do see the point that Debian doesn't want to be bound by a 2-year time frame to freeze Testing, but users could potentially benefit by knowing that Debian is at least on some kind of schedule, even if the time between freeze and release is not set.
As I said, overall Shuttleworth's ideas seem to be beneficial to linux as a whole IMO, so an opposition to change only because it will benefit Ubuntu makes little sense. There have been several DDs (non-Ubuntu employees) also expressing the opinion that it will benefit their packages since they can plan better for the features/version for the time-based freeze, so its not a one-way street as it has been made out to be.
Quoting: The sad part is, he wants every to change to suit him but wont even consider making the same adjustments he is requesting of others.In the current context I have yet to see anything reasonable being asked of Ubuntu. Shuttleworth has said that he is willing to either provide resources to meet the december freeze or push ubuntu release by a quarter but not both, so it cannot be said that he is not willing to make adjustment (in fact, now that even one of the RMs has stated that december freeze seems unrealistic, Shuttleworth may be forced to take the second option).
Quoting:One problem with this thought, Ubuntu releases every six months, for them to use Testing as a base for LTS they would essentially be releasing the same release as the one prior to LTS.
In package versions, yes. But the LTS base would be a whole lot more stable.
The difference between Debian testing and unstable is about 10 days when the release nears. After all, Debian testing is just the packages from Debian stable that have no new bugs reported against them in 10 days and which have all dependencies satisfied. By the time a Debian release nears, all dependencies will have been satfisfied and few new bugs are filed.
So, the Ubuntu release prior to LTS would be based on Debian near it's freeze, while the LTS itself would be based on testing just prior to release. It would mean that Ubuntu LTS has a stability that is pretty much as good as a Debian release. Just what you'd want for an LTS.
Also, don't forget that Ubuntu sources several large packages directly from upstream and not from Debian. Gnome for example comes directly from gnome.org and not from Debian. So these versions will change between LTS and it's prior release.
Quoting:Do you see the issue of demanding that...I see nobody making any demands in the "cadence" debate. I see a proposal. Mark Shuttleworth can do nothing to force the hands of the Debian developers; they can accept. reject, or counter-propose (I won't lay odds on which).
Contrast Microsoft, who will threaten you with court action if you don't abide by their interpretation of their licensing terms.
On the kernel issue being standardized, there is already an accepted standard kernel with regard to the corp-rat environment, that being 2.6.18 . The standard kernel argument that is being used here is rather the one of the latest greatest most bling kernel which happens to change with some regularity. Even on a 6 month release cycle the differernce of a few weeks in either direction can vary the kernel by 3 major revisions and/or multiple minor revisions.
Kernel argument == moot.
Gus can you tie EULA restrictions to release dates somehow? That might make that part of your argument relevant.
Sander I see the point and while it does make sense in that light, other than a few GUI click the pretty button configuration utilities and a paint job, whats to sell Ubuntu LTS over Debian stable? Debian Stable is notoriously more stable than anything that has ever come from canononical.
Kernel argument == moot.
Gus can you tie EULA restrictions to release dates somehow? That might make that part of your argument relevant.
Sander I see the point and while it does make sense in that light, other than a few GUI click the pretty button configuration utilities and a paint job, whats to sell Ubuntu LTS over Debian stable? Debian Stable is notoriously more stable than anything that has ever come from canononical.
Quoting:Gus can you tie EULA restrictions to release dates somehow? That might make that part of your argument relevant.The only thing in common is that you are using hyperbole to attribute behavior to Ubuntu/Canonical that would clearly not be hyperbole, but rather Standard Operating Procedure, were Microsoft the subject at hand.
2.6.18 has been very, very good to me. I believe that was the kernel in Debian Etch, PCLinuxOS 2007 and RHEL 5. I'm sure that wasn't planned ... but how did it happen?
Who among us hasn't seen a huge bug in Debian Stable that, due to the distro's "stability," will never be fixed? Same for Ubuntu ...
That was around the time a filesystem corruption bug cropped up in 2.6.19. Word got out real fast that 2.6.18 was a safer bet. The first release of 2.6.20 was a new branch that release managers didn't want to use without broader testing.
http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.19... (Search for "subtle race".)
http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.19... (Search for "subtle race".)
gus, MS is not the subject at hand your the one who has tried to draw the correlation. The only one mind you, I think your prejudices are showing.
Steven if you look closer you'll also see that t this point 2.6.18 also recieves the largest amount of backports. Many people just go with the flow or demand the latest greatest, but there is a standard and maintained that is out there, it may be unintentional but this is usually what ever RHEL is running at the time.
Gentoo even keeps the backports to 2.6.18 current, there are others I am sure.
