Jun 01 AT 3:20 PM Sean Riley 36 Comments

Android will move to a yearly update cycle

In a recent interview with Mercury News, Google’s VP of Engineering for Android, Andy Rubin, acknowledged that the current pace of Android OS development is not sustainable long term and that they will scale back to a yearly update when “things start settling down.”

Rubin explains in the interview that the speed of updates to date has been due to their need to bring the “product up to market spec.”

OEMs and developers will likely be the most gladdened by this news, but users will hopefully realize that long term this is necessary for the well-being of the platform. Talk of fragmentation constantly dogs Android and it will not stop until there are at most two versions of the OS, whereas the release of Froyo has us sitting again at four.

I suspect we all loved watching Google I/O and hearing of the innovation coming in Froyo and beyond — but each time there is a new OS release the joy is always tempered by the question of when current Android device owners are going to actually see these updates.

Some lag is inevitable in an update process that involves so many players, but it would be more tolerable if you didn’t have to see three more versions of the OS ship before you finally managed to get your update, just in time to be once again passed up by the newest version of the OS.

App developers are also likely as weary of hearing about their apps only supporting the most recent versions of the OS as users are of finding that the latest and greatest app they heard about doesn’t support their version of the OS. This should be a win for both sides of that equation.

Do you agree that this is a necessary move for Android and if you think that a year is too long to wait between updates what do you think the ideal update timeline would be?

So we launched it, and from our internal 0.8, we got to 1.0 pretty quickly, and we went through this iteration cycle. You’ve noticed, probably, that that’s slowed down a little bit. Our product cycle is now, basically twice a year, and it will probably end up being once a year when things start settling down, because a platform that’s moving – it’s hard for developers to keep up. I want developers to basically leverage the innovation. I don’t want developers to have to predict the innovation.Andy RubinVP of Engineering for Android at Google

Via: TechCrunch

Source: Mercury News

Sean has been with Android and Me for over 8 years and covering mobile for the last 9. He occasionally muses about gadgets and tech outside of the Android universe at Techgasms.

    Most Tweeted This Week

  • http://Website Kevin

    Ultimately, this is the right decision. Rooted users will be happy because they can have fun playing around with ROMS for 11 months, then go back to stock with the knowledge that an update is definitely coming. All android users will be able to make intelligent choices about what phones they want without sweating if their new phone will keep up with the next phone a month from now.

    • http://livinginagoogleworld.blogspot.com Jonathan Frederickson

      Or we could just keep them rooted and get the new versions through less official means. :P

  • http://Website david

    4-6 months between updates

  • http://www.fishmemory.net fractalbit

    As a nexus one owner i enjoy the pace we are getting the updates. But in the end i agree that a 1 per year cycle is more reasonable and will benefit that market and the platform. Meanwhile though, i can’t wait till we get gingerbread :D

  • http://Website ari-free

    we really needed app2sd as quickly as possible because there was a storage problem. Phones came with too little internal memory!

    • http://Website Xallies

      Froyo has that already

  • http://www.typhon4android.org/ Mike Leahy

    I too look forward to a yearly cycle as a dev. The 1st year (’09) for real time apps / games (OpenGL ES) was a little rough especially Q4 ’09. I think as long as there is solid and refined effort put into Gingerbread it could be possible to start a yearly cycle from that point onward. It would be fantastic if the yearly cycle coincides on release with the majority of recent gen devices such that there isn’t a long wait after release for various flagship devices to be updated. Hopefully a year lead time will also give various manufacturers who are providing custom UI implementations on top of Android to also have their systems well tested and ready to go when a new Android OS is ready to ship. I think it’s real important for more than the N1 to be upgraded immediately upon release of a new OS version. It will also be nice if a developers preview release is available a month or two before wide OTA release such that devs have a chance to update and test on device well in advance (this could be just for the N1 or follow up Nexus / Google affiliated devices). .

  • http://twitter.com/aaronhutchinson Aaron

    As a N1 owner i have to say I’m loving the speed of things at the moment, however i also understand that a more frequrnt update cycle is more stable for the Android community as a whole.

    I love FroYo, but knowing that I’m going to have to wait for months before i start to see any benefit in the market.

    Bring on annual updates!!! :)

  • http://wiredwonder.com agentlotek

    This will surely help the Android platform. They should start the annual update cycle to coincide with the release of gingerbread.

  • http://plankhead.com Zacqary Adam Green

    This is probably a good move, but I’d love it if Google took a page from Mozilla’s book and offered official betas of new versions. At least on the Nexus.

  • http://Website carig

    This is great as long as they keep up the pace with regards to the OS’s developments and improvements. Android needs to blast past the iPhone OS and not just be brought “up to market spec”.

    This should work out great in conjunction with the change that’s being said to be made in Gingerbread towards detaching some core apps from the OS update cycle and making their updates downloadable from the market.

    Now the question is: will they match the iPhone with yearly summer updates or stick with winter updates instead?

