Talk:Android (operating system)/Archive 6

Archive 1

←

Archive 4

Archive 5

Archive 6

Archive 7

Archive 8

Is Android Unix-like (without rooting)?

I took Unix-like out as, I define Unix-like as "Unix user program"-compatible, and Uniwersalista, reverted. I hesitate to revert him, as he could be right. In some sense it seems more Unix-like than I thought, but it might still be less wrong than right. See: Talk:Unix-like#What_constitutes Unix-like. Please discuss the general issue there (eg. we would agree Firefox OS is not Unix-like) but Android specific here. I installed some terminal (Dalvik, not "Unix"-program, some seem less capable without root) and could do:

ls

but not:

ls |wc -l

as wc was not available. I can not(?) install wc with Google play? apt-get? compile? How would I get my Unix programs in (without root)? Clearly the Unix way is at least not emphasized and while it seems in has pipes what can you pipe into? comp.arch (talk) 12:26, 31 March 2014 (UTC)

To me, Android is Unix-like, despite the fact it doesn't run standard shell commands out-of-the-box. If Android, as such, isn't Unix-like, then many bootable Linux installation CDs/DVDs aren't too, as many of them are missing some of the standard shell commands as not compiled into their busyboxes etc. Of course, all that can be added. Hope it makes sense. — Dsimic (talk | contribs) 22:20, 31 March 2014 (UTC)

But you can easily add the programs (if not might be based on Linux kernel but not an OS, see below). Maybe nobody just wants to with Android as it is not targeted as a command line OS. I'm just not sure you can without porting. If you need porting you may as well say its a different "OS family" (you can port Unix programs to Windows too). comp.arch (talk) 11:46, 1 April 2014 (UTC)

What do reliable sources say. If reliable sources say it is Unix-like we can report that in the article. If other reliable source say it is not, we can report that to. If they are silent on the subject then we should say nothing. —Jeremy (talk) 23:03, 31 March 2014 (UTC)

One reliable source even says that Android is a Linux distribution. It can't be that without being Unix-like, right? — Dsimic (talk | contribs) 23:09, 31 March 2014 (UTC)

I say it can. The Linux kernel only is not an "Operating system", you can't boot into one and run user supplied programs. Since the beginning Linus used bash with his kernel to make it Unix-like. All traditional Linux distributions have used bash or a some shell replacement to maintain it's Unix-likeness. If you take that out or something fundamental it can still be "Linux" but no longer Unix-like. Does Android allow installing "Unix programs" (runnable from a shell, leave X-Window programs out of this)? As programs have to use Dalvik, isn't there porting involved? comp.arch (talk) 11:37, 1 April 2014 (UTC)

Android programs don't have to strictly use Dalvik, that's why Android NDK exists, for example. You can run any Linux program on Android if you also install all of the dependencies, for example by chrooting into a whole Linux distribution; you can even get something like that from the Play Store. — Dsimic (talk | contribs) 20:04, 1 April 2014 (UTC)

I know about the NDK, but apps have to start in Dalvik first. I'm not sure if the porting is easy. Will only a generic "wrapper" do, that would work for (all) C standard Unix programs? Still would be a hassle and I wouldn't call it supporting Unix. Clearly it is not meant to (some C code yes, however). The Debian link you gave requires root so it's Android is not "supporting" Unix there. The other link is less clear but I clicked on a similar project from it that says: "THIS APP REQUIRES ROOT : ADVANCED USERS ONLY". In any case at least Android is not Unix like without hacking and/or modification. comp.arch (talk) 09:14, 2 April 2014 (UTC)

I wouldn't make decisions based on whether hacking or modifications are required or not; if we had no hacking around, we'd have no things such as Linux or various BSDs. If we'd take that route, many bootable Linux installation CDs/DVDs also couldn't be classified as Unix-like (what wouldn't make much sense, in my humble opinion), as they lack pretty much everything Android also lacks out-of-the-box; Ok, they don't run Java applications. :) By the way, this thread is quite interesting, and it pretty much concludes that Android isn't Unix-like. — Dsimic (talk | contribs) 14:41, 2 April 2014 (UTC)

I would say if sources do not agree with me they are not reliable :) The reliable sources would confirm this.. comp.arch (talk) 12:00, 1 April 2014 (UTC)

Czarkoff obviously disagrees. Perhaps we can discuss it further? — Dsimic (talk | contribs) 02:33, 16 April 2014 (UTC)

Yes, I obviously disagree:

1. One may use Android market app to install missing parts of Arch, Debian or Ubuntu userland on Android and use it as any other Unix-like system – even as a mobile failover node for production server.

2. There is "prior art" to Android+Dalvik combination: NeXTSTEP also had huge custom runtime on top of Berkley-style Unix, but even Jobs admitted it was still Unix.

— Dmitrij D. Czarkoff (talk•track) 06:59, 16 April 2014 (UTC)

1. I might be wrong, but a lot of the apps I see require root (that isn't available by design). Can you point me to one non-root app that adds any of these? Still is it Unix-like until then? Is Windows NT (already POSIX compatible) Unix-like before adding Windows Services for UNIX?

2. I have to look into NeXSTEP, did it only have a framework that most programs used or had available but was optional? Starting inside Dalvik in non-optional in Android for user-supplied programs. It seems JavaOS-like to me :) More than Unix-like.

3. Android doesn't have a CLI, where can I go in and compile my own Unix programs from source that is very much Unix-like to me..

I do not have much time now, but would really like to know more. What I found so far: [1] [2] [3]. comp.arch (talk) 09:18, 16 April 2014 (UTC)

1. Of course they require root for installation, as does every Unix-like system. FWIW installing software from Play Store also requires root, and corresponding permissions are granted by infrastructure.

2. NeXTSTEP is now MacOS X, so yes, it had framework that nearly every program used. Of course there were cat, sed, cp, rm and other utilities that are present on most Unix-like systems (including Android). MacOS X and Android are absolutely identical in this regard.

3. Android doesn't have pty support because it wasn't compiled in. You may compile custom Linux kernel without pty support as well. Nothing of difference. (FWIW my phone has terminal emulator as stock package.)

4. If you replace bionic with OpenBSD libc in year 2000 in that article, it will mostly hold true; if you replace bionic with System V libc, it will become a huge overstatement. Also note: one could replace bionic with modified musl, install toybox and enjoy reasonably POSIX-compliant Android.

— Dmitrij D. Czarkoff (talk•track) 11:38, 16 April 2014 (UTC)

1. Sorry, I meant to say without rooting (from section title) the device first. Android is NOT meant to be Unix-like, while it still uses the Linux kernel. While Android does have root-account (superuser) internally, it doesn't have the concept of root/superuser as a privileged status that the user can attain. Devices are locked down in that sense. You can (in some devices) install APK-files not from Google Play, but it is the only format of programs allowed. That, as far as I know, does not allow you root-access or allowing user supplied program to not start without using Dalvik (in a JVM-like environment). That is user programs (at least user supplied, not all internal) run in a JVM-like environment making the os more "JavaOS-like" than Unix-like, as the point of operating systems is to run programs. At least the OS would be a hybrid of both; more JavaOS-like than Unix-like. I believe NeXSTEP didn't try to exclude Unix-programs. Then we can also disagree on the Family-part of "OS Family", as a concept, elsewhere. I say we should stop using the field in Infoboxes.

2. NeXSTEP probably had theses and more utilities, but while Android has some, eg. cp, my phone does not find sed (or locate, for locate sed). I could add programs like this on OS X; and on Debian with apt-get but I believe I can't in Android with APK-files, weather from Google Play or not. And the OS allows no other way unless changed/broken by rooting (that is not "allowed") first.

3. What is compiled in stock Android, or more correctly, what is in AOSP, defines Android as an operatings system. Not what can be added, a custom-kernel, or other things requiring rooting first. BTW, what is your phone?

4. You seems to be saying Android is not "reasonably POSIX-compliant" unless some things are changed. If you change enough software (but I think you can't with stock unrooted Android) you get Windows Phone. If you add to Windows NT you get "Unix" (compatible, Unix-like) :) comp.arch (talk) 14:54, 16 April 2014 (UTC)

1. You overly concentrate on rooting issue, but it is merely security policy. Android has the same security model as every other Unix-like: you can't administrate it beyond some extent unless you gain root permissions. Things work exactly the same way with any other Unix-like system, where you can do whatever you want at $HOME and need special permissions to perform administrative tasks elsewhere. "Android rooting" as a topic is a strict subset of "privelege elevation on Linux".

2. Sure you can add whatever to your Android phone. I am not sure you can do it with standard software, but it has nothing to do with OS family – it's mere package management deficiency. FWIW older Unix systems didn't have any kind of package management, so Android isn't even the most limited in this regard.