Steven if you look closer you'll also see that t this point 2.6.18 also recieves the largest amount of backports. Many people just go with the flow or demand the latest greatest, but there is a standard and maintained that is out there, it may be unintentional but this is usually what ever RHEL is running at the time.
Gentoo even keeps the backports to 2.6.18 current, there are others I am sure.
Yes, I did bring Redmond into the conversation. I guess the point wasn't obvious enough.
Microsoft "standards" change, on a regular basis (OfficeOpenXML, anyone?). For a long time, they didn't care, because they assumed they had a ring in the public's nose. In a way, they did, but the public revolted as soon as the resentment hit critical mass. Between the swiss-cheese security and questionable upgrade policies, it was only a matter of time...
The FOSS philosophy is the complete opposite of that. My way, our way, no way, whatever. The limits of Linux's adaptability have yet to be discovered, and there's nothing in the pertinent GPL's to interfere with that process.
Microsoft hands down mandates From On High, and demands that Windows users either toe the line or not use the software (but good luck getting your M$ tax back).
FOSS negotiates. Sometimes rather boisterously, as this thread demonstrates. But even if two parties can't agree on whether A or B is more correct, neither party can stop the other from implementing A or B. Or both. Or neither.
And so it is with the "cadence" debate. Shuttleworth has raised an issue, as he sees it, and proposed two possible solutions (both at Canonical's expense) to the Debian team. They can accept one, or the other, or tell him to Papa Oscar. But nobody in this dust-up can force anybody else's hand.
As for my prejudices against Microsoft, I've never hidden them, at least on LXer.
Microsoft "standards" change, on a regular basis (OfficeOpenXML, anyone?). For a long time, they didn't care, because they assumed they had a ring in the public's nose. In a way, they did, but the public revolted as soon as the resentment hit critical mass. Between the swiss-cheese security and questionable upgrade policies, it was only a matter of time...
The FOSS philosophy is the complete opposite of that. My way, our way, no way, whatever. The limits of Linux's adaptability have yet to be discovered, and there's nothing in the pertinent GPL's to interfere with that process.
Microsoft hands down mandates From On High, and demands that Windows users either toe the line or not use the software (but good luck getting your M$ tax back).
FOSS negotiates. Sometimes rather boisterously, as this thread demonstrates. But even if two parties can't agree on whether A or B is more correct, neither party can stop the other from implementing A or B. Or both. Or neither.
And so it is with the "cadence" debate. Shuttleworth has raised an issue, as he sees it, and proposed two possible solutions (both at Canonical's expense) to the Debian team. They can accept one, or the other, or tell him to Papa Oscar. But nobody in this dust-up can force anybody else's hand.
As for my prejudices against Microsoft, I've never hidden them, at least on LXer.
That does claryify, thank you.
My position on this remains relatively unswayed. The entire premise for mated release cycles between Debian and Ubuntu is based on false logic. If the two meet at an approximate 2 year cycle between Debian stable (or the package freeze leading to it) and Ubuntu LTS, then short of a paint job, Ubuntu LTS becomes superfluous. Debain enjoys a reputation on stability that Ubuntu does not. For Ubuntu to maintain relevance in that contest it has to accept the stability reputation that it has earned on its own and attempt to surpass Debian in other areas.
The common kernel across distros, point moot, there is a corp-rat standard and Ubuntu has chosen to ignore it. Basing Ubuntu LTS on Debian pre stable freeze, point moot, Debain walks away with the stability crown nearly uncontested. Common package revisions across $DISTRO, both distro's lag behind even other binary distros.
The whole of this is, Shuttleworth wishes a homogeneous across the board environment. If you wish to stick with the Microsoft analogy, they have proven that this creates a lesser product not a greater product. Linux greatest advantage is its diversity, where someone like Ubuntu can come along and challenge the norm. dragging others kicking and screaming along with them if they wish to meet the bar. The homogeneous environment can only work for the established players, as it never encourages others to break the mold and challenge pre concieved 'truths' ar rasies the technical bar to a new level.
The whole of the argument is based upon false logic, that some will jump on the lemming bandwagon without examining the unintended or unrevealed consequences only goes to prove that there really is a sucker born every minute. The thing here is, is that there are a lot of dots that need to be connected to see the story. Getting outside of the cadence thread and seeing what he has said at other points on both similar and seemingly unrelated items shows his goals for canononcial and ubuntu. Achieve a percieved conclusive market share, and then hold on to it, period.
Never forget he is a business man first and benefactor second. That he tries to convince everyone else that his view is for the greater good of linux in general shows he is either more shrewd than I give him credit for, or is truly deluded enough to believe lockstep standards are truly adventageuos to technical progress.
