Google defrags Android by removing OEMs and Carriers from the loop

[Crono]

Your point about brand loyalty is sound, but it does indeed demonstrate the fragmentation issue. As Red Storm pointed out, it requires Android 4.3, which is not yet available on these devices. Why is the latest version of Android not yet available on Samsung's current and previous generation flagship smartphone models? Because it takes time and manpower to make a new OS work for each and every device with each and every combination of SoC, screen, camera, RAM sizes and speeds, WiFi/broadband modems, and God knows how many other chips and controllers inside.

Improving software to get better battery life is NOT a fragmentation problem, it's something required if you want to have more and more devices (like bluetooth-linked watches) connected to your smartphone. The real problem is this, which you quoted:

but it doesn’t look like Samsung’s going to support other phones, even those that run Android.

No one complains when new features are added to any other new traditional operating system or update, but for some reason on Android it's a major problem. I get that the pace of updates is much faster in the mobile world, but people need to get used to that.

 

[ControlD]

I think it's pretty telling that the most popular Android phone maker, with the most popular Android phones on the planet, cannot do a little better. Really, Samsung should be able to do day-and-date updates with the official Android releases. It's not like Android is a closed development project that is sprung on the world at random times with no advance notice.

It isn't quite that easy though. Unlike Apple, Samsung has to get each release approved and bloated with each carrier's flavor of that new release. Apple did it the right way in this regard, they don't allow the carriers any say in the process. Samsung may very well have Android 4.3 ready to roll for every device under the sun but you won't see it until the carriers feel like making it available (assuming you don't have a Nexus device).

 

[gorcorps]

It isn't quite that easy though. Unlike Apple, Samsung has to get each release approved and bloated with each carrier's flavor of that new release. Apple did it the right way in this regard, they don't allow the carriers any say in the process. Samsung may very well have Android 4.3 ready to roll for every device under the sun but you won't see it until the carriers feel like making it available (assuming you don't have a Nexus device).

But why can't it be that easy? Why can't Samsung do what they want with their phones like Apple does? At this point what is stopping Samsung from taking complete control of their updates? They need to invest in the server side infrastructure to host such things themselves (or find a good host/partner) and not be tied to the carriers anymore. Can you imagine the competitive edge for them if they had updates out for their top end devices just a little behind Nexus devices?

 

[slashbinslashbash]

Improving software to get better battery life is NOT a fragmentation problem, it's something required if you want to have more and more devices (like bluetooth-linked watches) connected to your smartphone. The real problem is this, which you quoted:



No one complains when new features are added to any other new traditional operating system or update, but for some reason on Android it's a major problem. I get that the pace of updates is much faster in the mobile world, but people need to get used to that.

Right, it's not a problem, we should all fall down on our knees and thank Samsung for adding functionality to their older devices that allow them to work with Samsung's newer devices.

Snark aside -- I'm not saying that it's necessarily a problem for consumers (I understand that it's adding a feature that they did not have before), I'm saying that it's a problem for Samsung because if they want to sell any of these watches, then they have a lot of work on their hands getting it to work with their older devices. Work which they could have saved by maybe not going so hog-wild with all the variations in their devices, making the same phone with multiple SoC vendors, etc. Which leads not just to problems for Samsung, but also with the development community -- who I'm sure in some cases has to test for which SoC is in (say) a Galaxy SIII, and write different code branches for each one.

 

[slashbinslashbash]

It isn't quite that easy though. Unlike Apple, Samsung has to get each release approved and bloated with each carrier's flavor of that new release. Apple did it the right way in this regard, they don't allow the carriers any say in the process. Samsung may very well have Android 4.3 ready to roll for every device under the sun but you won't see it until the carriers feel like making it available (assuming you don't have a Nexus device).

So Samsung should have their own devs and/or advocates, in-house at each of the major carriers, pushing their latest updates and working to have them approved as early as possible. Samsung is powerful enough to do this. They can play the carriers off each other just as Apple did (granting exclusivity to AT&T for a time period, while others were locked out -- but AT&T gave up all power over updates in return for that exclusivity). If AT&T started advertising that its Samsung phones will get updates within 2 weeks of their official release, while Verizon phones have to wait indefinitely, then Verizon will be forced to match it or start losing market share.

