LWN.net News from the source
How complete is the platform?
Posted Jul 26, 2015 14:09 UTC (Sun) by coriordan (guest, #7544) [Link]
Is is it just the UI layer? (I don't mean the word "just" in a negative sense. A free UI layer is great and I like their focus on privacy.)
I read the linked article and the project's FAQ and "General info" pages but they just say "It's a platform". And they say it will run Android apps but I didn't see if it's based on Android. I found one article that says it's based on Kubuntu but it's not clear if it's a Kubuntu-ised Android or a normal (free software) Kubuntu system that runs on smartphones.
How complete is the platform?
Posted Jul 26, 2015 14:39 UTC (Sun) by boudewijn (subscriber, #14185) [Link]
For libhybris, we're actually moving to not using Ubuntu's version, but upstream. We originally started this pSroject based on Mer with GreenIsland as the Wayland compositor, but Bluesystems is a Kubuntu shop, which is the reason that was scrapped. The most difficult bit was to get graphics running again, there were weird incompatibilities between libhybris and Qt 5.4.
We removed Mir and most of the Ubuntu-touch things, except what's needed for running apps. Then we made kwin the wayland compositor which was actually really helpful for kwin on the desktop as well.
Finally we took the unchanged Qt5.4/Plasma 5/KDE Frameworks 5 stack with a custom plasma desktop definition (made in QML).
How complete is the platform?
Posted Jul 26, 2015 16:42 UTC (Sun) by coriordan (guest, #7544) [Link]
How complete is the platform?
Posted Jul 26, 2015 19:41 UTC (Sun) by robclark (subscriber, #74945) [Link]
well, not really true..
https://plus.google.com/+RobClark/posts/a4qWWJUr1sG
that is linux-next running on a actual phone.. obviously still a lot of work to go (but display and apparently wlan working, gpu should be working very soon.. linaro fwd ported some drivers (audio/video-decode/etc) that are not quite upstream ready to 4.0 kernel for dragonboard 410c, so some of that could be re-used as temporary solution for some of the other missing drivers.
Obviously I'm not trying to claim that it is something ready to ship products today, but a *lot* of progress has been made and it would be nice to see some of the "open source" android alternatives get involved with that, at least as a parallel track, rather than just re-skinning android and pretending there is nothing else that is possible..
How complete is the platform?
Posted Jul 26, 2015 20:21 UTC (Sun) by boudewijn (subscriber, #14185) [Link]
How complete is the platform?
Posted Jul 26, 2015 20:41 UTC (Sun) by robclark (subscriber, #74945) [Link]
https://github.com/freedreno/kernel-msm/commits/ifc6540-4...
which provides a reasonable stop-gap solution for getting all the other drivers, with sane userspace (ie. mesa + drm/kms).. I could ofc do backports more frequently (although so much progress has been made for devices that I have on upstream, that the backporting tends to get neglected).
ofc, my intention is not to make light of the work done in userspace, which itself is a large chunk of work.. so I hope you don't take offence at the "skinning android" comment. But rather I'm trying to point out that there are better alternatives to working around the issues lower in the stack with duct tape like libhybris. We are getting tantalizingly close to upstream from the kernel up, but the backport kernel option makes things look upstream from userspace perspective so work there can proceed in parallel with the kernel work.
Anyways, I do sometimes find it odd that so many of the android alternatives (plasma-mobile, jolla, etc) are targeting snapdragon devices... which is basically the only phone SoC which has a valid upstream display/gpu solution, yet they all make do with hacks to use android display/gpu stack. I couldn't argue against that until recently, since we didn't have upstream support for DSI (which is what is used for display on pretty much every mobile device) until a couple kernel versions ago. But now, actually in fact thanks to contributions from qualcomm, we have upstream support not just for DSI but also dual-DSI panels. So now is the time that we can start thinking about doing something better than just hacking around the android base layers.
How complete is the platform?
Posted Jul 27, 2015 9:24 UTC (Mon) by sebas (subscriber, #51660) [Link]
So, thanks again for this work. We may get in touch with you for directions how to make use of it. :)
How complete is the platform?
Posted Jul 30, 2015 9:59 UTC (Thu) by pabs (subscriber, #43278) [Link]
How complete is the platform?