My position on this remains relatively unswayed. The entire premise for mated release cycles between Debian and Ubuntu is based on false logic. If the two meet at an approximate 2 year cycle between Debian stable (or the package freeze leading to it) and Ubuntu LTS, then short of a paint job, Ubuntu LTS becomes superfluous. Debain enjoys a reputation on stability that Ubuntu does not. For Ubuntu to maintain relevance in that contest it has to accept the stability reputation that it has earned on its own and attempt to surpass Debian in other areas.
The common kernel across distros, point moot, there is a corp-rat standard and Ubuntu has chosen to ignore it. Basing Ubuntu LTS on Debian pre stable freeze, point moot, Debain walks away with the stability crown nearly uncontested. Common package revisions across $DISTRO, both distro's lag behind even other binary distros.
The whole of this is, Shuttleworth wishes a homogeneous across the board environment. If you wish to stick with the Microsoft analogy, they have proven that this creates a lesser product not a greater product. Linux greatest advantage is its diversity, where someone like Ubuntu can come along and challenge the norm. dragging others kicking and screaming along with them if they wish to meet the bar. The homogeneous environment can only work for the established players, as it never encourages others to break the mold and challenge pre concieved 'truths' ar rasies the technical bar to a new level.
The whole of the argument is based upon false logic, that some will jump on the lemming bandwagon without examining the unintended or unrevealed consequences only goes to prove that there really is a sucker born every minute. The thing here is, is that there are a lot of dots that need to be connected to see the story. Getting outside of the cadence thread and seeing what he has said at other points on both similar and seemingly unrelated items shows his goals for canononcial and ubuntu. Achieve a percieved conclusive market share, and then hold on to it, period.
Never forget he is a business man first and benefactor second. That he tries to convince everyone else that his view is for the greater good of linux in general shows he is either more shrewd than I give him credit for, or is truly deluded enough to believe lockstep standards are truly adventageuos to technical progress.
Quoting:Never forget he is a business man first and benefactor second.Yes, as I echoed in my point #4 above (my second comment). I would expect a similar approach from Bob Young or Jim Whitehurst. OTOH, if I read something like that from Patrick Volkerding, I'd want confirmation that he actually wrote it.
Quoting:That he tries to convince everyone else that his view is for the greater good of linux in general shows he is either more shrewd than I give him credit for, or is truly deluded enough to believe lockstep standards are truly advantageous to technical progress.Or both. Are they mutually exclusive?
Quoting: If the two meet at an approximate 2 year cycle between Debian stable (or the package freeze leading to it) and Ubuntu LTS, then short of a paint job, Ubuntu LTS becomes superfluous.Sure, it could and Ubuntu will have to bring in more on the plate to be relevant. So how is that a bad thing? As I see it Ubuntu gains, and overall Debian gains.
Quoting: The whole of this is, Shuttleworth wishes a homogeneous across the board environment. If you wish to stick with the Microsoft analogy, they have proven that this creates a lesser product not a greater product. Linux greatest advantage is its diversity, where someone like Ubuntu can come along and challenge the norm. dragging others kicking and screaming along with them if they wish to meet the bar. The homogeneous environment can only work for the established players, as it never encourages others to break the mold and challenge pre concieved 'truths' ar rasies the technical bar to a new level.Oh well this is an over dramatization of the situation. One can point to multiple projects where it has been a proven success (e.g. eclipse). By your logic initiatives like from freedesktop or LSB etc. are bad things. Yes diversity of included components in a distro is good, but how do you see different upstream package versions for distro releases in approx the same time frame as a good thing? The proposal, as i see, is for cooperation to bring about a more uniform linux experience (with all its diversity) much in the same vein as freedesktop etc.
Quoting:If the two meet at an approximate 2 year cycle between Debian stable (or the package freeze leading to it) and Ubuntu LTS, then short of a paint job, Ubuntu LTS becomes superfluous.
There are quite a few differences between Debian and Ubuntu. As I said previously, a lot of Ubuntu packages come directly from upstream and not from Debian, such as Gnome. These will be much newer. There are also many other differences. A different base system (upstart v.s. init) a different default shell (dash v.s bash, although IIRC Debian now switched to dash as well) , different X server and configuration. Different kernel configuration. Different default services and packages. Different tools, and not just a few GUIs thrown in. Think of Bulletproof X (which I find annoying as hell, but is really useful for non-geeks).
Ubuntu is a lot more than Debian with a bit of lipstick.
Post a comment
A free account is needed to post. Write your comment here first: your text is kept in this browser while you sign up or log in, and is waiting for you when you come back.
Your message:Markdown: **bold**, *italic*, [text](https://link), - lists, > quotes, $x^2$ for math. Formatting help
Sign up and post (it is free) or Log in and post
One problem with this thought, Ubuntu releases every six months, for them to use Testing as a base for LTS they would essentially be releasing the same release as the one prior to LTS. Being that Debian goes into package freeze on testing several months prior to a stable release.