PinePhone
-
I am not sure we want to really [move to systemd] soon, and not rather stick with upstart. We dont need system atm in order to run a full working phone.
I'm not advocating for moving to systemd merely for the sake of moving to systemd. Rather, I'm advocating for moving to systemd because rebasing from 16.04 to 20.04 requires moving to systemd. (If 20.04 didn't require a move to systemd, I'd say stick with upstart for as long as you like.)
Also, just to be clear, I'm not suggesting that you should drop everything else and start migrating to systemd immediately. Certainly, transitioning to new Mir, transitioning to new Unity8, and transitioning to full Wayland, all should come first. But you need to start planning for the move to systemd soon, so that you can rebase to 20.04 (at least for the PinePhone) sometime early in 2020, hopefully by April or May.
transition to systemd will render all old devices unusable (kernel < 3.10)
Understood. And that's a big part of why I feel that this sort of work (preparing for 20.04) should be started first (and hopefully soon) for the PinePhone, and then later, where possible, applied to builds for older platforms. That is, because a transition to systemd (and rebasing on 20.04) could be done more efficiently on the PinePhone compared to other platforms, start with the PinePhone.
Although Ubuntu 16.04 may technically have a year of support left, at nearly four years old, 16.04 is showing its age. You skipped 18.04. Those buying a shiny new PinePhone will want a shiny new release (or at least not an ancient release) on it. (And you won't need to backport so much to a newer release.) Given that moving to 20.04 (including the requisite step of moving from upstart to systemd) will take a fair bit of work and time, better to start sooner than later.
If you're going to let the limitations of your old hardware restrict the pace of PinePhone development, you'll get left behind.
As I wrote before, if disappointed with UT on first impression, potential users on the PinePhone may become forever lost to one of the competing alternatives, or perhaps forgo the PinePhone altogether.
So, I'll repeat what I originally posted, but with one small edit: I hope that, once OTA-12 has been released, UBports will focus like a laser on PinePhone development for the next few months, even if that means other priorities are delayed for a short time.
-
@Flohack Indeed, I think that it is still a little early to "sacrifice" all the precursor phones that support UT well at the cost of evolution in all directions !!
-
@Flohack
Sadly some data gaps in this table:
http://docs.halium.org/en/latest/supplementary/devices/index.html
But it doesn't look too bad for Kernel 3.10+ support in my eyes.On the end UT doesn't seem to have the stated goal of Postmarket OS to maintain 10 year old devices.
-
Indeed, I think that it is still a little early to "sacrifice" all the precursor phones that support UT well at the cost of evolution in all directions !!
Just to clarify, I'm not suggesting that any phones be sacrificed. To the contrary, I hope that all currently supported phones (even those with kernel < 3.10) receive the enhancements currently slated for OTA-12, including new Mir and new Unity8.
However, we need to acknowledge that 16.04 is nearing EOL, and that UBports will, inevitably, need to rebase to something newer, probably Ubuntu 20.04. And, because rebasing to 20.04 necessitates transitioning to systemd, at least some old devices won't support the upgrade, whenever that happens.
What should happen to those old devices that can't support Ubuntu 20.04 after the inevitable rebasing? I don't know. Perhaps that should be discussed in another thread.
But regarding newer devices, we've been whining for years about the lack of phones that can run mainline Linux. Now we've got one: the PinePhone, which is the topic of this thread. But without a modern and polished OS to install on it, there will be no demand for the PinePhone, and production will likely end. I don't want that to happen.
UBports will only have one chance to make a first impression with potential users of the PinePhone. If UBports makes a bad first impression on the PinePhone, UBports (and the PinePhone) will have missed a huge opportunity. Again, I don't want that to happen.
While I feel that UBports should aim to upgrade all newer phones to Ubuntu 20.04 to as soon as feasible, I'm not proposing that here.
With regard to the PinePhone, I propose that UBports should release, no later than May of 2020, a UT build for the PinePhone that runs upstream Mir 1.6+ using full Wayland on an Ubuntu 20.04 base.
-
@GizmoChicken said in PinePhone:
However, we need to acknowledge that 16.04 is nearing EOL, and that UBports will, inevitably, need to rebase to something newer, probably Ubuntu 20.04. And, because rebasing to 20.04 necessitates transitioning to systemd, at least some old devices won't support the upgrade, whenever that happens.
