Minimum API level for apps (dev discussion) - Android Software Development

Hello
Asking all developers to start a discussion on what the minimum API level is that you support, and have you decided to drop support for older APIs and if so what the reasons are. The reason for starting this discussion comes from the frustrations caused by having to support older API levels. Yes it's important to support as many users as possible - not everybody has access to the latest and greatest devices / Android versions, my own day to day device is a Samsung Galaxy S6 with Android 6.0.1. Although I have access to a device with Android 7.1 and one with Lollipop. However there is only so far back you can go in terms of support before development starts to get frustrating and in the end the question has to be asked if it is really worth it.
Please first take a look at https://developer.android.com/about/dashboards/index.html for the latest device stats.
For some background, I'm the developer of this app http://forum.xda-developers.com/android/apps-games/quick-gallery-photo-editor-t3394976 and https://play.google.com/store/apps/details?id=com.mobato.gallery. Currently min API 16 is supported, but I am thinking of making this 19 (KitKat), or dropping JellyBean support all together and just supporting Lollipop onwards. I will aim to explain why.
Recently I have added SD Card support. Anybody who has implemented proper SD Card support in their apps knows how frustrating this can be, especially with the new permissions required from the user in Lollipop onwards as part of the storage access framework (this is so bad that I even made a tutorial with screenshots showing the user how to grant access the SD Card, I have seen real users get confused by the SAF permissions activity). On KitKat it is impossible to write to the SD Card unless the device is rooted. This resulted in lots of negative reviews for apps when this limitation was introduced. Prior to KitKat all that was required was the write access permission to be declared in the application manifest. As you can see there are three different scenarios to write code for and test (pre-KitKat, KitKat and Lollipop onwards) which adds to development time/costs and bug fixing can become extremely cumbersome having to do full regression tests. However KitKat is still very popular. Since I don't even own a KitKat device it's impossible to test this scenario, let alone a rooted version (emulators only go so far).
Here are some additional facts:
KitKat devices account for 24% of checkins to Google Play
The 24% of KitKat users combined with older API levels (down to API 2.2) account for a whopping 39.3% of users checking in to Google Play.
Memory management for bitmaps just isn't great prior to Lollipop (lots of OOM can occur even with careful memory management as I have seen from crash reports).
JellyBean 4.4.4 got released in October 31, 2013. That's just over 3 years ago. In the world of mobile, that's equivalent to an entire period of geology such as the Jurassic Period.
JellyBean is discontinued according to https://en.wikipedia.org/wiki/Android_version_history
Fact: most of the bad reviews for my app are coming from users with low-end devices running older versions of Android, Lollipop onwards users generally leave more favourable reviews.
So in conclusion I am thinking about dropping support for KitKat and older at some point unless anybody believes it is worth supporting still? The question is how likely are those users ever going to make in-app purchases, although from a perspective of just showing adverts it might be worth it.
Please share your experiences with your apps and contribute your ideas.