Posted Jul 30, 2015 10:14 UTC (Thu) by dlang (subscriber, #313) [Link]
all to frequently the answer is "nobody"
who _should_ be doing the work of getting support upstreamed is the SoC vendors, Device vendors, and Carriers. Linaro is working to help them do this, as are a bunch of random people.
From my dealings with the XDA crowd, they don't see any value in bothering to try and push anything upstream.
How complete is the platform?
Posted Jul 30, 2015 10:00 UTC (Thu) by pabs (subscriber, #43278) [Link]
How complete is the platform?
Posted Jul 26, 2015 14:46 UTC (Sun) by sebas (subscriber, #51660) [Link]
We (the Plasma team) have created a reference implementation that runs on an LG Nexus 5. This reference image uses an Android kernel, libhybris and an Ubuntu userspace. It uses Ofono for the phone stack, much like Mer / Sailfish and Ubuntu Touch.
Of course we want to run it on other devices as well, whether or not we can do that using exclusively Free drivers depends on their availability. Driver and hardware development are currently outside of the focus of Plasma Mobile. (If anybody is interested to work with us on that, we're happy to indulge.)
An important aspect is that this is a community-controlled and openly developed system. We designed it as an inclusive system, so we want maximum control in the hands of the users. The phone runs on Wayland, but we're supporting X11 apps, Ubuntu Touch apps and are currently working on getting Android apps to work as well, so we can pick up the user where she currently is but allow her to move to a more Free and less corporate-controlled system.
One driving aspect of this project is that we want to enable the user's privacy in the sense that the code is auditable, and that the user can choose the services she connects her phone with.
Also, developing phone and desktop in tandem will allow us to provide deep integration features between these devices, using for example KDE Connect ( https://community.kde.org/KDEConnect ).
We want this to be end-user ready by next summer. Blue Systems (my employer) sponsors developers and designers to work together with the community on this project, which has been put under KDE's infrastructure and governance yesterday when it was publically announced.
Let me know if you have more questions, happy to answer them. :)
Plasma Mobile launched
Posted Jul 26, 2015 21:28 UTC (Sun) by idupree (guest, #71169) [Link]
"Binder has a number of serious security issues when used outside of an Android environment, [Greg Kroah-Hartman] said, so he stressed that it should never be used by other Linux-based systems." https://lwn.net/Articles/551969/
What are these security issues and do they cause Ubuntu Phone / Plasma Mobile / Mer / etc to be unfixably insecure?
Plasma Mobile launched
Posted Jul 27, 2015 6:28 UTC (Mon) by Wol (guest, #4433) [Link]
Cheers,
Wol
Image creation guide
Posted Jul 28, 2015 10:14 UTC (Tue) by jreznik (guest, #61949) [Link]
Image creation guide
Posted Jul 29, 2015 9:06 UTC (Wed) by sebas (subscriber, #51660) [Link]
I'll look into this next week, if anyone beats me to it, that's of course most welcome. Please hop onto plasma-devel@kde.org and let us know so we can coordinate.
Plasma Mobile launched
Posted Jul 28, 2015 20:40 UTC (Tue) by gspyplex (guest, #103776) [Link]
Why does Google Chrome phone home several ways with encrypted data? Why can't I see the real source code of Chrome? Why is Google floating a NSA spy barge off the coast of California? Why should I trust my Android smartphone?
Plasma Mobile launched
Posted Jul 28, 2015 20:46 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]
Is this reason?
Posted Jul 29, 2015 8:23 UTC (Wed) by cyperpunks (subscriber, #39406) [Link]
Is this reason?
Posted Jul 29, 2015 9:09 UTC (Wed) by sebas (subscriber, #51660) [Link]
If you've concrete questions, I'm happy to indulge, but there's really not even a starting point for actual information in yours.
Is this reason?
Posted Jul 29, 2015 10:42 UTC (Wed) by cyperpunks (subscriber, #39406) [Link]
I am just very sad for recent changes in KDE camp.
KDE 4 was finally becoming stable and feature complete.
Now, for some reason, KDE 5 had to be rewritten from scratch.
All settings from KDE 4 is lost on upgrade.
KDE 5 Plasma is very unstable and almost impossible to use in a serious setting.
kwin is crashing if specific apps are started. drkonqi hangs and blocks desktop
completely when apps crash.
Lots of features have not be rewritten yet.
To top this it's impossible to have parallel install of KDE4 and KDE5.