3. Wrong. AOSP is just one of numerous Android distributions, just as Ubuntu is one of GNU/Linux distributions. You should also note, that if you want to pretend that Android operating system is defined with AOSP, you'll have to admit that most "Android phones" don't run Android – all of vendors ship more or less customised firmware, and this customization goes far beyond skins and vendor apps.

4. Nonsense. If I do the same on my Arch installation, it will still be Archlinux, Linux-based Unix-like operating system. Why same isn't true for Android? Sure, you may make up your own personal definition of Android that would have all the properties you want it to have, but reality won't change: Android is another Unix-like system.

— Dmitrij D. Czarkoff (talk•track) 17:28, 16 April 2014 (UTC)

Having read and considered the arguments above, I would say that Android is indeed Unix-like. A quick google found at least one RS source that seems to agree too: "both Apple‘s iOS (and OSX come to think of it) and Google‘s Android are variations of the basic Unix operating system." (Forbes) --Nigelj (talk) 18:09, 16 April 2014 (UTC)

1. The difference with regular Unix[-like], is that you are allowed (if you have permission) to gain root. On Adroid you are not allowed even with apps with all permissions they are can't do anything BY DESIGN, if they could break out of the sandbox it would be a privilege escalation: "exploiting a bug, design flaw".

2. Yes, you used to just compile from source, programs were (mostly at least) source code compatible, you can't compile (or cross-compile) regular Unix programs (let alone X Windows applications, but let's not go there) APK-files - the only "programs" allowed in. At least without porting: "porting is the process of adapting software so that an executable program can be created for a computing environment that is different from the one for which it was originally designed."

3. Google's Android (AOSP) is THE official "distribution". To be called Android (get the trademark) you have to be compatible, and I assume it also means with the security restrictions, see above. What is changed with the "clones", like Fire OS and Nokia's is just a new store. If anything they are more locked out (eg. Fire OS). There might be exceptions, "Android-based" but not "Android". I admit, this article is a little schizophrenic, trying to describe AOSP, the source code (FLOSS, not proprietary) and "practical Android", what you get as "binaries" (includes proprietary).

4. You can add a lot to an operating system (or other things) and it's still the "same thing". Doesn't mean you can remove stuff (or restrict) and still claim it's the same thing or similar enough. A typical Linux distribution is Unix-like, not if you remove everything except the Linux kernel, that alone is not Unix-like. They added stuff back, enough for some (most of) POSIX compatibility usually associated with Unix-like, but left out stuff eg. POSIX.2 it seems. The main point is that Android is meant to be different. See also Unix philosophy. comp.arch (talk) 22:35, 16 April 2014‎

I think what we have here is people saying "It's Unix-like" and one person replying, "But it's not full-blown UNIX". Yeah, it's been locked down, trimmed down, and turned into a phone or a tablet. The point still is that it has a Unix-like kernel, and trimmed-down Unix-like is still Unix-like in that all kinds of other Unix-like things could be added back (but why would you want to? It's a phone, FFS). Here's another ref: http://www.levenez.com/unix/ - you'll see Android appear, forked out of Linux 2.6.23, in 2007. --Nigelj (talk) 22:12, 16 April 2014 (UTC)

I actually make some use of command line on my phone. — Dmitrij D. Czarkoff (talk•track) 23:24, 16 April 2014 (UTC)

"Note 1: an arrow indicates an inheritance like a compatibility". Android is compatible with the Linux kernel (there might be a caveat, maybe not best link but would this break Unix-like illution?), that is not a full OS. "Note 2 : this diagram shows complete systems and [micro]kernels like Mach, Linux". The Linux kernel is not a complete system (OS), I have no objection to saying "Linux kernel-based". Would you say the Tivo is Unix-like? It can be made to if hacked, and the rest of the OS added. comp.arch (talk) 15:48, 17 April 2014 (UTC)

1. Plain wrong: superuser priveledge management app for Android is accepted on Google Play. Obviously Google disagrees with your point about no root by design.

2. FYI there are many Unix-like systems that have the same properties: all BSD install media, Linux and Busybox-based initrd combination, etc. All of these constitute proper Unix-like environments. Also note, you may install terminal emulator on any Android device from Play Store, compile any Unix application using Android NDK, package it into APK, install via default installer for third-party APKs and use this application.

3. You mistake branding for operating system. Ironically, the term "Unix-like" was coined specifically to circumvent "UNIX" branding while describing OSs.

4. Every OS ever was meant to be different. Most Unix-like OSs aren't fully POSIX-compatible. Android is less POSIX-compatible then MacOS X, but more compatible then AT&T Unix System V. In all other aspects Android is as much Unix-like as MacOS X is.

— Dmitrij D. Czarkoff (talk•track) 23:24, 16 April 2014 (UTC)

Thanks for pointing this out.

1. See about that app: "!!! SuperSU requires a rooted device !!!" comp.arch (talk) 15:48, 17 April 2014 (UTC)

You've missed the point: SuperSU is accepted to Google Play, resides there for ages and enjoys popularity. Obviously Google doesn't regard it as something going against Android design, because it would not be allowed otherwise (see 4.4 in Google Play Developer Distribution Agreement). — Dmitrij D. Czarkoff (talk•track) 23:50, 17 April 2014 (UTC)

I'm not sure what this program does on a non-rooted device ("Android"), probably nothing, it can't by design. On those devices it is no threat. It need a rooted device (modified Android, that in my view is not "Android"). "development or distribution of Products, that interferes with, disrupts, damages, or accesses in an unauthorized manner the devices" is about Google Play (that comes often with Android, but is not part of Android (AOSP)), not Android OS (design) per se. comp.arch (talk) 16:18, 18 April 2014 (UTC)

2. "you may [..] compile any Unix application using Android NDK, package it into APK". Is this really true? There are two possibilities: a) As Unix predates GUIs, we are talking about Unix-like programs that work on text files or streams, such as "wc", "sed", or applications such as "vi" or "emacs". [b) If you would count X Windows applications as Unix-like, there would be huge amount of porting.] Can you compile wc into an APK in install, and sed into an APK ind install (and maybe also vi). Can you then run "ls |wc -l" or "ls |sed "something" |wc -l" eg. ? That is the Unix way. My doubt is that you can do this. And in addition see the comment from the link you provided: "It's reasonably common to launch native executables from Android apps, but you should be aware that it's frowned upon by the Android architects. They keep making threats to disable that facility in future, as they're unable to regulate the resource usage of those processes as precisely as they do for the main app process." I suspect this running "Unix-binaries" only applies to limitted cases and only for porting, and then you are really running Android applications, supporting my assertion the Android is its own "OS Family" ["http://www.roman10.net/how-to-build-ffmpeg-for-android/]. comp.arch (talk) 17:51, 18 April 2014 (UTC)

1. You are still missing the point: if Android doesn't allow user privelage elevation, SuperSU is not Android software. The fact that Google accepted it to Play is Google's unambiguous statement that you are wrong.
2. AOSP includes mksh – POSIX-compliant Unix shell, – so you can do "ls | wc -l" and every other Unix shell scripting tricks. You can use vi. (It was preinstalled on all phones I used. I assume it is included with AOSP.) The fact that Google wants to remove support of direct executable calling from Dalvik says nothing about Android as OS – it only says something about Dalvik. (And AOSP ships with many software pieces that have nothing to do with Dalvik.) You don't need to port software to Android – only recompile it (for proper linking). And even if you had to port software to Android, it wouldn't be anyhow different from other Unix-like systems: you have to port software from Linux to BSDs or OSX, and porting graphical OSX software to other Unix-like systems means complete rewrite in most cases, which doesn't stop MacOS X from being fully POSIX-compatible certified Unix.