g313 said:
Hello
Asking all developers to start a discussion on what the minimum API level is that you support, and have you decided to drop support for older APIs and if so what the reasons are. The reason for starting this discussion comes from the frustrations caused by having to support older API levels. Yes it's important to support as many users as possible - not everybody has access to the latest and greatest devices / Android versions, my own day to day device is a Samsung Galaxy S6 with Android 6.0.1. Although I have access to a device with Android 7.1 and one with Lollipop. However there is only so far back you can go in terms of support before development starts to get frustrating and in the end the question has to be asked if it is really worth it.
Please first take a look at https://developer.android.com/about/dashboards/index.html for the latest device stats.
For some background, I'm the developer of this app http://forum.xda-developers.com/android/apps-games/quick-gallery-photo-editor-t3394976 and https://play.google.com/store/apps/details?id=com.mobato.gallery. Currently min API 16 is supported, but I am thinking of making this 19 (KitKat), or dropping JellyBean support all together and just supporting Lollipop onwards. I will aim to explain why.
Recently I have added SD Card support. Anybody who has implemented proper SD Card support in their apps knows how frustrating this can be, especially with the new permissions required from the user in Lollipop onwards as part of the storage access framework (this is so bad that I even made a tutorial with screenshots showing the user how to grant access the SD Card, I have seen real users get confused by the SAF permissions activity). On KitKat it is impossible to write to the SD Card unless the device is rooted. This resulted in lots of negative reviews for apps when this limitation was introduced. Prior to KitKat all that was required was the write access permission to be declared in the application manifest. As you can see there are three different scenarios to write code for and test (pre-KitKat, KitKat and Lollipop onwards) which adds to development time/costs and bug fixing can become extremely cumbersome having to do full regression tests. However KitKat is still very popular. Since I don't even own a KitKat device it's impossible to test this scenario, let alone a rooted version (emulators only go so far).
Here are some additional facts:
KitKat devices account for 24% of checkins to Google Play
The 24% of KitKat users combined with older API levels (down to API 2.2) account for a whopping 39.3% of users checking in to Google Play.
Memory management for bitmaps just isn't great prior to Lollipop (lots of OOM can occur even with careful memory management as I have seen from crash reports).
JellyBean 4.4.4 got released in October 31, 2013. That's just over 3 years ago. In the world of mobile, that's equivalent to an entire period of geology such as the Jurassic Period.
JellyBean is discontinued according to https://en.wikipedia.org/wiki/Android_version_history
Fact: most of the bad reviews for my app are coming from users with low-end devices running older versions of Android, Lollipop onwards users generally leave more favourable reviews.
So in conclusion I am thinking about dropping support for KitKat and older at some point unless anybody believes it is worth supporting still? The question is how likely are those users ever going to make in-app purchases, although from a perspective of just showing adverts it might be worth it.
Please share your experiences with your apps and contribute your ideas.
Click to expand...
Click to collapse
You're right about bitmap memory management (I had an infinite System UI FC on CM 11 when I got too many notifications (from YouTube, by the way)
Sent from my GT-S7580 using Tapatalk

i want to run marshmallow apps on jellybean 4.1.2
g313 said:
Hello
Asking all developers to start a discussion on what the minimum API level is that you support, and have you decided to drop support for older APIs and if so what the reasons are. The reason for starting this discussion comes from the frustrations caused by having to support older API levels. Yes it's important to support as many users as possible - not everybody has access to the latest and greatest devices / Android versions, my own day to day device is a Samsung Galaxy S6 with Android 6.0.1. Although I have access to a device with Android 7.1 and one with Lollipop. However there is only so far back you can go in terms of support before development starts to get frustrating and in the end the question has to be asked if it is really worth it.
Please first take a look at https://developer.android.com/about/dashboards/index.html for the latest device stats.
For some background, I'm the developer of this app http://forum.xda-developers.com/android/apps-games/quick-gallery-photo-editor-t3394976 and https://play.google.com/store/apps/details?id=com.mobato.gallery. Currently min API 16 is supported, but I am thinking of making this 19 (KitKat), or dropping JellyBean support all together and just supporting Lollipop onwards. I will aim to explain why.
Recently I have added SD Card support. Anybody who has implemented proper SD Card support in their apps knows how frustrating this can be, especially with the new permissions required from the user in Lollipop onwards as part of the storage access framework (this is so bad that I even made a tutorial with screenshots showing the user how to grant access the SD Card, I have seen real users get confused by the SAF permissions activity). On KitKat it is impossible to write to the SD Card unless the device is rooted. This resulted in lots of negative reviews for apps when this limitation was introduced. Prior to KitKat all that was required was the write access permission to be declared in the application manifest. As you can see there are three different scenarios to write code for and test (pre-KitKat, KitKat and Lollipop onwards) which adds to development time/costs and bug fixing can become extremely cumbersome having to do full regression tests. However KitKat is still very popular. Since I don't even own a KitKat device it's impossible to test this scenario, let alone a rooted version (emulators only go so far).
Here are some additional facts:
KitKat devices account for 24% of checkins to Google Play
The 24% of KitKat users combined with older API levels (down to API 2.2) account for a whopping 39.3% of users checking in to Google Play.
Memory management for bitmaps just isn't great prior to Lollipop (lots of OOM can occur even with careful memory management as I have seen from crash reports).
JellyBean 4.4.4 got released in October 31, 2013. That's just over 3 years ago. In the world of mobile, that's equivalent to an entire period of geology such as the Jurassic Period.
JellyBean is discontinued according to https://en.wikipedia.org/wiki/Android_version_history
Fact: most of the bad reviews for my app are coming from users with low-end devices running older versions of Android, Lollipop onwards users generally leave more favourable reviews.
So in conclusion I am thinking about dropping support for KitKat and older at some point unless anybody believes it is worth supporting still? The question is how likely are those users ever going to make in-app purchases, although from a perspective of just showing adverts it might be worth it.
Please share your experiences with your apps and contribute your ideas.
Click to expand...
Click to collapse
i want to run marshmallow apps on jellybean 4.1.2 by decompiling particular app and editing its manifest.xml then change the minsdkversion to 16. is it possible? i tried it by decompling the file but i havent found any minsdkversion in manifest file after decompiling the apk.i downloaded an app called solid explorer and there is an option in it to open the apk file but that apk is not decompiled.so i took compiled apk ,i opened compiled apk file and then to manifest file after opening it i can see the minsdk version and other platform versions remember that was compiled apk file.
please tell me how to do this.

