Running Android Apps in Pyra OS


Klumpen

Run away! Run away!
Joined
Nov 19, 2011
Messages
8,578
Location
Uncanny Valley
Alien Dalvik is a finished and working solutions to run Android apps in a normal Linux OS.
http://www.myriadgroup.com/device-solutions/device-software/myriad alien dalvik.aspx

Jollaphone is using it, for example, and it works without any issues.

I am planning to ask them how much it would cost to get it into the OS.
New link: http://www.myriadgroup.com/en/products/device-solutions/mobile-software/alien-dalvik/
IMO a lot better than running Android ;)
And all the Android games? Do they really run well without "real" Android? ^^"
As far as I am aware, it should run almost anything except for stuff using Google Services.
So probably 98 - 99% of that stuff.
Any news about that?
 
Last edited:
Haven't had the time to contact them yet, planning to do it this week.
Any news?

If Pyra won't be able to play Android games, I'd like an advice for which Android device (not phone) is best for playing Android ports from the Humble Bundles,

but I'd rather like to use Pyra instead of course, have enough batteries lying around in barely used devices.
 
Last edited by a moderator:
If Pyra won't be able to play Android games, I'd like an advice for which Android device (not phone) is best for playing Android ports from the Humble Bundles, but I'd rather like to use Pyra instead of course, have enough batteries lying around in barely used devices.
Okay if this Alien Dalvik doesn't pan out, The omap5 can boot Android if needed. I haven't played with it at all, but there is an official Ti Android image I could download and try out. 
 
Last edited by a moderator:
Is there a FOSS or at least gratis solution for this? I don't want to force people to pay for proprietary software.

Android is supposedly Open Source. GNU/Linux is FOSS. It feels really weird to have to pay some 3rd party for bridging the gap between the two.
 
Why does it always have to be gratis or FOSS? If it adds a lot more for little then why not?
I don't like to buy hardware which comes bundled with commercial software that you have to pay for whether you like it or not. Even if that software is actually useful to me -- which is not at all clear in this case.
 
_wb_, what keeps you from coding a FOSS version of Dalvik Alienware?
A lack of motivation and maybe a lack of skills I guess. Mostly I don't really know how much effort it would take, and I don't like to start such projects.

Maybe we can ask the Replicant guys to support the Pyra, so you just have to put a microSD card in the internal slot to turn your Pyra into a Android device?
 
I don't mind buying hardware with components which are not FOSS and I'm sure others wouldn't mind either
 
For most people that probably depends on actual cost vs. (subjective) value.


In this particular case I'd like to find out more about Alien Dalvik. Jolla seems to be the only source of information besides marketing material. See https://together.jolla.com/questions/scope:all/sort:activity-desc/tags:alien-dalvik


There seem to be the usual problems (access to host hardware, random crashes, some lag, excessive ressource usage for some users), but nothing spectacular and vendor support seems to be rather good. The most important questions is probably, how support works out mid to long term. Android 4.1 is nice now but possibly not in 3 years.


It would also be interesting to get more info about this Myriad Group, especially their engineering department. Anybody knows public lists/boards where the Alien Dalvik developers are active?
 
Mostly I don't really know how much effort it would take
A lot !
The effort have already been looked out in this community. google for it :)
Would it still be as difficult if we have a very recent kernel and much more space for libraries? I don't really know what the issues are, but whatever they are, I would expect things to be easier than on the Pandora.
 
Would it still be as difficult if we have a very recent kernel and much more space for libraries? I don't really know what the issues are, but whatever they are, I would expect things to be easier than on the Pandora.
The kernel is not a problem. Notaz have already demoed this on pandora (the android.pnd thingy).The problem is the number of java class to rewrote. Some member did dissect that for us a few years ago. I cant seems to find that post sadly
 
_wb_: Consider it a marketing cost. You may not benefit from it (if you don't want to), but it may bring more customers for ED. Like marketing costs, this too is factored into the final unit price. I'm not familiar enough with the pricing model to know if they sell the licenses per unit or per product though. If the cost per unit is low enough to be neglible, why deny others a use for Pyra no FOSS solution provides?

I suggest keeping alien dalvik separate from the base OS image for distribution permit ambiguity reasons (so third parties can host/distribute the OS image without fear of repercussions). Kinda like the codec pack is now.
 
_wb_: Consider it a marketing cost. You may not benefit from it (if you don't want to), but it may bring more customers for ED. Like marketing costs, this too is factored into the final unit price. I'm not familiar enough with the pricing model to know if they sell the licenses per unit or per product though. If the cost per unit is low enough to be neglible, why deny others a use for Pyra no FOSS solution provides?