Do your homework. — Dmitrij D. Czarkoff (talk•track) 18:16, 18 April 2014 (UTC)

Did anyone suggest just saying "Linux-like" instead? ViperSnake151  Talk  18:22, 15 June 2014 (UTC)

Of course not, "Linux-like" is meaningless. There were just 3 days remaining before this thread would get archived... — Dmitrij D. Czarkoff (talk•track) 00:37, 16 June 2014 (UTC)

Well I wasn't going to prolong this discussion, but more than a month ago my template proposal Wikipedia_talk:Articles_for_creation/Template:Unix_internals got accepted. I had been waiting a long time for it and forgotten about it. Would that be a middle ground? comp.arch (talk) 08:57, 27 June 2014 (UTC)

Sorry, what exactly do you mean? |family=Unix APIs? P.S.: I've started TfD discussion for your template, as it is redundant to {{Unix}}. — Dmitrij D. Czarkoff (talk•track) 09:29, 27 June 2014 (UTC)

I proposed this template at the time so that Android and iOS (and Firefox OS?) could be included in the same template with with Unix-like-OSes. Since then, "Without command-line interpreter" got added to {{Unix}} (by me for clarification, that they are a special case). By "Unix internals"/APIs I mean the Linux kernel or some other kernel as in iOS (that is excluding user-space). I would have liked the the other template to only include (out-of-the-box) full (with CLI) Unix-like. Firefox OS is still a problem(?) as while it has the Linux-kernel APIs, it's probably not accessible from any program by design (but it's still there so it's inclusion in the template is technically true..). comp.arch (talk) 09:56, 27 June 2014 (UTC)

Android uses the GPL version 3 license and Android 4.4.4 is Linux kernel 3.4.46 [4]. Android even runs SELinux for Mandatory access control and the Linux EXT4 file system. Renegadeviking 9/10/2014 11:26 am

Articles for individual versions

Well, if iOS can do it, than so can we. I've been starting work on seperate articles for each recent Android version, going by my work on Windows-related articles as an example (development, release, features, reception). I've done two so far, L, and Jelly Bean. Personally, I would prefer that these pages use their base version numbers as the title (i.e. Android 4.0), unless the specified version does not yet have a number (L), or a single codename covers multiple versions and the other versions do not contain enough substantial changes to justify a new article (Jelly Bean).

Think anyone could provide some help? ViperSnake151  Talk  00:10, 3 July 2014 (UTC)

Market share

Main article: Mobile operating system § Market share

Don't mean to be a market crier, but I think these diagrams are somewhat more impressive then a mere number in a text. Maybe the article should adopt them:

Android market share on smartphones

• 
• 

Cheers User:ScotXW^t@lk 20:59, 3 July 2014 (UTC)

I like graphs as much as anyone else but these are effectively unsourced... – Steel 19:23, 4 July 2014 (UTC)

Why is Android not a subsection of Linux?

I see no other Linux based OS's having their own entry? — Preceding unsigned comment added by 92.18.254.78 (talk) 20:15, 5 July 2014 (UTC)

What "Linux" do you mean? There is Linux kernel, which is currently used in several operating systems:

• loosely defined unnamed general purpose OS, which is referred to as Linux, GNU/Linux or (at least by rms) GNU+Linux; most operating systems based on Linux kernel are varieties of this one, eg. Access Linux Platform, Chrome OS, Firefox OS, Sailfish OS, WebOS;
• mobile OS, whose most prominent variety is called Android;
• mobile OS called MotoMagx by Motorola.

There are other operating systems based on Linux kernel which I don't know much about, so I am unsure whether they are varieties of GNU/Linux or not, eg. WebOS. There are Linux-based operating systems I did not recall while writing this response.

Did you notice the amount of blue links? — Dmitrij D. Czarkoff (talk•track) 22:11, 5 July 2014 (UTC)

 It was Linux 3 months ago, but now its UNIX-like. Android 4.4 is running Linux kernel 3.4.  You should fix the malfunctioning bots.

http://www.ibtimes.com/android-442-kitkat-update-samsung-releases-kernel-source-code-sprints-galaxy-s3-ota-1578851

https://source.android.com/devices/tech/security/enhancements44.html

renegadeviking

Semi-protected edit request on 29 September 2014

Srimanand (talk) 13:50, 29 September 2014 (UTC)

Not done: as you have not requested a change.
If you want to suggest a change, please request this in the form "Please replace XXX with YYY" or "Please add ZZZ between PPP and QQQ".
Please also cite reliable sources to back up your request, without which no information should be added to, or changed in, any article. - Arjayay (talk) 14:16, 29 September 2014 (UTC)

Unexplained delete - about Google Play (and F-Droid)

I noticed an edit that seemed to be an addition/rearrangement, but that deleted "Purchases of unwanted applications can be refunded within 15 minutes of the time of download". It may have been in error. I notice however the reference says (now at least) two hours.. Yes, Google Play isn't strictly Android.. however it is mentioned here a lot and maybe this should be in. Same user deleted an addition I made to F-Droid. Should all appstores be mostly out, or how much in and to which extent? comp.arch (talk) 11:28, 7 October 2014 (UTC)

I don't have a strong preference for having that content or not, but it might be better not to include "purchase timeouts" as they might be changed at some point in time, and who knows what else they depend on? Maybe even on the country from which the purchase was made? Just guessing. — Dsimic (talk | contribs) 11:34, 7 October 2014 (UTC)

Yes, memory requirements might also change (and have) for Android the OS (in new versions) :) Yes, those are more integral and timeouts seem to have changed (but only for the better?). I very much doubt they would change for the worse.. I wouldn't expect the timeout rules to be dependent on anything, e.g. country, in the past or the future. Unlike movies etc. apps are much less restricted by country and I see no good/logical reason to discriminate or any indication that they do[5]

Regarding F-Droid, info on non-proprietary app stores might be useful. I just put that in when correcting that alternative apps stores are proprietary. This article is a little schizophrenic with where to draw the line, some want it to be about the OS only ("it's open source") others the whole "ecosystem" - Android wouldn't be where it is at without Google Play/Market, how much of it or alternative app store info (and Google's other (API) addons) is ok here? comp.arch (talk) 13:39, 7 October 2014 (UTC)

Your comparison with the Android's memory requirements is a good one. Having that in mind, on second thought I'd say that restoring the info on "purchase refund timeouts" would be a good thing; maybe you could also try to find and provide one or two additional references. — Dsimic (talk | contribs) 18:26, 7 October 2014 (UTC)

Semi-protected edit request on 9 October 2014

The sources of the following sentence: "The chief beneficiary of Android has been Samsung, whose Galaxy brand has surpassed that of Android in terms of brand recognition since 2011.[199][200]" are:

[199]: http://www.slashgear.com/watch-out-google-samsungs-galaxy-brand-has-eclipsed-android-05268159/ However, this page only shows that in Google Trends, "Galaxy" got more attention than "Android". It is not a study on brand recognition.

[200]: http://appleinsider.com/articles/13/03/15/samsungs-galaxy-s4-distracts-attention-away-from-android/page/1 This is an article mainly on the topic of risks and weaknesses on the Android ecosystem. It mentions that Samsung hardly used the word "Android" during the Galaxy S4 launch: "In showing off the new Galaxy S 4, Samsung didn't draw attention to Google's search, mapping or translations services, nor even its ability to download Google Play content. It didn't feature Android apps." which however doesn't show the the *recognition* of Android is lower or higher than that of "Galaxy". There is only one mentioning of the recognition: "Many observers have noted that Samsung's Galaxy brand already has as much name recognition as Android itself. " However, the source her is "many observers".

Since the sources do not contain any evidence that Samsung is more (or less) recognised than Android, I request to add to these links.

78.42.228.24 (talk) 21:41, 9 October 2014 (UTC)

It's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format. EoRdE6 (talk) 02:12, 11 October 2014 (UTC)

I've removed this sentence on the basis that the current sourcing is probably insufficient to justify the claim that, amongst other things, Samsung is 'the chief beneficiary' from Android. Two other sentences in this paragraph are unsourced and have also been removed pending discussion. – Steel 17:11, 11 October 2014 (UTC)

My request was: "I request to add [not in citation given] to these links." Removing the line entirely would also be OK for me, but seems a bit risky since I did not prove it to be wrong. — Preceding unsigned comment added by 194.39.218.10 (talk) 07:43, 13 October 2014 (UTC)

Semi-protected edit request on 15 October 2014

Latest release: 5.0/Lollipop on 15 Oct 2014, today.

Nashonida (talk) 18:07, 15 October 2014 (UTC)

 Not done: Sorry, Lollipop isn't released yet, so it can't be listed as the latest release. — Dsimic (talk | contribs) 18:13, 15 October 2014 (UTC)

Semi-protected edit request on 16 October 2014

Update Android Version 5.0 180.151.225.3 (talk) 03:50, 16 October 2014 (UTC)

 Not done: Please see the request above; Lollipop isn't released yet, and until then it's covered in Android (operating system) § History section as the announced next major Android release. — Dsimic (talk | contribs) 03:55, 16 October 2014 (UTC)

Migrating apps? ("Nexus Player") - "out-of memory handling"

"we’re releasing the first device running Android TV: Nexus Player, a collaboration with Asus, is a streaming media player for movies, music and videos. It's also a first-of-its-kind Android gaming device. With Nexus Player you can play Android games on your HDTV with a gamepad, then keep playing on your phone while you're on the road."[6]

Now, this is, to me, one of the most interesting new development. And I don't play games much.. While the Nexus Player supports streaming/Chromecast functionality, it seems from your phone, it seems not to need a phone to run the game itself(?). What this all could mean is that you would have to install the same app on two devices and that the game developers have to make this all work for new apps. That is, you would save and continue at another device. To me, no additional programming (or minimal if hardware is different(ly capable)) seems necessary (in theory) and no need to "save"(?) directly - could migrate mid-game/add and be backward compatible. It's an idea I've had for a while, Android apps already have to handle being thrown out (eg. out-of-memory) and get restarted and then it's the app responsibility to make it work (transparent always?). What makes it hard to restart the app on another machine? Anyone seen this hacked together before in an alternative Android ROM? Is the how the Nexus Player does it?