Related

android 2.2 on the way according to some intell

not my artical just found it orig web page
androidandmeDOTcom/2010/04/news/android-2-2-is-coming-what-features-do-you-want/
replace dot = .
Android 2.2 is coming. What features do you want?
By Taylor Wimberly on Apr 20267
The majority of Android phones are still running an outdated firmware, but that hasn’t slowed Google from advancing their mobile operating system. Everyone knows the next version of Android (codenamed Froyo) is on the way and the rumors (1, 2) are beginning to pick up that the Nexus One will receive it soon.
Based on the data from our analytics reports it appears Google has already begun testing on their next firmware – Android 2.2. Google engineers have routinely given codenames to future builds of Android (Eclair, Froyo, Gingerbread, etc.), but they don’t receive a point release till they are finalized and nearing distribution.
We have shared this reporting data before and people are always quick to point out that it can be easily faked, but I have spoke with additional sources familiar with the matter who confirmed Android 2.2 is currently being tested.
The feature set for Android 2.2 was frozen long ago, but it remains mostly unknown. Google decided last year that they would no longer publish public roadmaps for future versions of Android. Based on the information we can gather, it appears Android 2.2 will mainly focus on performance enhancements.
Some of the rumored features for Android 2.2 include:
•JIT compiler
•Free additional RAM
•OpenGL ES 2.0 enhancements
•Flash 10.1 support
•Fixed problem with “crazy screen” / Resolution of cross multitouch
•Activation of Color Trackball
•Enable FM radio
Some of these features are plausible and others we have no clue.
The JIT compiler has already been confirmed by Google and we expect to find out new details during Google I/O. Android engineers will be leading a discussion that informs devs how to test and tune their apps to work with the new compiler.
Additional free RAM should come thanks to the new Linux kernel being used in Froyo. Google is moving from 2.6.29 to the newer 2.6.32 which makes it possible to address the extra RAM. This has already been done in custom ROMs like CyanogenMod so expect this feature for sure.
I’m no expert on OpenGL, but it has long been rumored that Froyo (or a future build) will add new APIs to the Android NDK which will let Java devs have full access to the OpenGL ES 2.0 library.
Flash 10.1 is coming in the first half of the year. It is entirely possible it could be tied to a future Android firmware and be included with an OTA update.
We have already addressed the HTC touch sensor issues. It sounds like a future software update could be used to clean this up a little.
Multicolor trackball notifications are definitely coming. Android hackers like ChainsDD have already unlocked this feature for users of custom ROMs. Google advertised the multicolor trackball during its launch of the Nexus One, but the feature was cut from the final Android 2.1 build.
When it comes to the FM radio, I have no idea. Other HTC phones on a similar Snapdragon platform (Desire and Incredible) include the FM radio so it’s not out of the question that the N1 has it too.
I spoke with Google’s Eric Tseng during CES and he told me there were many secrets left in the Nexus One that we would discover later.
When is it Android 2.2 coming out?
Circle May 19th on your calendar. This is the opening day of Google I/O and I’m anticipating the release of Android 2.2 (and maybe the Flash 10.1 beta) will be tied to this event. Google has provided every developer attending the event a free Droid or Nexus One, so it is possible to see a simultaneous release on both devices.
I’m really curious how first generation devices are going to play with this release. We are expecting most phones to receive Android 2.1 this month, so it will be interesting to see how Froyo fits in. Some of the Android engineers have been talking about Froyo on the Google boards and it is unclear if the HTC Dream and Magic will be able to support it since they only have 192 MB of RAM.
p.s. if in wrong spot please move it im a tootal newb

