I got a response regarding Linux Libre in Ubuntu from Canonical
- Anmelden oder Registrieren um Kommentare zu schreiben
I recently emailed Leann Ogasawara from the Ubuntu kernel team regarding the addition of the Linux Libre kernel in the main Ubuntu repositories. I CC'ed both DNS and jxself in the email and she made sure to CC them as well in the response I got back from her.
I get the feeling that they are already short staffed in maintaining their kernel, but are willing to look at alternative kernels if there is activity in a PPA and the owner of that PPA is in communication with the Ubuntu kernel team. So if DNS or jxself were to offer a Linux Libre kernel PPA that matched the Ubuntu releases (like what Ruben does with Trisquel), then they may consider it if DNS or jxself is willing to maintain in the future.
This is what she had to say:
"Thanks so much for the email. At this time, my team unfortunately does not have the resources or cycles to maintain another kernel flavor in the Main Ubuntu repository. The additional support commitment and maintenance burden involved (eg security updates, SRU's, bug fixes, testing etc) are just not feasible for my team at this time.
However, I would point out that we do have active Ubuntu kernel flavors maintained by community members. The examples I cite here would be the -lowlatency [1] and -ppc [2] kernel flavors. They closely follow our Ubuntu kernel releases and then rebase and upload almost in lock step with us. We are happy to coordinate with them as they are a responsive and active community. The coordination effort usually involves my team providing a notification of a pending upload and each of the kernel flavor maintainers handle the rebasing and uploading of their respective kernel flavor.
From your end, it seems you'd want to hit a few milestones first... Eg. Begin fostering a relationship with the Ubuntu kernel developers via their IRC channel and mailing list. Demonstrate your ability to rebase and upload (to your PPA initially) in a timely fashion. Work towards getting your kernel flavor packaged and available in the Universe repository.
Thanks,
Leann"
I don't know much about the subject but what I know is that, even knowing it is an excellent idea (from the technical and practical point of view) it can become complicated for DNS and jxself to maintain them. Because the whole idea is not to have it update and maintained side-by-side with Trisquel but with Ubuntu, so this would mean using Ubuntu in order to test the driver and be able to solve the problems in the moment. This can really be complicated. But it could also be a big help for Ruben it's true.
I think this is really important and there are two very talented Linux Libre kernel guys out there (DNS and jxself) that are ready to take their skills to the next level. They have experience working with Ubuntu repositories either standalone or in launchpad PPAs and I have a gut feeling that either one of them could make a real change in bringing this kernel to Ubuntu.
I can only hustle so much and open up so many doors for both of them. They can either run with it and show Canonical that there is a need for a Libre kernel and get it added to the ecosystem or choose not to and continue to have Linux Libre as a niche kernel. I'm thankful that I am able to convince someone from the kernel team to respond back to me about the kernel and maybe make it available to Ubuntu and also take the work load off of Ruben.
"are willing to look at alternative kernels"
Thanks for bringing this up, t3g but this has been raised with Canonical before through back channels and this response isn't actually all that different than what Canonical has stated in the past (in summary): Ubuntu distributes a kernel with proprietary software, Canonical really isn't interested in stopping, but they have no issue if someone wants to maintain a different kernel in a separate repository.
I should probably add, though, that DNS and I are discussing this matter amongst ourselves. As you've pointed out, it mean extra time and work.
While it is true they won't stop distributing their own kernel with the non-free bits, it cant hurt having a libre one alongside it as an option. She even said that if someone shows initiative and the skills to maintain an alternative kernel, then there is a possibility it could be added to universe down the road.
She listed the realtime kernel as an example of one maintained separately in a PPA by the Ubuntu Studio guys that eventually made its way into the main repos.
I have a feeling with enough elbow grease and drive this is possible. Just needed a little initiative.
I am a member!