All this due to support for phones or there any other good reason to break
the best Linux desktop avaiable and make it impossible to use the working version?
Is this reason?
Posted Jul 29, 2015 12:12 UTC (Wed) by distances (guest, #103785) [Link]
I did a clean install knowing that both Qt and KDE libraries were restructured. Maybe your old config is somehow clashing here since you went with the upgrade route?
I know this may be a subjective experience, and as sebas noted, this isn't really a place for this discussion anyway. I wish all the best for the mobile endeavour -- as Martin G. already pointed out in his blog, both the mobile and desktop efforts seem to nicely benefit from each other!
Is this reason?
Posted Jul 29, 2015 16:22 UTC (Wed) by dlang (subscriber, #313) [Link]
I know the claim is that this is an upstream bug, but with a bug like this (that's not fixed even using the 5.4 PPA), it hurts a LOT of people.
and telling people that they shouldn't upgrade to 15.04, they should stick with 14.04 if they want a working system (even if they were already on 14.10) does not improve the happyness of users.
the non LTS releases are not supposed to be "for expert users only, may break at any time", they are supposed to work, just not be supported for as long.
Is this reason?
Posted Jul 29, 2015 16:41 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]
That might be the intention but noone has basically achieved that in Linux land. The faster you move, the more prone you are to breakage. Automatic testing etc helps to some extend and I am sure everyone is trying to get better at it but there is no silver bullet.
Is this reason?
Posted Jul 29, 2015 16:47 UTC (Wed) by dlang (subscriber, #313) [Link]
the non-LTS releases are not rawhide/unstable/sid snapshots, they are supposed to be stable, not "let's push the unstable stuff into it so that we stabilize it before the next LTS release, we don't want to keep supporting the old version in the LTS release"
Is this reason?
Posted Jul 29, 2015 17:34 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]
They speak from experience.
>the non-LTS releases are not rawhide/unstable/sid snapshots, they are supposed to be stable, not "let's push the unstable stuff into it so that we stabilize it before the next LTS release, we don't want to keep supporting the old version in the LTS release"
That seems to be what they end up being in practice though. Not quite rawhide or unstable but the stabilization level is nowhere near long term releases.
Is this reason?
Posted Jul 29, 2015 20:03 UTC (Wed) by bronson (subscriber, #4806) [Link]
True. It just seems like the Linux desktop is speeding in circles.
The last ten years have seen a lot of fundamental change but not a lot of fundamental forward progress. Hoping the next ten years will be better.
Is this reason?
Posted Jul 29, 2015 21:34 UTC (Wed) by renox (subscriber, #23785) [Link]
I wouldn't hold my breath, I see no reason why this would change..
Is this reason?
Posted Jul 31, 2015 11:29 UTC (Fri) by sorpigal (subscriber, #36106) [Link]
I think desktop-style computing is 90% a solved problem. I've not seen anything I'd call forward progress anywhere for 10+ years. We're in the realm of "doing it again, but better," where better is not something users will care about. Microsoft has the same basic problem; apart from introducing some trivial window management improvements and cycling through modestly different launchers its desktop offering has not changed significantly. Apple has, again, the same problem. All changes I see are churn designed to bring big screen+mouse and small screen+finger UIs closer together.
Is this reason?
Posted Jul 31, 2015 12:44 UTC (Fri) by anselm (subscriber, #2796) [Link]
When it comes to desktop computing, one person's innovation is the next person's annoying and superfluous disruption. It is very difficult to be innovative when a large part of your userbase will be utterly flabbergasted and unable to get useful work done if the printer icon moved two inches to the right.
Often when somebody brings out something really “innovative”, there is a large hulabaloo and the next thing you see is them backpedaling towards the tried and true. Microsoft tried to revolutionise the PC UI in Windows 8 and look what Windows 10 is like. The only company that could pull this off (and did, in the past) is Apple, because many people don't buy Apple stuff for the software that comes with it – they buy Apple stuff in order to own Apple stuff. With Linux distributions, it doesn't matter as much because if desktop environment P doesn't appeal to you there are always Q, R, and S to try out, but the hulabaloo is still there.
Is this reason?