Marshmallow Defect Corrections Release?

It looks like Marshmallow is following the usual pattern of Android "x.0" alpha release to the public, followed by "x.0.1" beta release with initial defect ('bug") corrections starting with Nexus beta testers (I.e. Nexus users in general).
Reading about the MM 6.0 problems on MXPE, I'm sitting out the 6.0 alpha testing on the sideline with LP 5.1.1. Most trouble-free phone I've had yet, and I don't yet need the only compelling feature I see with MM on the MXPE (T-Mobile Band 12 support).
Any noises yet about MM beyond 6.0.1? (I know I can look for this elsewhere too, but thinking maybe some of the XDA community may have inside info from the Android community.)
TIA...
The marshmallow update give me some new features and better battery life (though I do own the X Style, not pure). Unless you are dead set on being intentionally obtuse, then this isn't considered an alpha update.
Also the 6.0.1 update is quite minor, the largest change being some ART performance improvements, the rest is adding bands to the Nexus line and some emoji's: http://www.androidheadlines.com/2015/12/google-posts-android-6-0-1-changelog.html
I know the label "alpha" is not not the official label for something like 6.0. But with so many substantial defects, and multiple forthcoming revisions to correct those defects a certainty, that's really what it is IMO. Maybe "public release alpha" would be a better description, since pre-release revisions go through even more defect-ridden levels including pre-release alpha, prior to public release.
Similar situation with previous Android versions, and in fact most software foisted on the public these days (I'm looking at you, Microsoft and Apple). Look at Lollipop and the multiple public release revisions it took to iron out most of the substantial defects, finally, with 5.1.1.
6.0.1 is not just "...some ART performance improvements, the rest is adding bands to the Nexus line and some emoji's...", it also includes defect corrections. (Bluetooth, anyone?) And if the changelog doesn't list a significant number of defect corrections, that doesn't necessarily mean it is already polished at 6.0.1. The fragmented Android ecosystem and separation between Google, phone manufacturers, carriers, and users guarantees a plethora of various non-trivial defects in the ecosystem, many of which Google will address only slowly or even never for most phones.
For example, the memory leak defect in LP was not fixed until 5.1.1. How may revisions and months did that take? How many phones still run pre-5.5.1 with this defect?
One reason I bought the MXPE was the idea that it would be one of the first to get the updates. That turned out to be overly optimistic. It looks like Nexus is the only one still close enough to the source to get timely updates, and it also looks like Google is not pursuing Android defect corrections with any kind of urgency at all nowadays, maybe because the hardware ecosystem is becoming way too diverse to adequately support any more (or maybe because the profits roll in no matter what). Motorola phones, with the Moto alterations to Android, outsider status with carriers, and now hollowed-out Motorola support, appear to be no closer to adequate Android support from Google than any other non-Nexus phone.
"Obtuse"? A "bug" is a euphemism for a defect. Let's stop being obtuse, and call it what it is.
Any other info also appreciated.
You're being obtuse by insisting that we're all public alpha testers.
You obviously have no idea about software development, nor about Android Open Source development. Not your fault, but running your mouth is.
You bemoan the memory leak fix took several revisions to fix. So, you think that Google dedicated the whole team to fixing that one bug? What then? No other bugfixes or features are introduced in the meantime? The likely case is (and this is from experience) that bug took some revisions to fix, in the meantime, Google were also pushing ahead with other fixes. Regardless to what the uneducated (about SW development), throwing 15 developers onto one problem doesn't solve it any quicker. 5.0.1 came, adn 5.0.2 came, then 5.1 came in the meantime. While that memory leak was being worked on, more releases come fixing other things. Be grateful they didn't listen to you and leave it at 5.0 for several months while they fixed one issue.
Whatever bluetooth fixes that you think are in 6.0.1 are pure fantasy, because none exist in AOSP 6.0.1: http://aosp.changelog.to/android-6.0.0_r5-to-android-6.0.1_r1.html <-- That's the FULL changelog of commits between 6.0.0_r5 and 6.0.1_r1.
It is not Google's job to fix a problem in anything other than their own devices. At all. Google's job is to make AOSP run smoothly on Nexus devices and release the source. Samsung, HTC, LG, Motorola et al all take the source code, just like CM, AICP, Slim and the rest do, and make modifications for their devices, using the sources given to them by their hardware partners and themselves. So if BT works in Nexus devices, but not others, then it's not Google's problem (usually). An AOSP issue will persist several devices, including Nexii devices.