Anyone know how other OSes handle throwing out apps (when memory is low)? This is not something regular Ubuntu/Linux apps have to handle. Anyone know if Ubuntu Touch added this requirement? iOS does similar? Not regular Windows. But added in Windows Phone I would guess (and newer combined app-model?).

While I've had an argument here about Android being Unix and found out Android can be make to run Unix apps, this still seems to be an "exception" (of course memory can run out in Linux and your app gets killed, you rely on enough memory and use virtual memory unlike Android, but if it runs out a random app can get killed). comp.arch (talk) 22:05, 18 October 2014 (UTC)

Hasn't Google released additional APIs required for storing game scores and whatnot? Google probably waited for those APIs to gain traction, and now it tries to capitalize on that by releasing a device that "provides" such a functionality. Regarding how out-of-memory conditions are handled, well, you need to terminate some application when there simply isn't enough RAM; what happens to applications is, in general, pretty much undefined. — Dsimic (talk | contribs) 07:26, 19 October 2014 (UTC)

I was getting at Android's "Activities": "The system will kill less important processes (the last ones) before it resorts to killing more important processes (the first ones). [..] so the system may safely kill its process to reclaim memory for other foreground or visible processes"[7][8] That is if the foreground process finishes memory, in regular Linux (foreground or not) it may get killed; under Android it is less likely to happen, first all other Android apps will be "killed" to free up memory. Unlike in regular Linux they are supposed to start again as nothing happened when switched to, right? For a game the same mechanisim could be used just on another machine? comp.arch (talk) 14:54, 19 October 2014 (UTC)

Well, there's no Android magic that would revive terminated applications back to the state before termination. Just try terminating an application yourself, and you'll see. For the Android games and "seamless" playing on multiple devices, I'd say that it's all about the games using Google APIs for saving scores and positions, nothing more. — Dsimic (talk | contribs) 20:54, 19 October 2014 (UTC)

Just not true, the Android "magic" is that you are supposed to handle being killed by the system (maybe not all (many?) apps are programmed as they should..). If the user askes for it I guess there is a different method. look at the Activity Lifecycle diagram and the text in the source I gave: "If an activity is paused or stopped, the system can drop the activity from memory by either asking it to finish, or simply killing its process. When it is displayed again to the user, it must be completely restarted and restored to its previous state."[9] comp.arch (talk) 13:57, 20 October 2014 (UTC)

You're right, I was wrong and I stand corrected. Moreover, please have a look at the Recreating an Activity document, here are a few quotations from it:

• The system may also destroy your activity if it's currently stopped and hasn't been used in a long time or the foreground activity requires more resources so the system must shut down background processes to recover memory.
• However, if the system destroys the activity due to system constraints (rather than normal app behavior), then although the actual Activity instance is gone, the system remembers that it existed such that if the user navigates back to it, the system creates a new instance of the activity using a set of saved data that describes the state of the activity when it was destroyed. The saved data that the system uses to restore the previous state is called the "instance state" and is a collection of key-value pairs stored in a Bundle object.
• By default, the system uses the Bundle instance state to save information about each View object in your activity layout (such as the text value entered into an EditText object). So, if your activity instance is destroyed and recreated, the state of the layout is restored to its previous state with no code required by you. However, your activity might have more state information that you'd like to restore, such as member variables that track the user's progress in the activity.
• To save additional data about the activity state, you must override the onSaveInstanceState() callback method. The system calls this method when the user is leaving your activity and passes it the Bundle object that will be saved in the event that your activity is destroyed unexpectedly. If the system must recreate the activity instance later, it passes the same Bundle object to both the onRestoreInstanceState() and onCreate() methods.

Thus, there is some magic involved with restarting activities killed due to OOM conditions, and it could be that Nexus Player will be using exactly that for "seamless" gaming on multiple devices. In addition, please have a look at the R.attr documentation, there's even more "magic" around persisting activities across device reboots; here are two quotations from it:

• public static final int persistableMode

Define how an activity persist across reboots. Activities defined as "never" will not be persisted. Those defined as "always" will be persisted. Those defined as "taskOnly" will persist the root activity of the task only. See below for more detail as to what gets persisted.

• persistAcrossReboots

If this activity forms the root of a task then the task and this activity will be persisted across reboots. If the activity above this activity is also tagged with the attribute "persist" then it will be persisted as well. And so on up the task stack until either an activity without the persistableMode="persistAcrossReboots" attribute or one that was launched with the flag Intent.FLAG_CLEAR_TASK_WHEN_RESET is encountered.

Activities that are declared with the persistAcrossReboots attribute will be provided with a PersistableBundle in onSavedInstanceState(), These activities may use this PeristableBundle to save their state. Then, following a reboot, that PersistableBundle will be provided back to the activity in its onCreate() method.

However, it's pretty much unclear what happens with other application components, as not only activites make up an application. — Dsimic (talk | contribs) 15:02, 20 October 2014 (UTC)

Thanks for the info. It just seems to work (or I have enough memory.. and the mechanism isn't activated, I have 1 GB (but 256 MB in an older 2.2 Froyo phone I no longer use)). It's one of my favorite Android features.. making it better than Unix/Linux. The $33 dollar 128 MB Cloud FX Firefox OS phone has memory troubles (background apps such as the alarm clock stopping and not restarting..). I assume web apps to not have to handle anything like this, at least not general JavaScript. Seems maybe FFOS is doomed.. Anyknow know about similar features in the other OSes; iOS, Unity Touch, Windows Phone? comp.arch (talk) 18:32, 20 October 2014 (UTC)

"Android holds the largest number of installed-base devices"

@Steel: already Gartner seems to say Andoid IS most popular: "In the OS market, Android continues to be the OS of choice across all devices (see Table 2). Gartner estimates that Android will reach 1.1 billion users in 2014, a 26 percent increase from 2013. "There is no doubt that there is a volume versus value equation, with Android users also purchasing lower-cost devices compared to Apple users. Android holds the largest number of installed-base devices, with 1.9 billion in use in 2014, compared with 682 million iOS/Mac OS installed-base devices," said Annette Zimmerman, principal analyst at Gartner."[10]. I say "probably" to be on the safe side (because the 2014 numbers are estimates). Note the 1.1 billion number from Gartners is number for sold devices in 2014 only (estimate) and the installed base of 1.9 billion then would only make sense by adding up 2014 and 2013, and all 2012 and older Androids thrown away!

According to the table Windows seems to have a larger installed base than iOS/Mac OS considering either two or three year old devices but not mentioned.. Their "682 million iOS/Mac OS installed-base devices" seem not to match the table if as old devices as Android are kept. Maybe they consider PCs to have a life of three year and mobile two year? Still, Windows PCs would then be fewer than Android (at end of 2014 at least). And 1.9 billion > 1.5 billion WxWorks.. (and sources seems to say more than other RTOS). comp.arch (talk) 11:42, 9 September 2014 (UTC)

Now Gartners's rapidly declining "Others" category is not enough to be bigger than Android if kept for three (or two) years. I can only assume that it's "feature phones". However, it is/was bigger than Windows! Now, it may not be only one OS (Symbian?), it's probably fragmented in at least two big chunks, S40-non-Symbian (eg. Asha) and S60/Symbian. Another thing to inflate Windows is piracy. I took that into account by trying to get installed base of PC not just Windows. Not all PCs however run Windows.. Another thing I'm not sure is counted under PC are servers (many run on PC category hardware but probably not most..). "PC servers" are split between mostly Linux and Windows, but I'm not sure the Windows server market is big enought to matter much (I would like numbers) compared to the much bigger client Windows market. Good sources are needed but I believe Android is most popular now and that it's verifiable (by my already given sources mostly?). Andoird vs. mobile PC (laptops) at least is interesting information. More recent estimates of PC installed base is needed. Mobile subscribers are over 6 billion (but not all smartphones). PCs ever produced are over 4 billion. comp.arch (talk) 12:02, 9 September 2014 (UTC)

Another possible error (if not counted towards PCs as I expect but did not consider) but not by much would by Windows tablets as of end-of-2013: 5,194,237=4,031,802 (2013) + 1,162,435 (2012) + 0 (2011).[11]

