Many of M's changes are "under the hood" and less obvious to the average user. If you compare a Lollipop handset and Marshmallow handset side by side, the user experience is very similar. M wins, but the differences are more subtle than, say, the Kitkat-to-Lollipop transition or, especially, the Ice Cream Sandwich-to-Kitkat transition.
The other factor at play is that people are using larger, more functional handsets than ever, with larger batteries that are going to take longer to wear out (albeit, increasingly non-replaceable batteries). Screens are more scratch-proof and cases are preventing breakage. A 5 inch phablet from a year and a half ago is still pretty functional today and has replaced people's 9.7" iPads for many day-to-day uses. People are not replacing their devices at the same pace as in 2011.
Plus, it's going to be months before OEM's get around to updating their huge installed base. I wouldn't hold my breath.
Has Android ecosystem ever given a hint that a device would be supported after 2 years? Certainly for phones you are supposed to get a new one from your carrier.
But .. why?
I mean - if that device works perfectly fine, why buy a new one for a _software_ upgrade?
And even that won't help in general (I was bitten by HTC, LG and Samsung in the past. Basically every Android vendor is crap and the GP calls out Google for being only slightly better).
Because they've figured that the group of users remaining on older devices and actually caring that they run the latest software is pretty small and they don't want to dedicate the resources needed to support them.
Because software upgrades more often than not includes a lot of bells and whistles. You can go ahead and try to update a Pentium 3 to Windows 10. Will probably work, but it wasn't intended to.
Okay, the Pentium 3 is not 3 years old, it's more like 20 (soonish). I'd bet that there are quite a lot of machines out there that are >= 3 years old, running Windows 10. Or more appropriately, running a recent Linux distribution.
Sure, at some point the hardware capabilities become a problem. But I just don't see that for a random operating system (Android in this case), after 2-3 years.
The real problem seems the infrastructure - every vendor claims that it is a hellish effort to move all their useless crap and customizations, all their closed source blobs and magic fluff to a new AOSP base. And _there_ lies the core problem. If vendors would stop messing with random UI changes to 'be different', if manufacturers wouldn't hide their drivers, only THEN would your hardware argument come into play.
If I technically _could_ upgrade my late HTC Hero to Android M I might find the performance lacking - although again: I fail to see why the basic operating system wouldn't work just fine, playing games or trying do watch HD videos on a tiny display notwithstanding.
Android L runs fine on a tepid single core ARM CPU with an equally unexciting GPU. I'm using a SoM like that in an embedded system. It should run fine for many years and several Android updates. High end handset specmanship is completely disconnected from actual system requirements.
Because it should be reasonable to have new software on two year old hardware?
I don't get why I should need to throw away a perfectly functional screen, battery, CPU, RAM, camera, etc. because I'd like an improved permissions system.
Isn't the problem with Marshmallow adoption due to the fact that you just probably can't get it on your phone due to carrier/manufacturer nonsense?
Apple pushes their OS to all devices, even if it bricks them. After iOS9 upgrade my iPhone 4s is practically unusable :)
In the end the consumers have very little choice in this matter. I can't go back to iOS8 (and the bloody phone was bugging me every single day to upgrade, so finally I did). Android users have not much say in this regard either.
This article would be more accurately titled "Only 1.2 Percent of Phones Are Allowed by the Telcos to Run Android Marshmallow". Do you think it's due to any kind of apathy on the user's behalf? I'd be thrilled to have Marshmallow. I upgraded to an HTC M9 from my M7 because HTC promised that the M9 would get Marshmallow.
Nonetheless, it's four months later, and I'm still waiting for Verizon to allow me the privilege.
I perform regular network audits and it always amazes me how many insecure, outdated android devices people are using - often they're even running services that are remotely exploitable. Quite often we find that such devices run phone-home beacons and it can be quite interesting what data is being sent back to Google and URLs presumably related to installed applications. On a semi related note last week I found that someone's brand new Samsung tablet was seeding torrents, I spoke with the person and we looked over the device carefully, it was not at all clear that although they thought exited the application it was still running and seeding out what looked to be updates to peers online without any indication.
Most smartphones manufacturers tend to stop to update them quickly (it doesn't go more than 2 years for non-major products). And low-end devices cannot run newer OS/softwares.
I stopped updating most of my softwares on my phone because it was making my phone too slow, even by deactivating notifications/background processes I did not care too much about.