Google also have taken on the quite large undertaking of monthly security updates for their devices, which I can tell you will be taking up some of the development teams time (it's what, 3-4 months into that project?).
No software ever released on this planet comes without bugs and issues. This is software development. You can check the status of AOSP development here: https://android-review.googlesource.com/#/q/status:open and feel free to download, code and submit your own features.
MattBooth said:
You're being obtuse by insisting that we're all public alpha testers.
You obviously have no idea about software development, nor about Android Open Source development. Not your fault, but running your mouth is.
You bemoan the memory leak fix took several revisions to fix. So, you think that Google dedicated the whole team to fixing that one bug? What then? No other bugfixes or features are introduced in the meantime? The likely case is (and this is from experience) that bug took some revisions to fix, in the meantime, Google were also pushing ahead with other fixes. Regardless to what the uneducated (about SW development), throwing 15 developers onto one problem doesn't solve it any quicker. 5.0.1 came, adn 5.0.2 came, then 5.1 came in the meantime. While that memory leak was being worked on, more releases come fixing other things. Be grateful they didn't listen to you and leave it at 5.0 for several months while they fixed one issue.
Whatever bluetooth fixes that you think are in 6.0.1 are pure fantasy, because none exist in AOSP 6.0.1: http://aosp.changelog.to/android-6.0.0_r5-to-android-6.0.1_r1.html <-- That's the FULL changelog of commits between 6.0.0_r5 and 6.0.1_r1.
It is not Google's job to fix a problem in anything other than their own devices. At all. Google's job is to make AOSP run smoothly on Nexus devices and release the source. Samsung, HTC, LG, Motorola et al all take the source code, just like CM, AICP, Slim and the rest do, and make modifications for their devices, using the sources given to them by their hardware partners and themselves. So if BT works in Nexus devices, but not others, then it's not Google's problem (usually). An AOSP issue will persist several devices, including Nexii devices.
Google also have taken on the quite large undertaking of monthly security updates for their devices, which I can tell you will be taking up some of the development teams time (it's what, 3-4 months into that project?).
No software ever released on this planet comes without bugs and issues. This is software development. You can check the status of AOSP development here: https://android-review.googlesource.com/#/q/status:open and feel free to download, code and submit your own features.
Click to expand...
Click to collapse
t
No new or useful information there. Thanks anyway, despite the ad hominem. I guess that comes with the territory (forums).
Yep, the Google-Android-(independent hardware makers) ecosystem is seriously flawed. Too much disconnect between the OS owner (Google), the hardware makers, the carriers, and the customer. And the first three in the chain (not including the customer) have different incentives/disincentives, and there are a bazillion hardware variations, of course it is broken. We know all this.
Reminds me of the original PC/Windows mess. Except worse because the carriers interpose an additional dysfunctional layer hindering OS updates/support. (Before anyone says "just DIY with one of the many available ROMs, I started this "Q" thread about stock MM, not third party ROMs.)
Still hoping for any useful information on anything happening to fix the MM defects, to get an idea when it might be past public beta and worth installing to MXPE.
TIA...
Tinkerer_ said:
t
No new or useful information there. Thanks anyway, despite the ad hominem. I guess that comes with the territory (forums).
Yep, the Google-Android-(independent hardware makers) ecosystem is seriously flawed. Too much disconnect between the OS owner (Google), the hardware makers, the carriers, and the customer. And the first three in the chain (not including the customer) have different incentives/disincentives, and there are a bazillion hardware variations, of course it is broken. We know all this.
Reminds me of the original PC/Windows mess. Except worse because the carriers interpose an additional dysfunctional layer hindering OS updates/support. (Before anyone says "just DIY with one of the many available ROMs, I started this "Q" thread about stock MM, not third party ROMs.)
Still hoping for any useful information on anything happening to fix the MM defects, to get an idea when it might be past public beta and worth installing to MXPE.
TIA...
Click to expand...
Click to collapse