Since there was an error anyway (fixed now but have to check with Q1-2 of 2014 is enough to make Android most popular or if Q3 is also needed), premature for lead: "Android is the most popular operating system in the world, by installed base, with 1.331.87 billion devices sold (there of 126.3 million tablets) in 2012 and 2013 only, making it the highest selling tablet operating system in that timeframe and combined with smartphones the installed base is also higher than the estimated installed base for the traditional desktop/laptop PC operating system Microsoft Windows." Any idea how to improve text (when/if numbers higher)? Installed base of Windows desktop and mobiles estimated 1.919 billion at end of 2013 comp.arch (talk) 16:58, 10 September 2014 (UTC)

We don't do original research. If some reference says that Android has a larger installed base than X then we can use that. Also this is way too much detail for the lead and should just go in the market share section. Another thing is your run-on sentences are unreadable. Sorry, but you don't seem to have the English skills to be editing English wikipedia. Bhny (talk) 17:10, 10 September 2014 (UTC)

Simple calculations are not WP:OR (may not be if adding across refs?). I'll add to market share first. And/or here in the talk page. Ok, only overall picture in the lead. comp.arch (talk) 17:26, 10 September 2014 (UTC)

Making calculations and then coming to a conclusion that isn't in the ref is WP:SYN. The source should have the conclusion, especially with fuzzy things like installed base. Bhny (talk) 17:35, 10 September 2014 (UTC)

comp.arch I don't doubt your enthusiasm or knowledge of the subject matter, but your recent edits have been unintelligible and backed up on this page by obvious original research. I don't know how everyone else feels here but I would appreciate it if you slowed down a little sometimes. You tend to make edits faster than people can decipher the previous ones, and justify them with lengthy, rambling talk page comments that not everyone reads for their own sanity. – Steel 21:20, 10 September 2014 (UTC)

Fair enough, I just find it to be major news if Android is number one operating system and at least doing OR for me or to convince myself is ok. I accidentally overestimated Android (added up a year, reading as a quarter) I now taking that into account Android sold 415.4 millions devices MORE than PC is Q1, Q2 and estimated Q3. As I made an error previously just a little bit bigger it seems Android is about as big now (even without adding tablets). I'm using Q3 2013 as proxy for Q3 2014 that may underestimate Android vs. PCs much. Let's see if the media catches on to this and does their own math. Probably after Q3 results are in. comp.arch (talk) 21:35, 10 September 2014 (UTC)

"Android is the most popular mobile OS". "mobile" is redundant. Most popular doesn't have to mean by installed base (but probably true by now) only by sales of new devices. How many quarters are needed? I added a source that compares Android directly to Windows showing it outsell Windows 2.8:1 in 2013[12]. "As of 2013, Android devices sell more than Windows, iOS, and Mac OS X devices combined", actually Android sold more than Windows in 2012, starting with Q1. "This Chart Shows Google's Incredible Domination Of The World's Computing Platforms"[13]. In Q4 2011 it was a little behind with 82% of the Windows sales volume. After this quarter it will be three whole years of Android outselling Windows, by a lot cumulative. comp.arch (talk) 12:21, 21 October 2014 (UTC)

Semi-protected edit request on 22 October 2014

Sreerag Sanilkumar (talk) 04:00, 22 October 2014 (UTC)

Not done: as you have not requested a change.
If you want to suggest a change, please request this in the form "Please replace XXX with YYY" or "Please add ZZZ between PPP and QQQ".
Please also cite reliable sources to back up your request, without which no information should be added to, or changed in, any article. - Arjayay (talk) 05:51, 22 October 2014 (UTC)

Android's popularity - how is recent devices defined

@Steel: Regarding this edit. Since 2012, onwards it sells more than any other OS. Says so in market share section, see Gartner ref and after ; in lead. comp.arch (talk) 08:08, 27 October 2014 (UTC)

See my counterchange to the lead. Note, the edit summary: "most popular mobile OS" implies not most popular overall - hate to start with that (as untrue?).. Of course most popular mobile OS is also true, "untrue?" does not refer to that. I meant to say "hate to start with that as the implication that another OS is more popular is untrue?" comp.arch (talk) 09:49, 27 October 2014 (UTC)

[Note: for those who view Android as "Unix-like" and the Linux kernel the OS (I don't) then the kernel is most popular.] comp.arch (talk) 09:51, 27 October 2014 (UTC)

Thanks for clarifying. I've made another edit, which I hope will be acceptable to everyone, because this sentence was a total nightmare. – Steel 20:27, 27 October 2014 (UTC)

Semi-protected edit request on 3 November 2014

At Marketing target add

• Tv's (Android TV)
• Cars (Android Auto)
• Wearable devices (Android Wear)

91.220.150.10 (talk) 09:42, 3 November 2014 (UTC)

 Done Luxure Σ 09:50, 3 November 2014 (UTC)

Semi-protected edit request on 14 November 2014

| latest release version = 5.0 Lollipop | latest release date = November 13, 2014 (2014-11-13) Lgiessler (talk) 12:03, 14 November 2014 (UTC)

 Not done: you have not provided a source for the material you want added. G S Palmer (talk • contribs) 15:41, 14 November 2014 (UTC)

Semi-protected edit request on 15 November 2014

(For Paragraph "Update schedule")

Google provides major upgrades, incremental in nature, to Android every six to nine months, which most devices are capable of receiving over the air. The latest major release is Android 5.0 "Lollipop". aRvInD G.R. 09:24, 15 November 2014 (UTC)

Done Stickee (talk) 10:50, 15 November 2014 (UTC)

References

"All antivirus is ineffective" is just plain wrong

I noticed this article was protected so...

The extreme-tech article for this is not a good source. It's also based on the 2013 AV-test, which had dramatically worse performance by all AV apps than it does now.

Here is the current 2014 AV-test:

http://www.av-test.org/en/news/news-single-view/36-security-apps-for-android-are-put-under-constant-fire/

Note that 8 AV apps achieved a perfect score of 13 points, and 14 had a perfect detection rate in the tests. I don't know how that would be possible if "Android's sandboxing prevents effective scanning." Maybe extreme-tech should stick to benchmarking gaming hardware.

Hello! I've had a look at the above linked web page, and it doesn't describe what was actually tested and compared between various antivirus solutions. In other words, what should those "perfect 13-point scores" actually mean? Unfortunately, that isn't described, what makes it pretty much unusable as a good source.

Based on the already used reference and another one (which also provides a good insight), sandboxing prevents antivirus applications on Android from being effective and doing what's actually written on their boxes. — Dsimic (talk | contribs) 07:25, 18 November 2014 (UTC)

Semi-protected edit request on 8 December 2014

Please change "Be together. Not the same." to "be together. not the same.". https://www.androidify.com/en/#/ 83.37.233.232 (talk) 10:22, 8 December 2014 (UTC)

Done -Thanks for the improvement suggested. Anupmehra -Let's talk! 19:18, 8 December 2014 (UTC)

Android on vanilla Linux kernel

"I have pushed the android-3.18 branch of kernel based on the vanilla kernel 3.18 to the git server."[14]

Is this a first? I do not really care about showing in this article what exact 3.x kernel is currently used. But a mainlined kernel might be an exception. But have all Android's up to most current used "non-vanilla" kernel? When where the last Android-specific bits merged into mainline? comp.arch (talk) 21:38, 2 January 2015 (UTC)

To me, that looks like it's a vanilla 3.18 version of the Linux kernel patched with Android-specific stuff. — Dsimic (talk | contribs) 21:55, 2 January 2015 (UTC)

"Java" on Android

as per http://www.oracle.com/us/legal/trademarks/index.html, "Java" is a trademark of Oracle, and can *not* be used in contexts not directly related to either certified Java implementations, VM, language etc. As such, I'm re-editing the dubious parts.

Poponuro (talk) 13:11, 3 January 2015 (UTC)

Also, due to legal dispute (Oracle vs Google) I'd *strongly* suggest we *don't* use the word "Java" to describe Android OS/SDK/Java language flavor etc. Any such use should have a primary source (NOR), be it from Android or Google, stating so. Secondary web sources ain't a good choice here.

NB, some quotes to back that up:

"Oracle and Java are registered trademarks of Oracle and/or its affiliates."

From "Correct use" "Proper use of Oracle trademarks reinforces their role as brands for our products and services, and helps prevent them from becoming generic names that can be used by anyone."

From "Prohibited Use"

"You may not use Oracle trademarks in a manner which could cause confusion as to Oracle sponsorship, affiliation or endorsement."

Android is *not* affiliated with Oracle - as such, its API language *musn't* be called "Java", otherwise we're breaching the trademark law.

tl;dr : Calling Android API language "Java" is like calling Bing a "googling engine".

BTW, calling Android API language "Java" kind of contradicts this statement: "On August 12, 2010, Oracle sued Google over claimed infringement of copyrights and patents related to the Java programming language."