That's a feature, not a bug. I want my seeding/downloading to run the background. I suspect he doesn't understand how to properly close the application.
Well, one is a mobile OS and the other is a desktop OS. And the reason Windows 8 got such bad press was because of the UI overhaul that was critically panned by the press and rejected by consumers. Metro/Modern UI or whatever they're calling it nowadays is still a failure.
You wouldn't believe how much data Windows 8 and 10 send and receive on a well populated network with Internet access, we observe a great deal of phone-home requests and often have found Windows connecting to external up addressees that appear to be within azure that are sending unencrypted or overly verbose data in the background.
I don't think Android-with-Google-Play-Services (not AOSP) any better in this regard. Never owned or observed network activity of an Apple device, so no idea about those.
<offtopic:rant>Gosh, I'd wish there'd be a wiki full of packet dumps and protocol analysis, showing what modern software really is. So when a topic "$software is spying on users" happens one link can show what's exactly is sent (sure, there'll be a lot "some identifier, meaning's unknown" stuff there, but still), and, possibly, refutes popular myths too.</offtopic:rant>
I also have that model, bought the Motorola because an update to M was promised to come "soon". Three months later I'm still waiting on it, it's disappointing it takes so long and nobody can tell you when it will come.
There are already many applications that won't run on my Cyanogenmod 4.4.4 because I refuse to have any Google binaries onboard... I hope that this is not the future.
It's disappointing there really isn't a great high-end Android phone these days that's competitive with the iPhone 6S. All the mainstream higher end units either have screens that are too big or lack something in features IMO.
The unit size is not proportional to display size.
iPhone 6S has display size comparable to Xperia Z5 Compact, yet the Xperia is a much smaller phone. iPhone is sized similarly to the original Nexus 5, yet the Nexus has much bigger display.
Even without any bezels at all, the 5"+ screens that seem to be the cost of entry to high-end Android devices would still be too large for many to use comfortably with a single hand.
Though the 6s and Z5 Compact actually have very similar ratios of screen coverage: 65.6% on the 6s[1] vs 68.9% on the Z5C[2], according to GSMArena.
Well, you might try the 5" phones. Two years ago, I also thought the same, and then I got the original N5. Last autumn I changed the phone to Z5, not Z5C, because I considered Z5 more comfortable to use.
> Though the 6s and Z5 Compact actually have very similar ratios of screen coverage: 65.6% on the 6s[1] vs 68.9% on the Z5C[2], according to GSMArena.
The 11 mm difference in length is subjectively huge. It's best to have both in the hands, at once.
Curious to get feedback from game developers, what their experience has been working with multiple devices, software version and now Android TV. What's the best way to get up to speed with shipping games on Google Play?
Android game developer here. The fragmentation is massive and annoying and very hard to deal with. I've found it best to try to work around the problem by minimising my OS dependencies where possible. This is actually easier for games than for apps -- games can take over the whole screen and have their own weird UI, while apps have to integrate nicely.
I'd say:
- Find a nice robust framework. I'm using the built-in NativeApplication and I regret it. I think SDL might be a good bet. Lots of people like Unity (but it seems very heavyweight to me for small games).
- For a portable game, C or C++ is your best bet right now. (Or Unity's C#.)
- Pick a nice baseline API level and stick to it. You can actually go right back to API 8 (Android 2.2!) and you get OpenGL ES 2.0, which is an important milestone for games. I'd like to update my games to ES 3 at some point, but that only works from Android 4.3 onwards (fewer than 50% of users) so I'm holding out for now.
- Avoid using Java if you can help it! Seriously.
- Avoid using the file system if you can help it. The APIs are very bad and the hardware is often bad too, many manufacturers seem to cut corners here.
- Figure out a good crash reporting / feedback system. I haven't actually found a good one for native apps, I'd love advice on that! HockeyApp is great on the Java side.
- Figure out a good testing system. I ended up just buying lots of old phones and tablets on eBay and Amazon. Getting a good mix of screen sizes and CPU types is key (there are quite a few x86 tablets out there, it turns out).
Ooh, good question... I just started using Steam for a PC port, and its killer feature is the way it only syncs the changes for each build. So I can fix a bug, upload an update, and get it in players' hands within minutes. I'd really like to see that on Google Play too.
In Play you currently have to split your app into "expansion files" if it's above the maximum APK size, upload each file in full, and then it takes hours to actually push the update out to users.