What ad hominem? Your uneducated state affects your ability to understand the nature of Android and software development. It's a perfectly legitimate response to your position. You lack the ability to understand and therefore your argument is flawed. I'm not attacking you, I actually tried to give you some insight into how it works, but you're not really interested and would rather insist on this "public beta" bull.
As far as fixing any "defects" you suppose, you haven't actually listed any so no-one is going to be able to help you with temporary work around without a list of what you feel is broken. I also showed you the changelog, so you can do your own homework to see if your supposed defects are fixed in 6.0.1.
The various hardware configurations doesn't even matter because Android is built to deal with it. So long as the hardware vendors of chips and modules support them properly and give out functioning binaries to OEM's, or proper source code, it's irrelevant. The exact opposite of what you said is true, Google has a very close relationship with it's partners (anyone signed up to their Google programs, to preinstall Google apps). The problem is carriers, who really shouldn't have a say in software on the phones, but that seems to be a chiefly North American problem.
Google doesn't need to have any connection to Android users as customers. Google does not sell Android, therefore you are not Google's customer unless you use a Nexus phone. Google sell the Google Experience, with the Nexus. You are Motorola's customer, and you are using Motorola's branched version of Android. Google doesn't owe Motorola any fixes or patches for their device. Motorola must maintain their own device tree and maintain their own relationships with their partners.
EDIT:
Also, Motorola's problem is resources. They have four version of the Moto X 2015 to deal with, three versions of the Moto X 2014, the new X Force, then the various versions of the G and E to deal with, along with two smart watches, and so forth. Their line up is increasing whist I imagine their development team is not. There was outrage (rightly so) when news broke that the Moto G 2015 wasn't getting the MM update, despite being a couple of months old, and Motorola listened and OTA's are rolling out.
I am asking if anyone can offer any info on anything being done to move toward MM revision with the many significant defects of 6.0 corrected. Read the forums, there are way too many defects with 6.0, it is patently a de facto public alpha, and we are tracking the usual pattern where it takes 3 to 5 revisions before an OS major rev is ironed out enough that upgrading will not cause more problems than it fixes.
There are always excuses made for why there are so many defects in software. There is a euphemism for "defect" everybody uses, "bug". Everyone has been making excuses for so long about shoddy workmanship and inadequate testing and correction of software, with the "bug" euphemism to minimize the reality that these are defects, that we are all just to suppose to accept systems ridden with faults without complaint. It's unacceptable. It can be done better. Part of why it doesn't get better is because everybody says "that's just the way it is, deal with it". With mountains of byzantine excuses and even ad hominem attacks (as here).
This thread was not started to start a tit for tat ad hominem back and forth, nor to post long essays detailing excuses for the pathetic status quo of the fragmented Android ecosystem with respect to defect causes and distributions. It was started looking for any info about work being done to fix the stock MM defects. Still seeking info.
TIA.
You should probably check the definition of ad hominem. There was no attack on you as a person, just pointing out that your uneducated state with regards to knowledge about software development affects your ability to call judgement on this.
But you haven't listened to a single word I've said and still maintain a shoddy position, so I would suggest to anyone else who reads this to simply ignore you as a troll.
Tinkerer_ said:
It looks like Marshmallow is following the usual pattern of Android "x.0" alpha release to the public, followed by "x.0.1" beta release with initial defect ('bug") corrections starting with Nexus beta testers (I.e. Nexus users in general).
Reading about the MM 6.0 problems on MXPE, I'm sitting out the 6.0 alpha testing on the sideline with LP 5.1.1. Most trouble-free phone I've had yet, and I don't yet need the only compelling feature I see with MM on the MXPE (T-Mobile Band 12 support).
Any noises yet about MM beyond 6.0.1? (I know I can look for this elsewhere too, but thinking maybe some of the XDA community may have inside info from the Android community.)
TIA...
Click to expand...
Click to collapse
I think we'll close this debate. There are no real "Android" insiders on XDA, so asking for update info which is privy to Google is perhaps somewhat futile.
On a related note, XDA have a few dedicated "Android Fora", such as this complete Category where non-device specific discussion and indeed conjecture takes place. Perhaps you could take a look there and see what transpires?
Thanks