-
@GizmoChicken sounds good, so what are you doing to make that happen?
IMHO for now a well working OTA12 with 16.04 base is good enough for what you are asking for on the PinePhone, but I agree the switch to 20.04 should probably be priority after that.
-
@ernest Thanks for providing a chart that emphases that, except for "extended security maintenance for customers" (which doesn't apply to UBports users), 16.04 is indeed nearing EOL.
sounds good, so what are you doing to make that happen?
Rallying the troops and breaking down resistance among the old guard.
IMHO for now a well working OTA12 with 16.04 base is good enough for what you are asking for on the PinePhone, but I agree the switch to 20.04 should probably be priority after that.
I'm proposing that UBports set a goal to deliver 20.04 to the PinePhone by sometime in May of 2020, which is almost five months from now. But sure, 16.04 will be fine on the PinePhone until then, and maybe even for a while after, if UBports can't make that goal.
-
Too little time right now to add the details to the OP (will update sometime later), but I highly recommend that anyone who has yet to see it read @PINE64 's December Update. Some very good news for fans of physical keyboards among other updates.
-
@GizmoChicken I'm pretty sure that timeline will be really hard to meet with the current workforce of UBports even if we decide to do it now. I'd love that as well but we have to be realistic though. Best case that I can think possible is having the new Mir/Unity8 and perhaps wayland. Rebasing to 20.04 would be a huge task to do.
And somehow, I don't agree that there's a rush in the development just to make a good first impression to future PinePhone users. I mean as far as I know, there's no OS yet that's fully functional on the PinePhone and I think UBports already has the advantage at the moment. From the current state of UBports on PInePhone, I'd say it would already give a good first impression
Lastly, not to be negative, but I predict that most future PinePhone users are looking for a linux-experience that are similar to desktops which UBports won't give them by default and that might be a bad first impression already that we can't do much about
But of course, i still believe UT is a great OS that many will still learn to love -
I'm pretty sure that timeline will be really hard to meet with the current workforce of UBports even if we decide to do it now. I'd love that as well but we have to be realistic though. Best case that I can think possible is having the new Mir/Unity8 and perhaps wayland. Rebasing to 20.04 would be a huge task to do.
You'd probably know better than I would. But please do keep in mind I'm not suggesting that phones requiring hybris be upgraded to 20.04 by about May of 2020. Rather, my suggestion is restricted to the PinePhone, which is much more like upgrading a release for a traditional desktop device. So I'm not ready to give up on my suggestion just yet.
And somehow, I don't agree that there's a rush in the development just to make a good first impression to future PinePhone users. I mean as far as I know, there's no OS yet that's fully functional on the PinePhone and I think UBports already has the advantage at the moment.
It's a bit late here, and so I'll provide a more complete response another time. But in the meantime, please consider the reaction of a potential user who's told that the OS he/she is about to install on his/her shiny new Linux phone is based a four year old release. Sure, you can justify sticking with 16.04 from a technical standpoint. But it's very difficult to market that justification to those who want the latest and greatest. Also, I think you'd agree that most devs would much prefer to develop on a newer platform. Here too, perhaps you can justify developing on a four year old release, but it's a tough sell.
Lastly, not to be negative, but I predict that most future PinePhone users are looking for a linux-experience that are similar to desktops which UBports won't give them by default and that might be a bad first impression already that we can't do much about But of course, i still believe UT is a great OS that many will still learn to love.
Actually, I think that UT has the capability to provide a Linux experience (through convergence) that even the most hardened Linux desktop user would welcome. But as mentioned before, an OS based on a four year old release will be, well, a bit off-putting to many such potential users.
Again, I know that sticking to 16.04 can be justified from a technical standpoint. But really, 16.04 is nearing EOL and upgrading is inevitable. Maybe I'm wrong, but I suspect that upgrading from 16.04 to 20.04 will be easier on the PinePhone than on any other device. If so, I hope that UBports will start the upgrade process for the PinePhone sooner rather than later, and certainly before starting the process for those devices requiring hybris.
-
This is a link to a guy putting a 1tb card in his Nintendo
I see the pinephone will take up to 2tb.
Who is brave enough to try ,they are not cheap.This is a bit ,tongue in cheek .
To keep you occupied ,while you wait for your device -
@GizmoChicken I share your dream of having version 20.04 on the Pinephone, that would be very exciting for the one who has a phone with Linux and without android. However, you have to be realistic, if there were more people donating to Ubports we would have more developers on the team.