And that's being generous by taking your assumption for granted, i.e. that "Samsung may have 4.3 ready for every device under the sun". If they did, then what is the benefit to the carriers for holding it back? Wouldn't they be fighting to be the first to release 4.3 for the S4? From what I've read (reading between the lines on the Samsung press statements), Samsung is simply not ready to release 4.3 on the SIII and S4 yet. If you've got evidence to the contrary, I'd like to see it.

 

[Crono]

Right, it's not a problem, we should all fall down on our knees and thank Samsung for adding functionality to their older devices that allow them to work with Samsung's newer devices.

Snark aside -- I'm not saying that it's necessarily a problem for consumers (I understand that it's adding a feature that they did not have before), I'm saying that it's a problem for Samsung because if they want to sell any of these watches, then they have a lot of work on their hands getting it to work with their older devices. Work which they could have saved by maybe not going so hog-wild with all the variations in their devices, making the same phone with multiple SoC vendors, etc. Which leads not just to problems for Samsung, but also with the development community -- who I'm sure in some cases has to test for which SoC is in (say) a Galaxy SIII, and write different code branches for each one.

Notice the word you are using quite often: Samsung.
That is a Samsung issue. I didn't disagree with what your point (read the post I made earlier, it's quite clear I don't like the direction Samsung is head in), just that people keep pointing to a "fragmentation problem". We need to stop using those two words together. There is definite fragmentation of Android - that's basically how it was designed, it isn't supposed to be a single vendor, rarely-updated platform -and there is the OEM problem, where large OEMs act as kings of their own domains, deciding what their products will and won't work with.

The real problem here isn't on Google, not really (well, maybe a little in terms of their influence over others). Which is why I keep bringing up in this thread that fragmentation isn't the problem, nor is Google "removing OEMs and Carriers from the loop" to address fragmentation, they are doing it to strengthen their own ability to quickly update their own service-oriented apps. That's their bread and butter. That's their source of income, via ads and upgrades in those services and data collection. They don't care about Android except as it relates to that, and it's the whole reason Android exists as far as they are concerned primarily. Everything else is secondary.

 

[slashbinslashbash]

Notice the word you are using quite often: Samsung.
That is a Samsung issue. I didn't disagree with what your point (read the post I made earlier, it's quite clear I don't like the direction Samsung is head in), just that people keep pointing to a "fragmentation problem". We need to stop using those two words together. There is definite fragmentation of Android - that's basically how it was designed, it isn't supposed to be a single vendor, rarely-updated platform -and there is the OEM problem, where large OEMs act as kings of their own domains, deciding what their products will and won't work with.

The real problem here isn't on Google, not really (well, maybe a little in terms of their influence over others). Which is why I keep bringing up in this thread that fragmentation isn't the problem, nor is Google "removing OEMs and Carriers from the loop" to address fragmentation, they are doing it to strengthen their own ability to quickly update their own service-oriented apps. That's their bread and butter. That's their source of income, via ads and upgrades in those services and data collection. They don't care about Android except as it relates to that, and it's the whole reason Android exists as far as they are concerned primarily. Everything else is secondary.

That's all well and good, but with Samsung being effectively the face of Android for several major markets, perception becomes reality.

Has Samsung effectively created a fork of Android? Perhaps that argument could be made. Perhaps Google should push to change the licensing terms, and disallow Samsung (and others) being able to advertise their phones as "Android" phones when they differ by X% of the codebase, or don't offer updates within 6 months of official Android updates.

Or something like that. We had the same thing going on with the PC world, Microsoft provided the OS while OEMs provided the devices (PCs). The big difference being that Microsoft's system was closed; it could be skinned, and applications could be run on top of it, but Compaq or HP or eMachines couldn't go and mess around with the kernel, or the driver APIs or something. Furthermore, Microsoft went to great lengths to maintain compatibility with hardware and software from one OS to the next, doing things like hard-coding exceptions for particular software packages that expected the OS to behave in a particular way. Microsoft put a ton of effort into maintaining Windows while still being successful letting others sell machines that ran Windows, without being able to change it fundamentally.

It's really a pretty fundamental difference, and it makes me wonder if Android isn't really just a noble experiment that's destined to fail. (Using "fail" here to mean something quite specific: becoming forked into multiple proprietary OSes maintained by the various OEMs, which sometimes integrate changes from the official branch of Android. We already see it with Amazon and the Kindle Fire... will we see it with Samsung?)