Why Xperia Z3 won't get official Android N

I've done some easy searching and found what might be the actual cause of the Xperia Z3 not being announced as getting Android N with other devices from the Sony line.
Even though we have Android 7.0 previews available and it works fine with a few software bugs that need to be ironed out, the Xperia Z3 will not be getting official Android N.
And heres why:
The problem lies in Google and their Android N requirements.
You might be asking, what do you mean by requirements the device ran the previews just fine!?
Well, Google has introduced new encryption methods in Android N which have certain hardware requirements.
The Snapdragon 800/801 don't have support for some of these newer encryption methods.
Another question you might be asking is, well we don't need these encryption methods, but Google says otherwise.
Their policy is that their encryption methods must be fully running by the time the device setup is done.
So the actual reason for the Z3 not getting Android N is both technical and legal.
In the end, its actually not Sony's fault, Sony is just following the rules which Google put.
Although I don't see any problems for Qualcomm releasing the drivers needed for further development.
We will probably see unofficial Android N on our devices, but at the moment it seems like a far stretch due to Qualcomm not wanting to release drivers.
Sources:
Android Authority
It's exactly full-disk encryption and GG haves Compatibility Definition Document (CDD) for all devices need to run Android N. So easily, Snap 800/801 have feature full-disk encryption, but with this, they do pass Compatibility Test Suite (CTS) but don't meet the CDD because read/write speed. Snap 805 is newer and have module for this feature, so can make it faster.
But, if Sony open our kernel, Snap public your driver, custom rook does not need full-disk encryption at all.
That's all I found about this.
This is just one of many reasons why we can't get N. Opengl 3.1 or Vulkan is also required for certification which only SD 805 and later have. Basically all devices older than SD 805 are sol for future updates. Now Sony could release stockbased build in a zip format through dev program without gapps which wouldn't cause issues with Google
Lets try and keep it in one thread
http://forum.xda-developers.com/z3/general/xperia-z3-wont-android-n-t3446158
Thread closed.

will project treble successed

As seen here
https://www.reddit.com/r/LineageOS/...ect_trebel_devices/?utm_source=reddit-android
LineageOS team state that project treble is in its baby shoes and completely dependant on google to optimize it even more since as of now gsi rom requires certain adjustments for each device, so will project treble successed?
any1 has an insight please share.
I think it needs more adjustment. The kernel should be universal and updateable along with the OS, it's pretty universal as it is at the moment. Drivers should also be standard and updateable, at least for standard items. There should be a driver model where possible to support other devices, and any phone specific changes could be done through manufacturer supplied drivers. There's really no reason why it can't be done, it would be along the lines of how Windows works. Of course, it can be tailored to suit phones.
System updates should realistically come from Google, it would mean all phones and devices would be up to date with the latest security updates. The phone can also check with the manufacturer for specific updates. If Google keeps them apprised of any changes they can update their specific updates in time. This model would mean individual service testing for a new OS update etc wouldn't be a problem since it should at least be compliant with the base model.
Don't forget Google tracks down security leaks in other OS like Windows, which isn't even a direct competitor, and releases the security leak information if it isn't patched within 30? days. How many Android devices are updated with security patches within 30 days by the manufacturer? It's very much a double standard. Google really needs to think of an even more universal model like I just depicted for Android 10 Quinoa Slice (or whatever they call it).
Will Treble succeed? It's a step in the right direction, but needs more work. Not only should something like I just described be done, it should be made mandatory for all new devices. It doesn't mean custom versions are out, but custom versions would have to have OTA updates and be updated quickly along with the standard OS.
To say whether it succeeded or not, I think you'd first need to define what's its goal.
I still don't even know the answer as of today.
Some people say that the goal is to have a system image controlled and updated by Google.
But I don't see this happening any time soon. Google would need to test their GSIs on many devices, and they didn't even test P GSI over O-MR1 Pixels!
It seems to me that their goal was simply to make updates easier to OEMs. Considering Essential PH-1 getting Pie day-one, this might seem a success.
But we'll need to compare Oreo adoption rate to Pie's to confirm.
Oooh someone made a thread so i can moan cheers
The main problem with treble for me is that it's splintered between a plethora of devices, so one dev will release a treble rom and a multitude of device owners will flash it, each with their own subjective problems and issues, requests and wants.
And there lies the problem, it's difficult even when it's a dedicated rom thread for a particular device to get help at times.
So when you have a bunch of users talking about completely different devices you haven't got a hope in hell.
I think there should be branches to each thread for each specific device, that way help threads can be more linear rather than the chaos that it is at the moment.
Least that's my thoughts
My only rom i've flashed is RR and besides a few missing features, Fingerprint, Stereo and NFC i think it's brilliant.

Story of Our Older Devices - Usages & Purposes

Do you have an older Android device? If yes, do you still use it in any way? ?, Share your story here.
Android OS is dynamic (I'd say unstable ?). The device you are buying today might become older enough in a couple of months in terms of the Android version. However if you're lucky to have a truly rich and supporting community, your device will be alive for an eternity.
In the revolution of Android ecosystem, you would possibly like to upgrade your device after major Android updates. Or you have to be satisfied and hanker after the newer Android features otherwise.
Now it's time to share mine ?. I got a local, cheap Jellybean device, when I was introduced to Android in early 2014. Now I have still that Android 4.2 device which I love more than the prettier material design or even the notification dots and others.
I have been using that device for my development purpose for a long time alongside my primary device. As I used it in my first step to Android root, mods, hacking and the system insight. I still remember the day I got this device into bootloop for the first time ?. Those days have passed away.
Now I am learning and developing Android apps. Since my laptop is even smaller than its small configuration (? You can call me a poor guy), it is quite impossible to test and debug apps on Android emulator while running Android studio in parallel with 4GB of RAM and i5. So, I started using my old friend as a helping hand in development. I keep it connected via USB while developing my apps, run and test them with it which frees nearly 700MB of RAM on my tiny machine and boost the build time also. For other Android versions, I have to do them on emulator separately ?.
What you're thinking of me I don't care ?. The phone is much more to me. Feel free to share yours...
What is your phone name and model?
Have you installed a custom rom on it yet?
What are your apps which you have developed?

Categories

Resources