If Android apps *were written* in Java programming language, Android couldn't infringe anything *related to* Java programming language. Either Android essentially *is* Java, and thus the copyright and trademark has been infringed, or it *isn't* Java, the copyright/trademark hasn't been infringed and it *can't* be called "Java" then. Either way, it's inconsistent with current article content. Also, the decision *if* the CR/TM has been infringed isn't *ours* to make - it's already in court. Calling the API language Java-like is a) basically true, regardless if the language is or isn't "Java", b) a fair use of Java name (describes similarity); calling the API language "Java" is a) mostly probably factually incorrect (for many reasons), b) an infringement of Oracle's copyright/trademark policy. As such, I'm heavily in favor of calling it "Java-like".

NB I've already edited the article, but since the edit was reverted, I've just tagged the offending quote appropriately, please don't revert those changes there without previous discussion here and without providing strong evidence/sources first.

If no statements/citations/facts to back up that Android language *is* Java appear in the next couple of days/weeks, I'll re-revert it to "Java-like".

Poponuro (talk) 13:04, 4 January 2015 (UTC)

We just use common names and what the sources say. We don't make legal decisions. Bhny (talk) 16:27, 4 January 2015 (UTC).

Not only do sources say Android is programmed in the Java *language*, it is also technically true (unless I'm given a counterexample that is not about the API confusion). That must count for something (but note the truth is more nuanced). I took "-like" out, but the language of the article can maybe be clarified. Note: Java the programming language and Java including that, the full "environment"/API/JVM/etc. are not the same thing.

For a primary source, see: "Dalvik had separate stacks for native and Java code"[15][16] but note a primary source is not preferred in Wikipedia. Java was meant to be more than a language, see "write-once-run-anywhere", that part of Java is broken w/Android but that issue is about more than the language. See also: [17].

Trademarks are meant to prevent confusion in the market place to restrict competitors. Wikipedia isn't a competitor and I'm not sure Oracle has any obligation to clear up any confusion from non-competitors. Maybe the article should do it though, but maybe not as Wikipeda is not a WP:HOWTO and technical Java-details here maybe be to technical for this artice..? comp.arch (talk) 19:24, 4 January 2015 (UTC)

The source that was linked (CNet news AFAIR) explicitly stated that Android diverged from Java and that it's not Java per se - that's exactly the reason I prefer a primary source here not because it proves that Android is Java, but because it constitutes a rationale to call it that way.

Also, Android and Oracle *are* direct competitors - as such, I'd say that Wiki should have a most chilled approach to this matter, not giving in to possible OR about Android being Java. CNet is not a technical source, it's a news source. It's neither a scientific article or an in-depth analysis. It's OK for me to state that if primary Android pages state that - even if it's possibly factually incorrect, we're just stating what Google states about it. "Unless restricted by another policy, reliable primary sources may be used in Wikipedia" - I consider developer.android.com a reliable source when it comes to things Android-related.

Anyway, thank you for the http://developer.android.com/guide/practices/verifying-apps-art.html#Stack_Size link, it was exactly what was needed here IMO.

See WP:SECONDARY for why I suggested adding possibly primary source, not removing secondary: "A secondary source provides an author's own thinking based on primary sources, generally at least one step removed from an event. [..] Policy: Wikipedia articles usually rely on material from reliable secondary sources." Primary sources can be bad, in case people - like you did - say they are wrong :) News articles are ok - when they are right as I think they are in this case. Note "Java" can mean different things, the platform including computer code, e.g. JVM but can also mean *just* the programming language and I view programming languages in the restricted (meaning: specification/Backus–Naur Form) computer science terms (and I guess you can't trademark a specification. It is not the (runnable software) product). In that sense the JVM is excluded. And I *think* the API/frameworks should also be excluded, but while I know there are standards for them I'm not sure how much is regarded as part of the "standard library" of Java or if that should be included as part of the "programming language". Maybe that is *considered* more a part of the language than for others. Anyway, all this doesn't matter I think as neither source you have say "Java-like" and you seem ok with "Java" now. comp.arch (talk) 17:21, 5 January 2015 (UTC)

Semi-protected edit request on 24 January 2015

Reference 2 (http://source.android.com/about/philosophy.html) gives a 404 error. I was going to replace it with http://source.android.com/source/index.html#governance-philosophy but I guess someone else has to. Bl184999 (talk) 20:31, 24 January 2015 (UTC)

 Done, updated the dead reference. Thank you for pointing that out! — Dsimic (talk | contribs) 06:50, 25 January 2015 (UTC)

Google Play and APK

Android package manager on the single Apk. Google Play Apk file to your device, download and setup in the background making this information is incorrect. — Preceding unsigned comment added by Tolgarda (talk • contribs) 18:14, 3 February 2015 (UTC)

Hello! Could you, please, elaborate a bit further what you are referring to? — Dsimic (talk | contribs) 18:28, 3 February 2015 (UTC)

Android as a "PC" - e.g. right-click mouse support, and scroll-bars..

I've been thinking of using Android on my main laptop. I know WP isn't a WP:HOWTO nor the talk pages. Still, would you think more information on that (minority) use would be appropriate here? I'm just not sure if I'm search for something that isn't there or just hard to find.. at least googling for right-click mouse support Android apps (say browsers)..

There isn't really any theoretical reason Android couldn't support native WIMP apps.[18] But I understand if the market isn't huge for that.. Browsers/apps could add scrollbars and pop-up menus if a mouse is plugged in. Know of any that do?

I noticed people have asked for right-click support (I assume without a mouse) in Chrome for Android.[19].

Microsoft's Remote Desktop Client on Android: "The right mouse button is supported on Android version 2.3 and newer. Mouse wheel is supported on Android version 3.2 and newer. External keyboards are supported."[20][21]

"Plenty of Remote Desktop software can "decode" right click so obviously right click and left click are registered differently from android but as android lacks "right-click functionality" (it) replaces both as a left click (touch) trigger. The trick would be to replace "long touch" for right click."[22][23] Note this info seems to be the default or at least outdated info as app support is not the "fault" of the OS; see the first ref above.

Besides remote desktop apps or similiar (and maybe emulators[24]) that are really for accessing non-touchscreens apps, I have trouble with finding touchscreen apps with optional mouse support.. I only find "the opposite".[25][26] For me personally I'm getting sick of Firefox crashing (in Ubuntu) and not handling lots of tabs/windows. I want the Firefox for Android experience in desktop use - e.g. the CPU only focusing on one tab at a time.

I remember seeing a review of an Android PC that complained about right-click support. It was a long time ago. Is it just an outdated complaint? For this one, as a touchscreen is included, I see this in reviews: "doubt the keyboard and mouse will get much use, but is included"[27][28] but noting on right-click. I'll have no touchscreen.. comp.arch (talk) 21:14, 4 January 2015 (UTC)

There is a section named 'Use outside of smartphones and tablets' that sounds fitting for mentioning usage of Android as a PC. I doubt right click support is that noteworthy. ☃ Unicodesnowman (talk) 09:08, 19 February 2015 (UTC)

Semi-protected edit request on 20 February 2015

59.90.235.81 (talk) 10:44, 20 February 2015 (UTC)

 Not done No request specified. Joseph2302 (talk) 10:59, 20 February 2015 (UTC)

Android's installed base - the highest of all operating systems

Are we ok, with including A and B? I was trying to add this information but B is here in a rewritten, more clear form (hopefully). A is unchanged from the text that was reverted. The article doesn't say much about installed base, note distinct from "market share" (usually the information on a sales in a quarter/year). What it does say, was a conservative language "close to the installed base of all PCs" from the past when Android wasn't clearly bigger than PCs. I think the citations that where removed show that PC/Windows installed base can't be bigger.

A. By Statistica's estimate, Android smartphones have an installed base, in 2014, of 1600 million Androids smartphones (75% of the estimated 2.2 billion);

B. To put the Statistica's numbers in context: by Strategy Analytics estimates, Windows the most popular "desktop" operating system, has an estimated installed base of about 1300 million at best; they also estimate the overall tablet installed base to be already of comparable size to the PC market and predict tablets will have surpassed them by 2018. Most of the tablet installed base consists of Androids that add to the above Android smartphone numbers.