I think it is better to have a Pinephone with a good software in 16.04 in which the phone and its applications work well, than a bad software in 20.04, with malfunctions, crashes or restarts. Furthermore, keep in mind that every app has to be updated.
Haste always brings problems.
When people buy a new phone they don't want problems, they want a software that is stable and works perfectly.
I opt for stability...
greetings...
-
Don't forget the original timeline of devices with OS ready for order in March next year. That's not only in our case UT being stable and working but all associated tooling, manufacturing processes etc. from everyone else involved as well. That's not many weeks away with seasons holidays and new year to factor in. I can't see any manufacturer waiting on one part of the process running late due to over stretching themselves and delivering late. It would be us missing out on the product launch, Which would still go ahead. Build on a good solid base. Updates will always follow.
-
-
if there were more people donating to Ubports we would have more developers on the team.
A transition from 16.04 to 20.04 will necessitate replacing upstart with systemd. From what I gather, UBports could use some help with such a project.
Has UBports issued a call-out requesting help from those in the Linux community having expertise with systemd?
I think it is better to have a Pinephone with a good software in 16.04 in which the phone and its applications work well, than a bad software in 20.04, with malfunctions, crashes or restarts. Furthermore, keep in mind that every app has to be updated.
You mentioned two potential outcomes, but forgot at least one potential: "good software in [20.04] in which the phone and its applications work well." That's the outcome that I prefer.
Haste always brings problems.
Developing on a platform nearing EOL brings its own set of problems. The upgrade from 16.04 to 20.04 must be done, eventually. The question is when. So I'll ask that question: if not by May of 2020, by when?
And no, I'm not asking facetiously. To be taken seriously, UBports needs to publish a roadmap, or at least a rough outline of a roadmap, discussing such matters.
@Lakotaubp said in PinePhone:
Don't forget the original timeline of devices with OS ready for order in March next year. That's not only in our case UT being stable and working but all associated tooling, manufacturing processes etc. from everyone else involved as well.
Okay, if you truly need a stable release by March of 2020, then yes, I'll concede that you are making an excellent case for first releasing on 16.04 in March, and updating later.
Even so, and sorry for being so pugnacious, but as I wrote above, UBports needs to publish a roadmap, or at least a rough outline of a roadmap, discussing your goals, when you hope to achieve them, and if applicable, what outside resources (beyond money) would be helpful in achieving your goals.
-
There were too many answers since my last post here, and I cannot honor all of them, just a few notes:
-
There is no requirement to use systemd on 20.04, if we fork upstart and adapt it, it can be used. In fact a lot of things already now on the phone have been forked, to either use newer or older versions. Ubuntu Touch is a heavily tweaked Ubuntu, in no way just using all packages as they are.
-
On the other hand, all forked stuff needs to be maintained and needs to get security updates. This will require more and more maintainer hands after some time has passed. Already now we are in troubles, like for example ofono for making phone calls etc. has diverged from the upstream, and we cannot just apply patches there. So basic rule is fork as less as possible.
-
Mir and unity8 are a stack that will stay for a longer time. We cannot and will not rip out Mir to replace it with Wayland, and in the same moment create 1000s of regression bugs. Mir is a supported platform, maintained by Canonical, and we can consume it for the time being. There is no added value in just using Wayland now.
-
With a small team there will be always shifting priorities. But with the pinephone everything will get much easier, true. It has a mainline kernel 5.x something already, and of course we will try to get to 20.04 with it as early as possible. But, we cannot just switch to that task in Feb or March. We promised to fix location/GPS and Bluetooth first for the old phones. There are approx 100 Bugs outstanding on those 2 topics only.
-
-
Along with what Flohack has just said above there was the call out for developers not too long ago https://forums.ubports.com/topic/3368/want-to-develop-ut-and-need-a-device/2 which should bring more developers on stream in the not to distant future and move the whole of UT development along.
Secondly I would like to just point out this thread is specifically about and for UT and PinePhone and not general UT future planning and development discussions. I think we are just beginning to drift slightly.