Android, as a for-profit enterprise by Google, is definitely an entirely new business model, and maybe one that we're not really well equipped to discuss in traditional business or software-development terms. How does Google define the "failure" or "success" of Android? You're right: their business model relies entirely on the search traffic and other user-generated information, ads, etc. In that regard, it doesn't matter at all whether Android is forked, so "failure" as I defined it above might not really be a big concern for Google.

 

[Eug]

Huh? With a billion activations, Android is far from having failed.

 

[slashbinslashbash]

Huh? With a billion activations, Android is far from having failed.

Granted. You didn't read far enough.

What if it is forked into 20 different incompatible distros, maintained by 20 different OEMs, but thereby obtains 20 billion activations, most of which still use Google search, Google maps, Google ads, etc.? While that might be a "failure" in terms of traditional software development and business models (which was where I labeled it a potential "failure"), it could very well be counted as a "success" for Google financially. Which is what matters in the end, really.

 

[OS]

Blocking Google Services via firewall and staying away from actual Google services (apps) give insane battery life for your Nexus 4 in case anyone is interested. It is not realistic, but not impossible.

I've noticed going to cyanogenmod seems to delete the play store, and judging from this thread, i am totally ok with that.

 

[Eug]

Granted. You didn't read far enough.

What if it is forked into 20 different incompatible distros, maintained by 20 different OEMs, but thereby obtains 20 billion activations, most of which still use Google search, Google maps, Google ads, etc.? While that might be a "failure" in terms of traditional software development and business models (which was where I labeled it a potential "failure"), it could very well be counted as a "success" for Google financially. Which is what matters in the end, really.

That's a "what if", not an "is".

 

[Crono]

That's all well and good, but with Samsung being effectively the face of Android for several major markets, perception becomes reality.

Has Samsung effectively created a fork of Android? Perhaps that argument could be made. Perhaps Google should push to change the licensing terms, and disallow Samsung (and others) being able to advertise their phones as "Android" phones when they differ by X% of the codebase, or don't offer updates within 6 months of official Android updates.

Or something like that. We had the same thing going on with the PC world, Microsoft provided the OS while OEMs provided the devices (PCs). The big difference being that Microsoft's system was closed; it could be skinned, and applications could be run on top of it, but Compaq or HP or eMachines couldn't go and mess around with the kernel, or the driver APIs or something. Furthermore, Microsoft went to great lengths to maintain compatibility with hardware and software from one OS to the next, doing things like hard-coding exceptions for particular software packages that expected the OS to behave in a particular way. Microsoft put a ton of effort into maintaining Windows while still being successful letting others sell machines that ran Windows, without being able to change it fundamentally.

It's really a pretty fundamental difference, and it makes me wonder if Android isn't really just a noble experiment that's destined to fail. (Using "fail" here to mean something quite specific: becoming forked into multiple proprietary OSes maintained by the various OEMs, which sometimes integrate changes from the official branch of Android. We already see it with Amazon and the Kindle Fire... will we see it with Samsung?)

Android, as a for-profit enterprise by Google, is definitely an entirely new business model, and maybe one that we're not really well equipped to discuss in traditional business or software-development terms. How does Google define the "failure" or "success" of Android? You're right: their business model relies entirely on the search traffic and other user-generated information, ads, etc. In that regard, it doesn't matter at all whether Android is forked, so "failure" as I defined it above might not really be a big concern for Google.

This is why I'm hoping people buy from Samsung's competition whenever the products are just as good if not better. I don't think Samsung quite yet has forked Android (not technically and not in practice, either), but looking at their UI and dozens of features (most of which is what I would call bloatware), they definitely wouldn't mind getting there. Keep in mind that Samsung and other OEMs are looking at other operating systems and their own versions, too, in addition to customizing Android. They are practically entirely vertically integrated, so given a year or two it isn't unimaginable that they start a transition to their own OS.

There's a reason newer OS' are popping up like Tizen, Firefox OS, and Ubuntu. It's not Android that is heading toward fragmentation as it is the entire global market. Not that you won't see 2 or 3 big ecosystems together dominate, but there are going to be a lot of smaller players, too.

 

[slashbinslashbash]

That's a "what if", not an "is".

And my original statement: "it makes me wonder if Android isn't really just a noble experiment that's destined to fail. "

Not "has failed", not "is failing", but "is destined to fail." Future possibility, aka "what if".

