Linked by Amjith Ramanujam on Sun 21st Sep 2008 06:46 UTC, submitted by Rahul
Linux Greg KH, Linux kernel developer delivered a keynote in the Linux plumbing conference about the health of the ecosystem. His message was essentially that distributions that don't contribute to the ecosystem have to rely on the whims of others which is unhealthy for them. Here is an introduction the development model and some interesting statistics about the Linux kernel code. Update by TH: Rebuttals are appearing all over the web, like this one by Canonical's Matt Zimmerman ("He's refuting a claim which has, quite simply, never been made. [...] When this sort of thing happens on mailing lists, it's called trolling."), or this one by another Canonical employee, Dustin Kirkland.
Thread beginning with comment 331001
To view parent comment, click here.
To read all comments associated with this story, please click here.
VistaUser
Member since:
2008-03-08

Look for example at the fact that Ubuntu has usually better hardware support, if we all were on the same kernel the others could take the drivers we put in there and have hardware support that is just as good as Ubuntu.


http://lwn.net/Articles/290156/

As for the second, yes it is my cynical reading of the proposal from http://www.markshuttleworth.com/archives/162 (specifically the reply to the first comment.) and other blogs on planet gnome where developers were asking for the code.

To the cynic in me, even the talk of synchronisation talked of flailing - Ubuntu suddenly was using a kernel not being used by another distro, an xserver with the same issue and a couple of other important bits too. (since that time, Debian 5 and OpenSuse 11 have also made the same choices, so Ubuntu is no longer alone).

And for Ubuntu 8.10, there is the "late" switch to kernel 2.6.27 - something Fedora had already been planning to go with. TO me this smacked of synchronisation to cut the workload. Not a bad thing in itself, but how large is the iceberg?

I have also noticed that in some places, there is massive resistance to change in Ubuntu, as was seen in the discussion over whether to go with Empathy by default over pidgin for Ubuntu 8.10. This caution is good in some places, but bad in others. Me personally, I like a little bit of Chaos.

Reply Parent Score: 3

Kokopelli Member since:
2005-07-06

"Look for example at the fact that Ubuntu has usually better hardware support, if we all were on the same kernel the others could take the drivers we put in there and have hardware support that is just as good as Ubuntu.


http://lwn.net/Articles/290156/
"
uh.... read the original interview, not a response from a Red Hat developer?

He is responding to the point that with release synchronization Ubuntu gains the stability of the Red Hat/Novell kernels without contributing anything back.

Shuttleworth's argument is that Ubuntu provides better security updates (which I disagree with but that is another matter) and hardware comapatibility. Thus all parties gain from the common release, Ubuntu from Red Hat's stability patches and Red Hat from Ubuntu's hardware compatibility. Nowhere does he say others should use Ubuntu's Kernel, but instead that all parties would benefit from a unified release schedule and patch set.

Read into it as you like but he does not say "Everyone should use Ubuntu's Kernel."


As for the second, yes it is my cynical reading of the proposal from http://www.markshuttleworth.com/archives/162 (specifically the reply to the first comment.) and other blogs on planet gnome where developers were asking for the code.

To the cynic in me, even the talk of synchronisation talked of flailing - Ubuntu suddenly was using a kernel not being used by another distro, an xserver with the same issue and a couple of other important bits too. (since that time, Debian 5 and OpenSuse 11 have also made the same choices, so Ubuntu is no longer alone).

OK, obviously after reading the blog entry and responses I came away with a different impression but so be it. Ubuntu is still working on getting into the game and they seem to want to play by different rules. Time will tell if this form of contribution can be effective.

Launchpad can be used as an effective tracking tool with the ability to push bugs upstream to bugzilla among other tracking systems. This does give some benefit of a single point of entry for entering bugs. Personally till they open source it completely the idea of launchpad is "meh, so what" to me. Have no real need for it personally, though if it works with upstream as advertised it does make life simpler by providing a single point of entry for problems in Ubuntu.

On the other hand I am quite fond of bzr and use it for my projects. I use the svn connector to push change sets up to the company repository and can definitely see the benefit in using zr for tracking internal changes prior to pushing upstream.


And for Ubuntu 8.10, there is the "late" switch to kernel 2.6.27 - something Fedora had already been planning to go with. TO me this smacked of synchronisation to cut the workload. Not a bad thing in itself, but how large is the iceberg?

You see it as a late switch. I see it as a reasoned decision based on stability and hardware support. It would have taken more work to backport and test than to update to .27 and test. Shrug, different views I guess. Personally I would have been critical of Canonical had they not done the update to .27.

I have also noticed that in some places, there is massive resistance to change in Ubuntu, as was seen in the discussion over whether to go with Empathy by default over pidgin for Ubuntu 8.10. This caution is good in some places, but bad in others. Me personally, I like a little bit of Chaos.


I do not think there is much debate that it will happen, just if the 8.10 release was the right time for it. The counterpoint to the resistance to change are some of the unwise changes that Ubuntu did too early, like PulseAudio.

Reply Parent Score: 2