No harm at all in that line of thought and discussion but probably better off in a new thread somewhere else. Thank you for your understanding. -
@GizmoChicken said in PinePhone:
please consider the reaction of a potential user who's told that the OS he/she is about to install on his/her shiny new Linux phone is based a four year old release
Honestly? They wouldn't care about it at all, if the phone works great and allow them to do what they want. There are tons of missing features and even more bugs in UT, do you really think that a user will care more about a 20 being displayed on the About page instead of a 16, or what the phone actually does? Why put pressure to switch to 20.04 when the current state of development is to do a successful phone call?
-
Just when I thought I was out...
@Fla, maybe you didn't read my response to @Lakotaubp in which I wrote:
"Okay, if you truly need a stable release by March of 2020, then yes, I'll concede that you are making an excellent case for first releasing on 16.04 in March, and updating later."
But okay, if you want to continue the discussion, here's my response to your inquiries:
There are tons of missing features and even more bugs in UT, do you really think that a user will care more about a 20 being displayed on the About page instead of a 16, or what the phone actually does?
I'm tempted to say, yes, they'll care because "[20] is [4] louder than [16]."
But in all seriousness, yes, I'd rather have tons of missing features and bugs on a new platform, instead tons of missing features and bugs on a soon-to-be-obsolete platform.
On a soon-to-be-obsolete platform, I'd be asking myself why I'm suffering through these bugs now, only to suffer through an all new batch of bugs next year, or whenever upgraded to 20.04. But maybe that's just me.
In any case, as I conceded to @Lakotaubp , perhaps it's better to ship with 16.04 in March, and then upgrade to 20.04 as soon as feasible.
Why put pressure to switch to 20.04 when the current state of development is to do a successful phone call?
I'll answer that question with a question: Why do you want to strive for a successful phone call on a soon-to-be-obsolete platform, when, instead, you could switch to 20.04 soon, and strive for a successful phone call on a platform that you won't be forced to upgrade in about one year?
But again, as I conceded to @Lakotaubp , perhaps it's better to ship with 16.04 in March, and then upgrade to 20.04 as soon as feasible.
@Lakotaubp and @Flohack I promise that I had planned merely to upvote your comments and then move on. But as long as I'm responding to @Fla , I'll provide a few brief comments responding to each of you.
Mir and unity8 are a stack that will stay for a longer time. We cannot and will not rip out Mir to replace it with Wayland, and in the same moment create 1000s of regression bugs. Mir is a supported platform, maintained by Canonical, and we can consume it for the time being. There is no added value in just using Wayland now.
I truly do appreciate your response. But to be blunt, provided I'm correctly interpreting what you meant when you wrote that "[UBports] cannot and will not rip out Mir to replace it with Wayland," I think what you wrote in the above snippet is pretty far off-the-mark. The whole point of upgrading from "old" Mir to "new" Mir is that "new" Mir is Wayland compliant. For now, in addition to the Wayland protocol, "new" Mir also supports the protocol used by "old" Mir. But that support for the old protocol is not enabled by default in current versions of Mir, and will eventually be removed from Mir altogether. Hence the need for migrating fully to Wayland, eventually.
There is no requirement to use systemd on 20.04, if we fork upstart and adapt it, it can be used. In fact a lot of things already now on the phone have been forked, to either use newer or older versions. Ubuntu Touch is a heavily tweaked Ubuntu, in no way just using all packages as they are.
I'm less sure about this topic than I am about Mir/Wayland stuff. While it would be possible to fork upstart and maintain it yourselves, to me, that sounds like a terrible idea. So no, my guess is that, unless you move away from Ubuntu to something like Devuan (which I doubt will happen), you'll be using systemd on 20.04. But maybe I'm wrong about this. Hope this will be discussed elsewhere.
@Lakotaubp said in PinePhone:
Along with what Flohack has just said above there was the call out for developers not too long ago https://forums.ubports.com/topic/3368/want-to-develop-ut-and-need-a-device/2 which should bring more developers on stream in the not to distant future and move the whole of UT development along.
The linked posting isn't really what I had in mind. I was thinking you could identify a few discrete projects (and the skill sets needed for completing them) and posting that information.
@Lakotaubp said in PinePhone:
Secondly I would like to just point out this thread is specifically about and for UT and PinePhone and not general UT future planning and development discussions. I think we are just beginning to drift slightly.
No harm at all in that line of thought and discussion but probably better off in a new thread somewhere else. Thank you for your understanding.100% agreed. Provided no one pulls me back in, I'm out
-
BTW, will the PinePhone support USB OTG? I'd sometimes like a hard-attached USB keyboard....