And, again, using the word "fail" only in an narrow sense that may not really be applicable, when the broader context of profitability is considered.

 

[Roland00Address]

I think the issue is illustrated by the fact that Samsung's new Galaxy Gear smartwatch is not yet compatible with its own flagship S4 or SIII. I mean, how hard could it be? I realize the phone would need updating or a special app, but why isn't that update/app ready the day that they launch the watch? I read this and smacked my forehead. What a stupid move not to support your own flagship phones with this super-hyped product! At least have a set date for support, say "October 15 for S4, October 31 for SIII" or something like that, not just "whenever we get around to it."

According to Samsung (source is Endgadget) the S3 and S4 will get the 4.3 update sometime in October. (Carriers will of course delay this, for they will do their own testing.)
http://www.engadget.com/2013/09/04/android-4-3-arriving-for-galaxy-s-3-and-galaxy-s-4-in-october/

 

[ControlD]

So Samsung should have their own devs and/or advocates, in-house at each of the major carriers, pushing their latest updates and working to have them approved as early as possible. Samsung is powerful enough to do this. They can play the carriers off each other just as Apple did (granting exclusivity to AT&T for a time period, while others were locked out -- but AT&T gave up all power over updates in return for that exclusivity). If AT&T started advertising that its Samsung phones will get updates within 2 weeks of their official release, while Verizon phones have to wait indefinitely, then Verizon will be forced to match it or start losing market share.

And that's being generous by taking your assumption for granted, i.e. that "Samsung may have 4.3 ready for every device under the sun". If they did, then what is the benefit to the carriers for holding it back? Wouldn't they be fighting to be the first to release 4.3 for the S4? From what I've read (reading between the lines on the Samsung press statements), Samsung is simply not ready to release 4.3 on the SIII and S4 yet. If you've got evidence to the contrary, I'd like to see it.

I'm not assuming Samsung has 4.3 ready at all, I was only addressing how OS updates for everything non-iPhone work. I agree the system is crappy but that's the business relationship between the carriers and those that make the phones. Allowing the carriers to bloat the phones is part of the "subsidizing" that goes on, so it will take a shift in business philosophy before that changes. Right now I'm not sure either the carriers or the manufacturers feel a need to make such a shift.

Nobody has the kind of leverage Apple did with the iPhone, and I doubt we ever see such an example again.

 

[slashbinslashbash]

I'm not assuming Samsung has 4.3 ready at all, I was only addressing how OS updates for everything non-iPhone work. I agree the system is crappy but that's the business relationship between the carriers and those that make the phones. Allowing the carriers to bloat the phones is part of the "subsidizing" that goes on, so it will take a shift in business philosophy before that changes. Right now I'm not sure either the carriers or the manufacturers feel a need to make such a shift.

Nobody has the kind of leverage Apple did with the iPhone, and I doubt we ever see such an example again.

Take this scenario: Samsung sends execs to both Verizon and to AT&T. They tell AT&T that they've been negotiating with Verizon to launch Android updates early. They tell Verizon that they've been negotiating with AT&T to launch Android updates early. They ask each to make offers and say that it could be a competitive advantage that they could advertise, and they'd be willing to guarantee exclusive early-access to updates. If you strike this deal, you'll get Android updates for all of our supported phones, 1 month earlier than anybody of your competitors, guaranteed for a 2 year period. We'll work with your internal dev team to have all of your carrier-specific BS built-in so you don't have to worry about it. But you have to guarantee to push the updates out to customers within 2 weeks of submission.

Do you think that this couldn't happen? Do you think that neither AT&T nor Verizon would go for it? It is a competitive environment, and both companies are basically selling the same phones now. The "timely updates are problematic" issue probably rings true for a good proportion of the existing Android userbase, who might be enticed to switch carriers, all else being equal.

Of course, this just brings us back to Samsung, which apparently doesn't care enough to actually try to do this. IMO the carriers (at least in the US) may be happy with their positions right now, and they're definitely not going to advocate to change the status quo, but I think they would come around fairly quickly if Samsung or another big Android OEM started playing hardball.

They probably wouldn't even have to go through that negotiating tactic that I outlined above; they could just simply publicize when they release updates to the carriers, and tell users "it's out of our hands, go complain to Verizon" and basically shaming them into releasing the updates in a timely manner. But again, this would require the desire on the part of the OEM to do this, which I don't think is really there.