I suggest keeping alien dalvik separate from the base OS image for distribution permit ambiguity reasons (so third parties can host/distribute the OS image without fear of repercussions). Kinda like the codec pack is now.
It's not just the cost I'm worried about. If the cost is low enough (~1 EUR per device or so) I wouldn't really care about that. I'm more worried about long-term support and things like: will we be able to upgrade stuff (in particular, the kernel) without trouble? It would really suck if maintaining compatibility with some proprietary software would become something that holds us back from moving forwards in the future.

Why is a dual-boot between GNU/Linux and Replicant not a good enough solution for now?

Also I'm opposed to forcing people to buy proprietary software if it can be avoided. GPU drivers and stuff like that is one thing -- there is no alternative at the moment, so not much can realistically be done. But in this case there are plausible alternatives.

I don't mind having Alien Dalvik as an optional thing you can choose to have pre-installed on your device, but I don't want it to be something everyone has to pay for even if they are not interested in running fart apps without having to reboot. (Just kidding about the fart apps, I realize that there is more than that in Android ;) )
 
It's not just the cost I'm worried about. If the cost is low enough (~1 EUR per device or so) I wouldn't really care about that. I'm more worried about long-term support and things like: will we be able to upgrade stuff (in particular, the kernel) without trouble? It would really suck if maintaining compatibility with some proprietary software would become something that holds us back from moving forwards in the future.
AFAICT, Alien Dalvik is a userspace application that just happens to run android apps. Why would it be trouble? Or have I missed something?
Why is a dual-boot between GNU/Linux and Replicant not a good enough solution for now?
You can't multitask android and non-android applications with dual boot. As an analog, consider you had to switch to KDE every time you wanted to run Qt apps or Gnome for GTK+ apps. Not very convienient and takes away a lot of the point for being able to run both. As a direct comparison: why have Wine at all when you can just boot to Windows?
Also I'm opposed to forcing people to buy proprietary software if it can be avoided. GPU drivers and stuff like that is one thing -- there is no alternative at the moment, so not much can realistically be done. But in this case there are plausible alternatives.
I'm too, but given the downside for the user is neglible, I don't see any alternatives with comparable compatibility and/or performance. I consider the situation less than ideal, but better than nothing (or substantially less).
I don't mind having Alien Dalvik as an optional thing you can choose to have pre-installed on your device, but I don't want it to be something everyone has to pay for even if they are not interested in running fart apps without having to reboot. (Just kidding about the fart apps, I realize that there is more than that in Android ;) )
I understand this stance, but consider the actual absolute magnitude of the downside per user who doesn't use it (from zero [no cost] to small negative [small cost]). Now consider the actual absolute magnitude of upside per user for those that will use it (from zero [not able to run android apps] to big positive [able to run most andorid apps]). Compare these two. I'll draw a parallel to hardware features people don't want to pay for because they aren't going to use them. Everyone pays for them, they may be proprietary, some won't use for them, others will benefit from them. There may or may not be non-proprietary solutions of varying applicability and/or performance that may or may not entirely make the entire feature useless even for those that wanted it.
I consider "practically good, ideologically neutral, neglible cost" a win. (legend: "practically good" - brings new uses in addition to existing ones, "ideologically neutral" - Doesn't require the rest of the system to be less free, but isn't free itself, "neglible cost" - per unit, maybe)

(disclaimer: I do not have any idea on the pricing of AD. All this is based on the assumption that it can't be enormous given that a small company like Jolla can manage it. I'm guessing the amortized price per unit is neglible)
 
It's not just the cost I'm worried about. If the cost is low enough (~1 EUR per device or so) I wouldn't really care about that. I'm more worried about long-term support and things like: will we be able to upgrade stuff (in particular, the kernel) without trouble? It would really suck if maintaining compatibility with some proprietary software would become something that holds us back from moving forwards in the future.
AFAICT, Alien Dalvik is a userspace application that just happens to run android apps. Why would it be trouble? Or have I missed something?
The potential trouble could be that it may have certain dependencies (e.g. only works with a specific kernel version) and when at some point in the future, some non-backwards compatible potential upgrade breaks the dependency, we end up not having the upgrade because we don't want to break the dependency and the Alien Dalvik company no longer exists or charges a lot of money for an upgrade to fix the problem.

I agree that being able to multi-task between Android and Linux applications is nice, but I don't think it's really that important -- at least not for playing full-screen games. I think it would make more sense to put Replicant on a separate partition for now, so at least you get a separate Android that works well instead of an integrated Android that may have issues. I would much prefer to send some dev boards and/or money to the Replicant devs or some other FOSS devs to make an integrated Android work in Debian ARM, than to pay some company to give us some binary-only black box that magically works now but does not necessarily get updated and does not allow to be tinkered with.
 
Back
Top