This looks better. Still a bit concerned about statistics overload since this will be the 10th paragraph in this section. How many paragraphs does it take to get across the point that Android has outsold a number of other platforms, and who is still reading at this point? But at least this time it makes sense. – Steel 18:22, 19 February 2015 (UTC)

Hello! I tend to agree with Steel – do we really need so much of the statistics? What would be our possibilities to make that section of the article a bit more compact, maybe by moving some of the content into {{Efn}} notes? If compacted that way, maybe this addition (which migth also use {{Efn}}) would fit better? — Dsimic (talk | contribs) 02:52, 20 February 2015 (UTC)

I completely agree, I was thinking that was my next step, if the above is accepted - really that Android is the most popular OS. This is saying something different, stronger than before, and then the old info doesn't matter as much. @Dsimic:, if you agree we the above then, why this edit? Maybe I let this sink in and wait with reverting but I'm not sure most people realize that Andoid is the most popular OS - it just happens to be a mobile one. Still, I would want the section to market share section to show the progression, 1. Android getting most popular in smartphones (really by now all mobile phones), 2. Android also most popular on tablets (beating iPad), 3. Most popular (while Android combined vs. Windows the perceived dominant OS, while its ok to mention that Windows is still dominant on PCs). comp.arch (talk) 09:10, 20 February 2015 (UTC)

Speaking of that edit, saying that "Android is the most widely used mobile operating system" is quite different from saying "Android is the most widely used operating system" – the key is what "used" should mean. "Used" as in the install base, "used" as in how many busy CPU hours it serves, or "used" as in how much data flows through it? Having "mobile" as part of the sentence actually improves clarity by restricting the domain of application. Hope it makes sense, and that's why I've re-added "mobile" there. Also, what should be the difference between "Android is the most widely used operating system" and "highest-selling operating system overall"? Should that be the overall user base vs. new installations in a selected time period? Maybe I'm nitpicking here, so please let me know if I should just let it go. :) — Dsimic (talk | contribs) 09:30, 20 February 2015 (UTC)

Maybe you (or I) are nitpicking. I believe I originally put in "most popular" - thinking is it appropriate. Popular (I'm told..) means the number of (computers) - installed base. Anyway, at least I have my phone always running (at has an alarm clock I use.. and of course is a phone). I in fact also have my desktop computer always running (my laptop could be in sleep/hibernate). Yes, Android will sleep but that means "connected standby" (Windows also starting to use it). Windows may very well be actively more used and web use may support that (62.67%). I'm not sure if it is my place to judge active or inactive use.. Maybe you want to suggest some wording? comp.arch (talk) 13:01, 20 February 2015 (UTC)

Hm, maybe we could simply say "As of MM YYYY, Android has the largest user base of all mobile and general-purpose operating systems" instead of separating that into potentially confusing "the most widely used mobile operating system" and "highest-selling operating system overall"? — Dsimic (talk | contribs) 01:40, 21 February 2015 (UTC)

Since we all agree on the length issue, can we have a rule that any edit that introduces more statistics must be accompanied by the removal of a comparable amount of less important or older statistics? That way comp.arch can keep the article updated, but without the section ballooning out of proportion more so than it has already. – Steel 20:08, 20 February 2015 (UTC)

That would make sense, though I'd suggest that we try not to remove old content entirely, but to "compress" it while using {{Efn}} notes where appropriate. Having a historical overview can only be a good thing, if you agree. — Dsimic (talk | contribs) 01:40, 21 February 2015 (UTC)

Semi-protected edit request on 2 March 2015

37.237.120.72 (talk) 19:44, 2 March 2015 (UTC)

Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format. NiciVampireHeart 20:30, 2 March 2015 (UTC)

Android fact: Bugdroid/Andy the Android wasn't Google's creation.

https://www.youtube.com/1gUQhN5joSQ?t=3m8s Fun fact: Andy the Android Mascot, or Bugdroid, was based on the Android that appeared on Gauntlet: The Third Encounter. You can play as Android on that game. It wasn't a original creation by Google. Rather, it was Tengen's creation at the time back in 1989, before Google used it as their mascot. CyanoTex (talk) 14:31, 7 March 2015 (UTC)

Hello! Hm, that YouTube link doesn't seem to be valid? — Dsimic (talk | contribs) 14:40, 13 March 2015 (UTC)

Update suggestion: Last version is now 5.1 Lollipop, not 5.0 Lollipop.

Under Development - Update Schedule HoschiCZ (talk) 18:03, 27 March 2015 (UTC)

Hello! Could you, please, explain a bit further to which part of the article you're referring? — Dsimic (talk | contribs) 18:24, 27 March 2015 (UTC)

Semi-protected edit request on 1 April 2015

Replace the following sentence with the updated sentence provided below: =

The officially supported integrated development environment (IDE) is Eclipse using the Android Development Tools (ADT) plugin.

Replacement/updated sentence

Initially, Google's supported integrated development environment (IDE) was Eclipse using the Android Development Tools (ADT) plugin; in December 2014 Google released Android Studio V1.0 as their primary IDE for native Android application development.

AndrewHulme (talk) 23:37, 1 April 2015 (UTC)

Done Stickee (talk) 00:14, 2 April 2015 (UTC)

Semi-protected edit request on 18 April 2015

Saro1991 (talk) 13:36, 18 April 2015 (UTC)

Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format. Edgars2007 (talk/contribs) 19:54, 18 April 2015 (UTC)

Added a new external link to popular Android APIs in GitHub

External links

• Official website
• Android Developers YouTube Channel
• Android Demo on YouTube, November 11, 2007
• Android: Building a Mobile Platform to Change the Industry: lecture given by Google Mobile Platforms Manager, Richard Miner at Stanford University (video archive)
• Android Internals: Fragment of a course detailing the architecture of Android and interaction of its components
• Android Popular APIs in GitHub

Andrehoraa (talk) 01:33, 2 May 2015 (UTC)

Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format. --I am k6ka ^{Talk to me!} _{See what I have done} 15:35, 2 May 2015 (UTC)

Separate section for desktop/laptop?

See edits, I put this in and now it is not a separate section. And a fact-template.

I think it is not in dispute that Android runs on laptops and desktops. There are I think at least an example of either/both (or was?) that was sold, in the article. That is not the issue however. Android is software. It only matters if Android is capable of running on these devices , and at least one reviewer says it is great to do..

I acknowledge that most do not do install (me included.. AOSP links here) Android themselves (ignorant?) or buy as a desktop or laptop. There are not that many. This page is also for device makers that also may be ignorant..

I note also that Chrome OS already "run" Android apps (while that OS is not "Android"). That is premature here.. but I see these OSes merging in the future.. Even been stated right?

Addition: I can't remember whether the hardware (review) was about the actual laptop or the actual desktop(and must have been removed from the article – as "promotional?) – shouldn't matter, as both considered the "same" platform for OS reasons. This text I quoted , was very intentional, as he must have used it on either or both – and it indicates a not-much known feature ("Alt-Tab") in Android (you never would on a smartphone) – about the software – that is Android:

"You can even Alt-Tab between running applications, something you don’t really notice until you use Android on a laptop or desktop." comp.arch (talk) 18:42, 17 May 2015 (UTC)

Hello! IMHO, a separate section on desktop/laptop use is justified, but not in its original form that consisted primarily of a few rather long citations. It should provide more descriptions and facts, and as such would fit very well. — Dsimic (talk | contribs) 05:22, 18 May 2015 (UTC)

I don't have an objection in principle to a section or subsection on desktops and laptops. I only removed the section heading because once the long quote had been removed there wasn't enough content to justify a whole section. Regardless a description/review of the user experience on desktops and laptops is clearly not 'Market share', so I've gone ahead and moved it back to 'Use outside of smartphones and tablets'. The long quote is still an issue. – Steel 17:12, 18 May 2015 (UTC)

Most popular "of any kind" - including RTOS and non-RTOS embedded OS?

Note, I did not put in "of any kind". I'm not really objecting, but I like the truth more than sources. But this one may be true..

I do not like to take the word of newspapers (even the New York Times) on technical matters: "Android is now not just the globe’s most popular smartphone operating system but the most popular operating system of any kind."

I have a source (a reliable survey) saying Android is most popular in embedded (at 16%), above FreeRTOS at 13%, but after "inhouse/custom" (at 24%).[29] Note, inhouse are many OSes - if that..

Now, embedded is not only RTOS (that is also included in (most) Android, while Linux there isn't, right?).

Android is not the most popular RTOS as it isn't - but it might be more popular than any single RTOS there is (or all combined?). Similarily, while not a server OS (can't be, w/o rooting), Android is probably more popular than all server OSes combined - at least any one. Or all desktop OSes.. comp.arch (talk) 11:21, 2 June 2015 (UTC)

Largest installed base?

The source does not state that Android has the largest installed base, but rather that it has sold more than Windows and iOS in the previous year. Please, update the source or the statement. — Preceding unsigned comment added by Pabloalmeidaff9 (talk • contribs) 02:39, 6 June 2015 (UTC)

[Please don't look at your (possibly) biased non-global or US/North American viewpoint. Yes, even iOS just got a slight edge there in sales in a quarter and web use (and Win 7 more popular than both, at least by web use).[30] Neither take the other extreme, Asia, Android most popular (yes, not one version..) at 34.7%, Win 7 (yes only one version..) 29.13%[31] OR the whole world, as web use is also biased showing Win 7 at 32.45%, then Android at 23.55%[32] and web use is not the same as installed base.]