 

[ControlD]

Take this scenario ... snipped ...

I agree with almost everything you are saying. I don't think the desire is there on the part of Samsung or the carriers. Samsung is happy to be selling gobs of phones and the carriers are happy to be getting / holding customers for two years for each of those sales. Until something comes along to shift the industry (like a strong pre-paid option perhaps) I'm afraid things probably aren't going to change which is unfortunate.

What I don't know is what kind of contracts or other business dealings are at work. Can Samsung simply come in tomorrow and tell Verizon they are changing the way they handle OS updates? Are there existing contracts and agreements already in place?

 

[slashbinslashbash]

I agree with almost everything you are saying. I don't think the desire is there on the part of Samsung or the carriers. Samsung is happy to be selling gobs of phones and the carriers are happy to be getting / holding customers for two years for each of those sales. Until something comes along to shift the industry (like a strong pre-paid option perhaps) I'm afraid things probably aren't going to change which is unfortunate.

What I don't know is what kind of contracts or other business dealings are at work. Can Samsung simply come in tomorrow and tell Verizon they are changing the way they handle OS updates? Are there existing contracts and agreements already in place?

Well, in the long run, I think that spells big problems for Samsung and other non-Google Android device makers. There's no way that they can remain competitive, long-term, with Apple, MicroNokia, and GoogMoto, all of which are going to be able to tightly couple the software to the hardware, push OS updates when they feel like it, and create better experiences because of it.

It's almost like... wow, that would be really interesting if it were to be shown that Google has planned all of this, and they gain, say, 50% of Android market share within the next 3-4 years. Maybe Google is just using Samsung et al as temporary placeholders before brushing them aside once they get some new Motorola devices pushed through the pipeline? Kind of like how MS pushed Windows RT to various OEMs, then cut them off at the knees by releasing Surface.... that would be true evil genius stuff if it happened. And maybe it won't happen. But if I were Samsung, I would be watching my back and building up a strong internal OS development team. Maybe even buy BlackBerry, or try to purchase Palm's assets from HP.

 

[BrightCandle]

I actually don't think the issue is Samsung or any of the manufacturers really. The carriers have the longest time been the ones slowing down the update process. They have a different business model based on selling phones at premiums along with bundled deals and upselling phones mid contract when a new shiny thing comes out. It is definitely not in their interest to provide upgrades continuously to their customers.

Google originally released the Nexus phone away from the carriers for a reason, to try and start a revolution in the mobile world away from the grasp of the providers. It didn't sell very well because of the large capital price up front. They changed tactics, sold through the carriers with the deals and the carriers along with the manufacturers created a fragmentation issue that they felt brought them more sales. Ultimately Google cares about winning the war for mobile search and other apps of its interest so fragmentation is largely a none issue for them.

But you have to realise that fragmentation is a temporary problem. We all have different Windows versions on our machines with different hardware but in the end it isn't very important. In the end what will happen is the APIs will have matured, the hardware will have matured and the open market approach with an open source OS that can be customised but still runs any Android based software will prove to be great. It is significantly easier to put an app out for Android than it is for iOS. The very fact that Google does no testing at all and leaves it open to the market ensures a smoother and quicker release process, in essence platforms rather than closed ecosystems have historically won in the tech industry.

So while as a developer I have an issue with how annoying supporting old devices is, pretty soon it wont be an issue at all, everyone will be on modern versions. The latest versions differ a lot less than say 1.0 to 1.6 and 1.6 to 2.0. From 2.0 onwards there are some great features (animation system, alpha blending and lots of others) but they matter less than those initial jumps. The same was true of Windows. The early jumps were enormous but since XP there hasn't been a massive advance that makes the grand majority of applications really want to require the latest version (beyond games).

For now the fragmentation is in my opinion is mostly the fault of the carriers and in the end the OS difference wont be worth talking about. The future is still many different devices and OS' or perhaps no packaged OS at all. But just as fragmentation isn't a user issue from now on in the near future it wont be a developer issue either as little extra of value needs to be added to the APIs.

 

[crystalcollins]

Yes, I agree with you all.

 

[poofyhairguy]

Has Samsung effectively created a fork of Android?

I think of it like Ubuntu and Debian:

Samsung starts a new "fork" for every Android release.