Ubuntu 14.04 LTS Planning A Linux 3.13 Franken Kernel

[rdnetto]

Because some companies plan there server upgrade ahead of the release.

If you have a server where stability is important enough that you're running the LTS version, you should probably be waiting until a month or two after the release anyway. The last LTS is supported until 2017, so it's not like you need to worry about vulnerabilities or anything.

[Uqbar]

Useless marketing

While I do understand that the kernel is an important part of a distribution, surely among the most important ones, that debate is pointless.
Ubuntu has tons to bugs introduced by either themselves or the Debian team and none seems to be taking care. As well as the ones inherited by the original software developers.
There are bugs in Ubuntu staying there since quite long time and hindering "users' experience".
Ubuntu should focus more on the overall distribution status more than the kernel itsels, the theme colors or even the window system.
Choosing between 3.13 or 3.14 is a matter of flipping a coin. You can go 3.13 and then weeks later jump into the PI version.
That version bump is hardly going to kill a distribution, not more than the bugs they introduce in packages.
But this is just my opinion.

[Pajn]

Yeah, I don't know why they won't just delay it a little bit and release it as Ubuntu 14.06 LTS or something. Why does it HAVE to come out in April? I can tell you I just tried out Ubuntu 13.10, and it could really have used an extra few months of polish (I realize it isn't an LTS, but still). I went to 13.04 because of too many bugs.

Because companies like System 76 and Google does plan their update cycle after Ubuntu?s release cycle. Delaying the release would piss them of.

I don't really get it. They don't want to use a 3.14 kernel because they fear it will not be thoroughly tested by the time of the release, but they'd prefer to do a back port of big amounts of 3.14 code to 3.13, getting an even less tested solution. Does anybody else think that just going 3.14 makes more sense, even without the delay? I mean, they'll have the support they want, it will actually have more test coverage than a 3.13 with bock ported code.

By backporting Broadwell support only Broadwell support is untested, by using 3.14 everything is untested.

[Vim_User]

Because companies like System 76 and Google does plan their update cycle after Ubuntu?s release cycle. Delaying the release would piss them of.

The question is, what pisses companies like System76 more of: Having to wait 1-2 months for Ubuntu and having full hardware support or delivering the hardware with an OS that doesn't fully support the hardware.

By backporting Broadwell support only Broadwell support is untested, by using 3.14 everything is untested.

By using 3.14 and delaying the release 1-2 months you have all the hardware supported and a tested kernel. But that won't happen, since for Canonical the release date is more important than anything else, nowadays.

[Pajn]

The question is, what pisses companies like System76 more of: Having to wait 1-2 months for Ubuntu and having full hardware support or delivering the hardware with an OS that doesn't fully support the hardware.
By using 3.14 and delaying the release 1-2 months you have all the hardware supported and a tested kernel. But that won't happen, since for Canonical the release date is more important than anything else, nowadays.

System76 said in an interview that (this were when Mir was announced) they trust Canonical and because of their release schedule they can plan new stuff very good.

[Andrecorreia]

???

The question is, what pisses companies like System76 more of: Having to wait 1-2 months for Ubuntu and having full hardware support or delivering the hardware with an OS that doesn't fully support the hardware.
By using 3.14 and delaying the release 1-2 months you have all the hardware supported and a tested kernel. But that won't happen, since for Canonical the release date is more important than anything else, nowadays.

you think is better delay 1/2 months is a good thing? for a lts? not kidding.

the better way for stablity is using a 3.13 kernel and in august upgrade the kernel to 3.15 in 14.04.1 release. a bad distro launch is the end of this distro like 12.10, or fedora 19 for example. the first impression is important.

i really don t care about this, when i need i upgrade the kernel.

[dh04000]

Here's a question, WHY DOES ANYONE EVEN CARE?!? I use a ppa's to get newer drivers as well as install binaries from time to time. How's that any different? Non-story here.

[Vim_User]

you think is better delay 1/2 months is a good thing? for a lts? not kidding.

Yes, I think that. Especially for an LTS. Canonical's track record when it comes to release stable and relative bug-free isn't that good (they even recommended to their LTS users not to release to 12.04 LTS when it was released, but to wait until they fixed the bugs), so this could even be used for marketing.

the better way for stablity is using a 3.13 kernel and in august upgrade the kernel to 3.15 in 14.04.1 release.

Sure, business users will love the idea of replacing the kernel with a newer, not well tested, version by an upgrade instead of just waiting 1-2 months and go with a well tested kernel that supports the hardware they are aiming at from the beginning.

[nils_]

By backporting Broadwell support only Broadwell support is untested, by using 3.14 everything is untested.

Everything is a bit harsh, it's not like they're rewriting the whole kernel for every release. I think it would be more prudent to focus on testing the newer kernel release instead of wasting so much effort backporting things.

[JanC]

The question is, what pisses companies like System76 more of: Having to wait 1-2 months for Ubuntu and having full hardware support or delivering the hardware with an OS that doesn't fully support the hardware.
By using 3.14 and delaying the release 1-2 months you have all the hardware supported and a tested kernel.

But if you release with 3.15 even more hardware will be supported, so maybe we should wait another 3 months, and then we see 3.16 will have even more or better hardware support, and we add some more months to the schedule, and ..., and ..., and after 60 years you haven't seen a new release yet.