More popular for years, not just one year. Is the installed base somehow unbelievably large(r) anymore? In the article: "Three billion Android smartphones are estimated to be sold by the end of 2014 (including previous years). According to Gartner research company, Android-based devices outsold all contenders, every year since 2012.[228] In 2013, it outsold Windows 2.8:1 or by 573 million.[229][230][231] As of 2015, Android has the largest installed base of all operating systems;[11] as of 2013, devices running it also sell more than Windows, iOS and Mac OS X devices combined[citation needed]." This had a citation (for combined, but 228 is also a good one) when I put it in.., somehow got lost in reorganizations..

Source 11, The New York Times: "Android is now not just the globe’s most popular smartphone operating system but the most popular operating system of any kind. More than a billion Android devices were sold in 2014, a c cording to the research firm Gartner. That’s about five times the number of Apple iOS devices sold, and about three times the number of Windows machines sold." WP article says only "Android shipped over four times more than iOS and OS X combined".

The "Three billion" estimated is of course, lower than all PCs ever made. Those, do not however constitute "installed base", most (only in 2007 two billion PCs had been made in decades, who keeps those..?[33]) of them are in landfills (some but not most of Android are also..). Getting clear estimates of installed base is difficult (and people apply different metrics for when old ones are thrown out (justifiably, at least in the past) for Android vs. iOS (vs. PCs)..), at least a *one* source for PC vs. smartphone compared (but see next ref). I went over this previously (should be in talk-archives). I think the onus in now in on those thinking PCs more popular giving a source for close to three billion (PCs estimated 1.5 billion in December 2013[34] (look at the graph there - projected 2 billion in 2014). Even Microsoft says they have a 20% market share in a WP:PRIMARY-source (from memory, echoing Goldman Sachs below). Android is more popular in embedded than FreeRTOS (the next-most popular (non-proprietary/non-commercial), by a survey I've provided and also more popular than proprietary VxWorks who say they are most popular commercial RTOS ("Deployed in over 1.5 billion devices"[35]).

What years (since 2012) of high market share (Android) and low market share (Windows) gives you is high installed base (Android): "The Goldman Sachs report considers traditional computers, tablets, and smartphones as all being computers, and takes a look at the market share of Microsoft, Google, and Apple. Back in Microsoft's glory days, in 2000, the report says, 97% of all computers ran a Microsoft operating system. By the end of 2012, the report says, Microsoft will have a 20% market share, trailing Google, whose Android operating operating system will have a 42% market share [..]

Goldman Sachs doesn't see Microsoft becoming the king of computers again for the foreseeable future. It projects that by 2016, Microsoft will have a 26% market share"[36] comp.arch (talk) 13:42, 8 June 2015 (UTC)

Platform usage pie chart colors

I realize it is cute in some way to use the Android green for the colors in this chart, but it makes it very, very hard to read. Why not just use a series of readily distinguished colors? I'd be bold and do so, but I am concerned that there may have been a consensus formed long ago that agreed to make it like this. I must say, it took me minutes to process information that should have taken seconds as I tried to compare shades and distinguish them. Huw Powell (talk) 04:03, 9 April 2015 (UTC)

Just change it, I don't think there has been any such consensus. The colours are changed back and forth semi-regularly between those who prefer using Google's colours for some reason (this part of the article is copied almost exactly from android.com) and those who find Google's colours unhelpful. – Steel 16:57, 9 April 2015 (UTC)

Hello! Just as a suggestion, we might want to use colors from the Android version history article, for some cross-article consistency. — Dsimic (talk | contribs) 00:58, 10 April 2015 (UTC)

I'm in favor of matching the colors from the official Android dashboard (https://developer.android.com/about/dashboards/index.html). I agree it's difficult to read but if you want consistency, it's better to be consistent with Android itself. SSherris (talk) 13:02, 5 May 2015 (UTC)

"Consistency with Android" is of no arguable value to the encyclopedia reader; ease of use is. The graph at the dashboard has each pie slice labelled, so it works fine there. Here on WP we strive to use no text in charts and let the caption match the user language, which makes the translation fail. But anyway, someone changed it to a series of contrasting colors (for now?) that appear to be the same as the version history article linked above, and I really appreciate that! Huw Powell (talk) 20:30, 12 August 2015 (UTC)

I agree, contrasting colors are much more usable. — Dsimic (talk | contribs) 06:19, 13 August 2015 (UTC)

1.0 Eclair

Please mention about the first version which is named after chocolate flavor - Eclair — Preceding unsigned comment added by 125.16.12.194 (talk) 12:48, 19 August 2015 (UTC)

Proposed merge with Android Gingerbread

See also: Talk:Android version history § Proposed merge with Android Gingerbread, and Talk:Android Gingerbread § Merger

This is a short enough paragraph to just be included under "versions". Jerod Lycett (talk) 08:40, 29 August 2015 (UTC)

• No merger: Nah. The solution here is to expand that page instead. There's nothing stopping it from being a fully fledged article like Android Lollipop, Gingerbread is just an older release so hasn't received as much attention (yet). The different releases were given their own articles precisely because there's too much to say about each release than can be covered under a subheading in some other article. – Steel 12:21, 29 August 2015 (UTC)
• No merger: Nope, we should expand the Android Gingerbread article further, as Steel already explained. — Dsimic (talk | contribs) 19:37, 29 August 2015 (UTC)
• I think not, this is still tracked, by Google (usage statistics), but seems to be EOLed by Google. Including recent versions.. Maybe they reconsidred with new security issues.. Anyway, Just as with Windows that has a family article, this one is that and needs to be mostly the current state, and about recent (non-EOLed) versions and brief history about older. The article is already 168,651 bytes (over the preferred 100,000 byte limit). If anything, if possible, move stuff from this article to version articles and/or the history article. A question where to draw the line, as EOLed status may not be clear. comp.arch (talk) 20:19, 29 August 2015 (UTC)
• Merge - If they're both about OSs, why not save space and include it in the main article? Socialistguy (talk) 20:25, 29 August 2015 (UTC)

  I assume you mean to here, but would you consider to the history article, like the other redirects? See at the Gingerbread page, edits I just did. comp.arch (talk) 20:37, 29 August 2015 (UTC)

  If the Android Gingerbread article is going to remain a brief list of Gingerbread features, merging it into Android version history would make sense; otherwise, the solution is to have a separate article. However, merging it into Android (operating system) wouldn't make much sense. — Dsimic (talk | contribs) 20:43, 29 August 2015 (UTC)

• My apologies if I pointed it at the wrong article. Just doesn't seem like quite enough to be worth an entire article, and I thought I had seen a "versions" section here, must have been Android version history I was thinking of. Should I change the merger to there, or is the idea of a merger a failure? Jerod Lycett (talk) 22:16, 29 August 2015 (UTC)

  No worries. If the Android Gingerbread article is going to grow further (which requires time, of course), then the merger wouldn't make sense. If it isn't going to grow, merging it into Android version history would make sense. — Dsimic (talk | contribs) 22:20, 29 August 2015 (UTC)

Android is not the most popular tablet platform - yet..

see edit summary: "Could we, please, have a reference for the "Android is the most popular tablet platform" part?" Thanks, for editing the "Adoption on tablets"-section, subsequently, then I assume you've seen I didn't say "most popular", only in many countries. The citations to statistics I put there are the only ones I have, that I summarized in the lead. Some more:

African continent 60.23%[37](majority since December, something I missed.., Nigeria 57.83%[38], Kenya 54.75%[39])

Poland 77.57% (one of many European countries with Android more popular than iPad)[40]

Two of the biggest countries (in addition to Nigeria) in the world: India, 65.73%[41], Indonesia 65.9%[42]. Asia is still only 45.93% (but many over 50%)[43] comp.arch (talk) 21:14, 31 July 2015 (UTC)

Hello! You're right, it's Ok not to have a specific reference in the lead section because, as a summary, it's covered by the references provided later in the Android (operating system) § Adoption on tablets section. — Dsimic (talk | contribs) 08:33, 1 August 2015 (UTC)

Not that it matters much, but in South America, Android briefly gained tablet majority in June and July, and then lost in August, and clearly in September (that in not yet over[44]). That made two continents (or three with Antartica.. but tablet use there is erratic.. iOS and Android switch places at 100%.. :) with Africa where Android is most the most popular operating system. Europe is also divided, while as a whole iPad majority, most Eastern European countries use Android.[45] comp.arch (talk) 12:03, 15 September 2015 (UTC)