Going after getting an Ubuntu derived kernel without blobs into Ubuntu's own universe repository sounds like a good thing although I don't know what the value here is. Those who care about free software are probably not using Ubuntu and those basing a distribution off Ubuntu would just as easily add it to there repository (ie or say a Trisquel user who wants to update to a newer kernel would just add it separately).
I do see a slight advantage if it isn't a significant amount of time though and the Ubuntu/Trisquel repositories remain sufficiently in sync. It would make it slightly easier to direct users to a newer free kernel within Trisquel.
Ultimately I think back porting with an Ubuntu kernel for an LTS release would benefit most provided it was done with a few other critical packages. This would make it easier to move novice users to GNU/Linux and in particular Trisquel.
I think such a setup would be easier to move people to and be better polished than the current release schedule of a new version every 6 months.
After talking to a few developers in the past day it seems like my thinking is spot on and a project worth shooting for. However I think any such project needs to take a backseat to the current approach. The current approach is appealing to the current audience. So this thinking will have to wait until there are more funds available. Asking Rubén and others to work on it whom are already overstretched would be asking too much.
I think there is always value in proposing free software. Even on non-free systems. Even on Windows!
In the case here, you may encounter Ubuntu users who value their freedoms, know that the Linux kernel infringes them, and are not aware of a fantastic distribution called Trisquel. :-)
You may encounter Ubuntu users that say they cannot work without the proprietary X application (X not being a driver) although they want their system to be as free as possible.
Linux-libre in Ubuntu could also ease the transition to Trisquel: the user can test whether their hardware work without having to download an ISO, burn it and boot it.
Magic's viewpoint matches mine and why I made the push to contact the kernel head at Canonical. It benefits Ubuntu in addition to Trisquel. A Linux Mint or Elementary OS user can use it. Heck, maybe down the road if someone makes an Ubuntu based phone that isn't made by Canonical can use it.
I'm thinking big picture here. If jxself or Ruben is compiling a libre kernel and uploading to a repository, why not Ubuntu? It reaches a broader audience while still having a libre kernel for everyone and not just Trisquel.
I am a member!
As I mentioned... time and energy. This might be a good idea if it doesn't add to the work load. The problem is even just looking into it (even after you have all the info ready) is going to eat up time. Is it worth bringing this up to him?
I don't have the answer to that question.
I think we should be back porting a number of critical pieces to the LTS release of Trisquel and abandoning the STS releases. However it isn't a move we should take right now. And if it is ever to be a smart move it would probably be alongside the STS releases. In other words more work.
I thought it was worth a few minutes of Rubén's time only because there is a real possibility we (ThinkPenguin) could fund such a project eventually and it would have a beneficial impact (enable new GNU/Linux users). There are other bigger issues to solve first though. A flash replacement or at least some better alternatives to the entertainment options. This is really a HUGE problem to solve though.
I'm idealistic in that jxself builds the kernel once on Launchpad and then moves those packages to Trisquel while still leaving his originals in a Linux Libre PPA so others can add them to their existing Ubuntu or Mint installs. Its what I would do, but I cannot control how jxself thinks or how he works.
Focusing on the LTS releases and backporting newer packages has always been something I favor. I am doing it right now on the 12.04/Trisquel 6 branch with a handful of PPAs that offer the most recent software that I use on a daily basis like nginx, Git, MongoDB, MariaDB, PHP, VLC, Transmission, LibreOffice, Pidgin, and Gstreamer to name a few. I have not moved to a newer release and it has not hurt me one bit. I like the stability.
Plus, Trisquel can have point releases throughout the years like 6.1 and 6.2 like how Ubuntu does and re-authors the ISO with newer packages for those who need to install it.
I am a member!
I agree with the notion that the libre kernel should be distributed to as many places as possible, but at the same time depending on how much time and resources that would take I would hate to see less attention devoted to refining trisquel and being spent on providing a kernel that may or may not be used on a big name distro.
- Anmelden oder Registrieren um Kommentare zu schreiben