Apple and Android have very different strategies here. The result of Apple waiting until the OS works on most hardware means they never have articles like this published.
The real headline should be this:
Android has 1.2% of users using a brand new OS that Apple would never ever release until >50% of users could be using it.
It's not that other users and phones are behind. It's that Android itself is so far ahead.
One of the reasons I bought my particular variant of my LG G4 was the unlock-able boot-loader. Running ROMs of my choice helps me stay up to date. I have been on marshmallow since this past Dec and am glad to report everything runs very well. I no longer force 2D acceleration and the UI is very smooth and responsive.
I kind of liked some of the 3rd party ROM solutions to permissions.
With Google's new system you can only grant or deny right?
... then the application can behave differently depending on what the user chose.
The 3rd party ROMs I remember had another option to supply the application with no contacts, or fake locations, etc but the app was none the wiser.
Kind of bad when you're running a stock factory image on a Nexus device (Nexus 5) and your battery can't last a day at the office with next to ZERO usage.
Interesting, one of M's big innovations is supposed to be better battery usage. At least one report I've seen[1] suggests that standby time on a Nexus 5 is improved by Marshmallow. Have you studied the battery usage report to see what's sucking up all the wattage?
The N5 is a terrible phone in general. Even with the OS it launched with, battery life was subpar. Worse than even budget phones of the time.
I'm glad the Nexus product line has given up on the 'budget' end of things. Battery life and camera quality in the 6, 6p, and 5x are much better. They cost a bit more, but its more than worth it.
When I had Nexus 5, the battery lasted for two days of normal office usage. (I.e. no games, no permanent facebooking, no permanent browsing, having good signal and the only "optimization" was turning Google Now off).
Hardly surprising: an OS with a major revision number, released only 4 months prior (including the busy Christmas season), hasn't yet been tested-for-release by telcos and pushed into shops or upgrade servers, then finally filtered out to the lazy 98.8% who couldn't be bothered disposing of their phone and buying a new one (or rooting and upgrading their phone manually or abandoning their contract to get the marginal improvements over now-mature products).
Must remember to write an article on this strange phenomenon.
> the lazy 98.8% who couldn't be bothered disposing of their phone and buying a new one
Or who don't want to spend hundreds of dollars on a new phone every year. My phone (S3) is 3 years old and works fine. Why should I need to buy a new one?
vacri's comment made me chuckle. I do not understand why someone thinks it is an acceptable practice to force the replacement of adequate hardware due to the lack of software updates.
When security should be a concern in this ecosystem, vacri tells users to root their phone and presumably either load some amatuer dev's ROM or port and compile their own.
I just don't see any of that being helpful for the average user.
then finally filtered out to the lazy 98.8% who couldn't be bothered disposing of their phone and buying a new one (or>rooting and upgrading their phone manually or abandoning their contract to get the marginal improvements over now-mature products).
This is the bulk of what I was referring to in the comment. The first line is a blatant attack against anyone who is not willing to part with their phone. By even telling users that is what they should do with their Android phone, the author is supporting this viscous cycle.
I read that as sarcasm, it's making fun of the idea of writing a story about limited adoption after such a short time (and given the mentioned hurdles to switching), not actually mocking the 98.8%.
To me, the real story is that if you cover off Kitkat and Lollipop, you've got 70% of the Android market. Add Jellybean to the mix and you've basically got the entire userbase covered.
For a platform known for fragmentation, that's not nearly as bad as I was expecting.
There's also Android and Android, you need also to test for each popular product of Samsung since it's so different from stock and then hope there is no carrier-specific problem. And you also need test with a lot of different screens, but it's not like in web, you can't resize the emulator like this so it takes longer. Not trying to criticize, but it's still not easy to get an Android app to work mostly everywhere.
In essence. This is easier than it sounds, though, as there are a number of support libraries out there that provide emulations of newer APIs for older versions of Android.
The other factor at play is that people are using larger, more functional handsets than ever, with larger batteries that are going to take longer to wear out (albeit, increasingly non-replaceable batteries). Screens are more scratch-proof and cases are preventing breakage. A 5 inch phablet from a year and a half ago is still pretty functional today and has replaced people's 9.7" iPads for many day-to-day uses. People are not replacing their devices at the same pace as in 2011.
Plus, it's going to be months before OEM's get around to updating their huge installed base. I wouldn't hold my breath.