Posted Jul 31, 2015 15:48 UTC (Fri) by boudewijn (subscriber, #14185) [Link]
That's just your perception. I'm pretty convinced that the reason you don't see progress is that you started using desktops ten, maybe twenty years ago. Your habits got stuck with what was available then, and that's still available now, so you're doing what you did then and fail to see how a modern desktop like plasma five could help you get your stuff done more efficiently -- but you're still demanding progress in the abstract. So, tell us -- _what_ would you call "fundamental" progress? Have you got any idea about that?
As for me, search, information integration, activities, scripting, widgets -- I don't use all of that either. Just like with KDE 1.x, I've got a panel, I use one or two terminals, a text editor or an IDE and krunner to do my work. After all, ten years ago I spent all my waking hours working on Krita, and I'm still doing that. I notice that krunner in Plasma 5 is vastly better than in Plasma 4, though. And at least I'm aware that if I took the time to learn the new stuff, like for instance my wife did, I might be more efficient. But I'm too stuck in my habits...
Is this reason?
Posted Jul 31, 2015 22:27 UTC (Fri) by bronson (subscriber, #4806) [Link]
I'm absolutely happy to adapt to new and clearly better ways of working. I'm just not happy to lose a bunch of features and wait 4 years while they slowly get added back. And I'm not happy to relearn basic stuff just for the sake of being different. And then relearn it again 5 years later, and again 5 years after that, etc.
The two big Linux desktops are always replacing working things with half-baked nerd experiments. Microsoft does this around once a decade and then apologizes profoundly in the very next OS release. Apple almost never does this and, if they do, they provide the old way as a configuration option for 5 or 10 years. That's why I do paid work on Macs now -- I can't afford the downtime. :(
So, what would be fundamental progress? Me no longer having to dick with Linux desktops every few years. A consistent design that, though imperfect, grows fluidly (again, Mac). More working, less coming up with new fads. Just stop going backward!
Is this reason?
Posted Aug 1, 2015 8:06 UTC (Sat) by anselm (subscriber, #2796) [Link]
The two big Linux desktops are always replacing working things with half-baked nerd experiments.
This should come as no big surprise given that their main purpose is to provide entertainment for the nerds involved in making them. Building a convenient and reliable user experience for third parties is a side effect.
This is an obvious consequence of the observation that maintaining complex stuff which you didn't write yourself, which probably isn't documented all that well, and which doesn't work quite right is tedious and boring. Building new stuff from scratch according to your own design is what is interesting and fun. Hence the bias towards throwing old stuff out and reinventing things over again.
Is this reason?
Posted Aug 10, 2015 22:32 UTC (Mon) by renox (subscriber, #23785) [Link]
Easy! For free desktop environement a fundamental progress would be stability:
when you replace a part by another one, introduce the replacement as an additional option first
and only when the new part replace 100% of the old one, change the default (but provide still the old one to ensure that users who have issues with the new part can still work until the issues are solved).
This is very unlikely to happen because stability isn't fun for developers.
Plasma Mobile launched
Posted Jul 29, 2015 10:29 UTC (Wed) by mmind00 (guest, #103779) [Link]
What I have in mind are devices like the recent Chromebook Flip (http://reviews.gizmodo.com/asus-chromebook-flip-review-wh...) which can be both laptop and tablet and already is able to run a regular GLES-accelerated X11 userspace (Debian or Arch-Linux currently). So ideally it would have a regular plasma-desktop in laptop mode and use plasma-mobile in tablet mode.
Plasma Mobile launched
Posted Jul 29, 2015 16:29 UTC (Wed) by boudewijn (subscriber, #14185) [Link]
Plasma Mobile launched
Posted Jul 29, 2015 20:14 UTC (Wed) by bronson (subscriber, #4806) [Link]
Reminds me of my thinkpad X61t misadventure... I could make peoples' jaws drop by flipping the screen, whipping out the stylus, penning some notes, flipping the screen back... It was truly the stuff of technology ads.
In practice, though, the thing was slow, inconsistent, cranky, and the stylus hardly worked outside Microsoft's demo apps. Linux was only marginally better (at least it wasn't always hinting that I should use the stylus, and pressure-sensitive Gimp was pretty fun). After the novelty wore off a few months later, I couldn't get rid of it fast enough.
If I tried again on modern hardware, think my experience would be better?
Plasma Mobile launched
Posted Jul 30, 2015 6:19 UTC (Thu) by boudewijn (subscriber, #14185) [Link]
Copyright © 2015, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds