Database Info

My thoughts on why Google couldn't avoid OS version fragmentation

Guys,
I think everyone nowadays is quick to judge Android and Google for fragmenting the OS with multiple (4) versions but I think it was a necessary for success side effect. Without the humongous push that Google has done with Android, it wouldn't have been where it is now.
If you haven't heard, Google shared its plans to battle fragmentation which I mentioned as well.
What do you think? Do you think if Google slowed down back then (1.5/1.6), they would have ended up with Android that is as awesome as today?

Right now, Google has set up Android with a low-end version (1.5/1.6) and a high-end version (2.0, 2.1) and the market is sorting itself out. I think the two options were to fragment or be unsuccessful. With the fragmentation comes some issues, but it also comes with a wider range of devices that are capable of running it, which pushed it's popularity. Fragmentation can be easily solved farther down the line when Google takes the updating into it's own hands and stops letting carriers and manufacturers screw with everything.
In the end, we'll see Android itself being updated via the Market, I'd bet.

thats why google is releasing separate packages for the awesomeness theyre releasing from now on. the browser, ui kits, etc will come in separate apk's on the android market so you can have fun with whatever kernel and not be binded to the manufacturers limits. at least thats what the article today from engadget said.

Good idea for Google to redesign Android for better/easier OEM skin integration?

I couldn't fit the concept completely in the title, but here it is:
Would it be a good idea for Google to redesign Android in such a way that it allows for OEMs as HTC and Samsung to more easily add their layers Sense and TouchWiz upon it - perhaps even in the form of a single apk?
Considering the fact updates take extremely long because the newly released Android versions, in this case Jelly Bean, need to be altered by HTC and Samsung and others before Jelly Bean can be even released onto the brand devices, not to mention an even longer delay for those who decide to purchase their phone by the major source of destruction of innovation, that is, the carrier. If Key Lime Pie, the next Android version, is designed in such way that the OEM skins and Android itself remain seperate, would it be possible to update to a newer revision of Android without having to be dependent on the OEM and carrier? It would then be the OEMs responsibility to update their skin as soon as possible to maintain a proper experience, as obviously a new Android revision has new elements that could possibly remain unskinned yet at release.
I personally think a strategy in this way forces OEMs more to bring out updates as soon as possible, thereby driving competition up - which is ALWAYS a good thing, while Android updates remain independent of OEM and carrier.
Of course, there will be issues with such strategy, such as required updates to drivers. Skin updates are non-essential, while it would still be nice to at least have a functioning device after an update. How would that work out for devices by HTC automatically receiving updates to Android? Remember, the whole point of this idea is to bypass the need for OEMs and carriers altogether. While I even doubt thát is possible, I'd like to hear a discussion trying to achieve this point as closely as possible.
I even believe this thread, if it does result in great ideas, could be forwarded to Google to look at - but I might just be completely wrong with such idea and as such, I'd also like to hear why.
Discuss away.

They're already working on it.
[Note: This information is ~18 months old, probably. I haven't kept up to date on the projects.]
T-Mobile built a theme engine, which is included in CyanogenMod. However, there are aspects of it that Google does not quite like. Sony-Ericson is also putting forth an alternative approach, which Google likes better. So, T-Mobile is working to merge the best of both worlds.
The "big difference" between the approaches is:
- T-Mobile's themes are user selectable at run-time.
- SE's themes are set by OEM (need to flash a new ROM to change)... but they are more deeply/cleanly/??? integrated with the Android core.
It's both a technical and political matter.

Ars: Google’s iron grip on Android: Controlling open source by any means necessary

A few weeks ago, I posted a very unfortunate Google+ post of the creator of Focal and why it was removed from the CM codebase. It was a depressing story and it really started to make you wonder about where CM is going.
This time, after reading an extremely well-written article, I've come to a similarly depressing conclusion: Android by Google is slowly becoming as locked down as iOS, but not in the sense that you think; it's not about what apps let you do what, it's the developers.
We've finally arrived at a critical flaw with the way Android is developed and these days, I can no longer claim that Android (by Google) is "open" anymore.
Feel free to give this a read (Disclaimer: I am not affiliated with Ars Technica in any way).
http://arstechnica.com/gadgets/2013...ntrolling-open-source-by-any-means-necessary/
It's not just about Amazon's version of Android; CyanogenMod is for all intents and purposes a "fork" of Android. It is designed to work without Google Apps and as we all know, we flash those seperately. But that's the problem, the answer isn't just "Well, I'll just flash the Gapps and it will work like it should". What will happen if new Play Store apps start referring to features in the framework that don't exist in a form that we can flash? What if the license to flash the Gapps gets revoked?
How will CyanogenMod start adding features to apps that were originally AOSP but are now closed source? What will happen when the open source Messaging app is abandoned and turns into a Hangouts feature? How can CM stay on top of that?
It's not as simple as "take the source we currently have and work with it", because what will happen when Google adds a killer feature to an app that depends on some API that is no longer open source?
These are some rather frightening questions to deal with. I don't know where Android is going, but I'm certainly starting to wonder what's going to happen to it.
I'd appreciate any and all input on this.

Not very continuous, but here's my thoughts about the article:
The Gapps license is meant to lock the makers of Android phones into Google, so users get locked within Google and Google can gain revenue from the users. After going to that extent to make sure Google gets to keep the device's user, what's to gain if Google users of the device who flash CM to be locked out of the system instead of keeping them "trapped" with the Google ecosystem even with a non Google ROM? Doesn't make any sense does it?
I suppose we will still have to flash them like we flash the Play Store now. Unlike Amazon, CM (for now) actually still relies on Google and doesn't "divert" revenue to another company and therefore Google would be more than happy to let their apps be used. But if CM does start going the Amazon way, I believe Google may lock CM out.
Those APIs take time to develop, take the Maps API for example - you think they spent millions, if not billions mapping the entire world and even roaming every street just to make sure you can find your way around for free? They'll need to recoup their costs somehow.
While Android is open source and contributed by Google for free, don't forget Google is a company, not a charity. They have to make money or their shareholders won't be happy. Even if their shareholders are massive fans of open source they also have thousands of employees to pay, and all that costs money. And don't forget, when a company is providing free stuff for you to use, you are not their customer - you are their product. Android will change in ways that will keep Google profitable and keep competitiors unprofitable, while keeping the users as comfortable as possible so they will continue to be their product.

cccy said:
Not very continuous, but here's my thoughts about the article:
The Gapps license is meant to lock the makers of Android phones into Google, so users get locked within Google and Google can gain revenue from the users. After going to that extent to make sure Google gets to keep the device's user, what's to gain if Google users of the device who flash CM to be locked out of the system instead of keeping them "trapped" with the Google ecosystem even with a non Google ROM? Doesn't make any sense does it?
I suppose we will still have to flash them like we flash the Play Store now. Unlike Amazon, CM (for now) actually still relies on Google and doesn't "divert" revenue to another company and therefore Google would be more than happy to let their apps be used. But if CM does start going the Amazon way, I believe Google may lock CM out.
Those APIs take time to develop, take the Maps API for example - you think they spent millions, if not billions mapping the entire world and even roaming every street just to make sure you can find your way around for free? They'll need to recoup their costs somehow.
While Android is open source and contributed by Google for free, don't forget Google is a company, not a charity. They have to make money or their shareholders won't be happy. Even if their shareholders are massive fans of open source they also have thousands of employees to pay, and all that costs money. And don't forget, when a company is providing free stuff for you to use, you are not their customer - you are their product. Android will change in ways that will keep Google profitable and keep competitiors unprofitable, while keeping the users as comfortable as possible so they will continue to be their product.
Click to expand...
Click to collapse
First, I appreciate the input! I was looking forward to intelligent discussion and it's great that the first reply is just that.
I would like to clarify though; my concern is not so much about Google making money; they are a business and deserve to make money in whatever way they see fit. We have something they want (ad clicks and search history) and as long as they provide an experience worth using, I don't mind that transaction at all.
My worries start with what the custom development scene will look like one or two years from now if the base apps that make Android useful on its own (and by extension, useful to custom developers) have been molded into Google Play apps or frameworks or APIs.
In parallel, it's also starting to make sense why Cyanogen continues to put effort into alternate applications such as Apollo and Focal; they saw this coming way before we did.

LiquidSolstice said:
First, I appreciate the input! I was looking forward to intelligent discussion and it's great that the first reply is just that.
I would like to clarify though; my concern is not so much about Google making money; they are a business and deserve to make money in whatever way they see fit. We have something they want (ad clicks and search history) and as long as they provide an experience worth using, I don't mind that transaction at all.
My worries start with what the custom development scene will look like one or two years from now if the base apps that make Android useful on its own (and by extension, useful to custom developers) have been molded into Google Play apps or frameworks or APIs.
In parallel, it's also starting to make sense why Cyanogen continues to put effort into alternate applications such as Apollo and Focal; they saw this coming way before we did.
Click to expand...
Click to collapse
I believe the custom development scene wouldn't get affected much. After all, remember the old XDA-Developers? Windows was all locked down, but the cooks still managed to make customized ROMs. What's more, Google wouldn't want to lose their "products" - Google wants us to continue to use their services so they can earn money, they wouldn't lock us out.
What competitors lack is the capability to access Google's services (Frameworks, APIs, etc) as Google has ways to block them (Which is why we had circumvents like device spoofing). If you had a device designed for Google's version of Android, I am sure Google would still enable access if you use a custom ROM. The point of locking those competitors out is to force them to embrace Google's version of Android and not use their own forks which would keep Google out of certain aspects of the user's phone, decreasing revenue. Therefore, if you could roll your own custom ROM, it makes sense for Google to continue supporting you so you still completely rely on them instead of "outsourcing" to other competitors.
CM puts effort into alternate applications because as you can see right now, CM's starting to roll their own commercial forked devices - what happens after that? If you have seen the ways of other commercial versions of Android (Amazon, China brands, etc), they start replacing certain revenue generating aspects of the phone to use their own service instead of Google's. Certainly not what Google wants.
In short, I would say, if you are a small custom ROM user, Google isn't going to come after you, they want you to use their services! But if you are a competing company, expect your devices to be locked out from Google in the hopes that they eventually force you to bow to them and convert all your users completely to Google's "products".

Why isn't there custom opensource bootloaders like custom recoveries for android phones ?

This may be stupid, but I couldn't find any resources regarding this. We have custom recoveries for android devices but why isn't there custom bootloaders like there is for PCs ? Like in the PC space we have the likes of reFind and gnu grub.
Thanks

There are some instances of alternate bootloader projects. Just that they are not popular,
[Bootloader] LK for Xperia T
LK for Xperia T LT30p Only - Unlocked Bootloader Required WARNING 1: This modification makes changes to the devices partition table. I (lilstevie) am not responsible for any damage to your device or data loss that may occur. WARNING 2: ICS...
forum.xda-developers.com
EFIDroid
EFIDroid is a easy to use, powerful 2ndstage-bootloader based on EDKII(UEFI). It can be installed one-click with the EFIDroidManager app. You can add/remove/edit multiboot ROM's. There's no special support needed by ROM's or RecoveryTools(no...
forum.xda-developers.com
The developer of EFIdroid stopped developing in 2019.
efidroid on Android 9 and 10 devices ? · Issue #152 · efidroid/projectmanagement
Hi, I just want to know if efidroid supports devices with 6 GB RAM and 64/128 GB Storage devices running Android 9 and Android 10 ? thanks.
github.com

Not to mention you would need OEM's to cooperate....

Thanks @karandpr for that github comment a lot of info there. Thanks @galaxys too. So a quick summary would be that the reason is that for the bootloader to work smoothly there has to be support from the kernel too, which the OEMs should do and probably would not. But I didn't think about the support in the kernel was an issue. That does seem to be a lot of work and I see the reason now.

al_l_en said:
Thanks @karandpr for that github comment a lot of info there. Thanks @galaxys too. So a quick summary would be that the reason is that for the bootloader to work smoothly there has to be support from the kernel too, which the OEMs should do and probably would not. But I didn't think about the support in the kernel was an issue. That does seem to be a lot of work and I see the reason now.
Click to expand...
Click to collapse
I don't think Google intends to open up android anymore. They want restrictions like iOS but pretend to be open source for the "goodwill". What's the use of AOSP if you cant effectively install it on a device or your important apps don't work?
I believe PinePhones are the ones that can have truly open-source compatible hardware. The specs are underwhelming but the community is really good.
You can get spares easily and the battery is removable.
Only thing is they are mostly out of stock.

karandpr said:
I don't think Google intends to open up android anymore. They want restrictions like iOS but pretend to be open source for the "goodwill". What's the use of AOSP if you cant effectively install it on a device or your important apps don't work?
I believe PinePhones are the ones that can have truly open-source compatible hardware. The specs are underwhelming but the community is really good.
You can get spares easily and the battery is removable.
Only thing is they are mostly out of stock.
Click to expand...
Click to collapse
Yeah those are great but the problem is that they are not usable for "normies" which will prevent mass adoption and hence cannot have a sustainable business model.
But I think google is not the only one to blame, like couldn't the OEMs actually provide bootloaders that can boot signed os images. Or is there any technical or security difficuties in doing that.

al_l_en said:
Yeah those are great but the problem is that they are not usable for "normies" which will prevent mass adoption and hence cannot have a sustainable business model.
But I think google is not the only one to blame, like couldn't the OEMs actually provide bootloaders that can boot signed os images. Or is there any technical or security difficuties in doing that.
Click to expand...
Click to collapse
Normies are afraid to change the default browser, so bootloader is really out of their leagues.
Phone tinkering is a hobby, not a necessity. Phone tinkering itself is not a sustainable model.
Google is to blame primarily. Because they have a stringent list of requirements for devices to pass CTS. You can read the bootloader requirement and judge yourself.
Android 11 Compatibility Definition | Android Open Source Project
source.android.com
Without passing CTS, devices cannot use Google apps, they cannot get push notifications and they cannot pass SafetyNet checks used by most banking apps.
At the end of the day do I want to spend 100s of hours to bring a feature to an android phone which will probably be used by 10 users and deprecated by the time I finish doing it?
or do I want to buy a phone which will allow me to tinker freely in a community and ecosystem which allows modification?
For our tinkering pleasures, Pinephone is the way to go for now. They have support from Manjaro, Debian and KDE. Which is a big thing IMO.
Or else there you can roll your thing in RaspberryPi?

While going through related details I found an article about google probably switching to hardware based safetynet checks which could be ending google play compatibility on custom roms.
It really seems like google is using security as an excuse to make sure that there are no competitors in their business space.
Maybe this is because I have been only doing web development and only started learning app dev, but the reasons google use for CTS like for enforcing DRM, is also handled on websites while allowing openness and being neutral (or maybe the web is not as secure as something like this, so forgive me if I am wrong). Android could really take pages off the web ecosystem for being a neutral platform.
I really appreciate the patience for hearing out and also the references(and the rabbit holes that it was followed by) really taught me a lot about general android architecture.

al_l_en said:
While going through related details I found an article about google probably switching to hardware based safetynet checks which could be ending google play compatibility on custom roms.
It really seems like google is using security as an excuse to make sure that there are no competitors in their business space.
Maybe this is because I have been only doing web development and only started learning app dev, but the reasons google use for CTS like for enforcing DRM, is also handled on websites while allowing openness and being neutral (or maybe the web is not as secure as something like this, so forgive me if I am wrong). Android could really take pages off the web ecosystem for being a neutral platform.
I really appreciate the patience for hearing out and also the references(and the rabbit holes that it was followed by) really taught me a lot about general android architecture.
Click to expand...
Click to collapse
Theoretically, Google can end GPlay compatibility on Custom ROMs anytime they wish. It's just that lot of App Developers don't use SafetyNet the way it is intended and Google doesn't roll out its strict check. They do it once in a while.
They don't have any competitors in their business space. It's a very well-thought monopoly.
CTS restricts Google Play API access to vendor operating systems. So vendors like Samsung, OnePlus and others have to play by their rules. IIRC, the cost of Play API is around 15$ per device but it is subsidized for large quantities.
End users don't really care about Play API. But App Developers do.
Without Play services, there is no easy way to integrate push notifications, ads, maps, analytics, metrics, and so on. Rolling your own thing will take years to develop and won't work as seamlessly as the play service counterparts.
I don't think Google will ever cede their monetary interests for open collaboration.

karandpr said:
I don't think Google will ever cede their monetary interests for open collaboration.
Click to expand...
Click to collapse
Yeah that's for sure. The only way this monopoly can break is when an opensource alternative to google play services and other apis exist and while doing that it must be compatible with the existing google apis. And that is probably not going to happen in a long time. Although microg does solve this to some extent, but still it is a second citizen.
Some of the functionality is already there, like most of the google apps like docs and drive could replaced by nextcloud and then maps could be replaced by osmand. If some company, preferably an OEM, comes and integrates all of these into a package maybe there's hope. I think /e/ os tries to do this to some extent.

You might find this resource useful. As they have gone over a comprehensive set of bootloader software and tried to outline their primary features in detail. Hopefully, you’ll be able to determine the best one for your use case. https://www.ubuntupit.com/best-linux-bootloader-for-home-and-embedded-systems/

The future of android

Hi, as some will obviously know, google is forcing a change in android development to be more like ios. Some developers and users wont even notice or care. Others may find the changes fundamental and devastating.
Some of the changes have come about in version 11 but will be fully implemented by 12-13. These changes are going to limit access to android file system. The way apps work and limit what you can install, copy, write to external usb etc. Others will mean total lockdown of security from installing apps and google spyware controlling what you can change.
Over the years we have all seen many versions of android in countless devices with as many custom iterations and mods. In a way. Us users and the developers have shown what's possible with imagination skill & ingenuity and happily let google lead us down the garden path making billions in revenue from our devoted support. Not everyone could see the control manipulation, development and exploitation. Not everyone even cared.
But it seems now that free reign google has given us in ability for hacks & mods and and the devices android can can be used on is coming to an end. Google is yanking its chain and reeling us in.
If you think scoped storage, or more play store control will just be an inconvenience think again.
Developers and genuine android experts will know this and will probably already be aware of some solutions. I hope so. As the thread count and discussions on this balloon maybe some will consider a fork of huawei's stripped down versions of android might be an option, however we feel about china. Let's hope some options will come to light soon.