    • SGB101

      i like winter updates, gives me something to look forward to when its cold and raining, plus all the bugs will be ironed out for the sunshine days out :-)

  • http://Website Jarl

    i dont believe changing the update cycle alone will fix the fragmentation, some devices simply dont get updated, because of choices made by carriers and manufacturers. the magic and the hero are fine examples of that

    i guess thats the thanks you get for being an early adopter

    i think google should apply more pressure on manufacturers/carriers to update their devices

  • http://Website B

    Im blessed enough to have a Nexus, but I agree that this is a necessary move for the growth/adoption of Android ad a whole. I just hope that slowing things down doesn’t make the innovation stagnant as a result.

    Also, wasn’t there word that devs now have the ability to update core apps straight to the market, effectively allowing ‘outdated’ users to at least somewhat keep the pace without actually running the newest build of Android? Did I read into something wrong?

  • http://Website Leonard

    I think a year is TOO MUCH time, twice a year seems totally reasonable to me ala ubuntu, hope this isnt a final decission =/

  • http://www.typhon4android.org/ Mike Leahy

    Ok this gets slightly technical, but I think it’d be rad if it were built into Android soon. I might mention too that I think one more 6 month cycle with a June ’11 release is reasonable to move innovation well ahead on a whole and to be absolutely sure that core architecture is really refined.

    Personally I wish, and this is dev centered (everyone benefits though), that there was a way to facilitate potential non-core API / OS / device specific updates in a component oriented fashion. OSGi is the Java tech I’m using for the middleware, TyphonRT, that I’m putting out soon. But technically as long as similar infrastructure was built into Android (Gingerbread or next release) to be able to update non-core OS level functionality / Android app API level functionality and perhaps even store such additions on the SD card or increased internal storage (non-ROM based). The general idea is that non-core OS level functionality targeted towards app development can be updated at will for any given app that requires that functionality. Once this additional functionality is dynamically installed on the device it’s available for any app to use and only the 1st app that requires such functionality will trigger a download of new OS level API functionality and each subsequent app that requires it would have immediate access. Yeah.. a dev idea / concern for sure, but this would allow changes during the year in between major OS releases without having to push out a whole OTA release.

    To give a small example on the real time app / game dev front of how something like this would be a benefit to everyone and allow further innovation between official releases . Consider that all 2nd gen devices (Droid, N1, etc.) support OpenGL 2.x. On the Java side of things an official Java binding for OpenGL 2.x isn’t available until 2.2 / FroYo even though all 2nd gen devices that may be running 2.0 / 2.1 can utilize OpenGL 2.x. Once the Java bindings for OpenGL 2.x were available they could be pushed out to all Android devices that support it in hardware (Droid / OMAP3, N1 / Snapdragon based devices, etc.) regardless if they are running running the latest and greatest OS (in this case the Droid still on 2.1) if there was functionality built into the Android OS architecture to facilitate updates as described in the paragraph above app developers could take advantage of official APIs across devices with a variety of OS versions rather than waiting for all devices to reach X OS release (in our example FroYo).

    I know a little technical perhaps, but that would be my dev dream for device specific Android OS / API level enhancements between official OS releases especially if the release cycle became yearly. Think of it in a similar was as Google separating their apps from the firmware/ROM and putting them in the market; therefore non-core and even device specific OS API level functionality too could be updated between major yearly OS releases. A system like that would go a long way to allowing constant innovation without having to force an entire major OS update and / or delay useful app / API level functionality for a year!

    • http://www.typhon4android.org/ Mike Leahy

      Heh heh.. I went a little mad scientist as things go, but don’t neg it just for that as I did try to explain a solution to larger fragmentation concerns; I’m also aware this is not a dev blog. Those that negged are free to provide their solution.. ;P It really is a catch-22 and I do believe we’ll see a new and more onerous type of fragmentation when innovation with Android is perhaps shifted to the manufacturers during a yearly major release cycle for the OS as devices proliferate thus forcing innovation to require new custom APIs that are going to be really device specific; the onerous kind of fragland especially when functionality is duplicated between devices, but different (HDMI support & APIs varying across devices is a 1st potential example of this). On one side for real time app & gaming a yearly release is great because we have all we need to do great work really now after Froyo (sans a totally complete bluetooth stack and universal HDMI support). On the other is continued innovation by device manufacturers adding new functionality and really fragmenting Android. Now we may also see that a yearly release cycle will discourage manufacturers from adding new hardware functionality that simply is not supported by the OS and choose not to add custom APIs or innovate. If a yearly release cycle slows down innovation across the board then well.. It is what it is at that point. Just saying.. ;P

  • http://Website joe huff

    Hello everyone. I would say they should bring the os update as often as needed. Maybe they could offer the customers a option too upgrade or not . I believe that would be a lot better. As for me I’m running motorola cliq with 1.5 and have been for awhile. I would love just to see one update, nevermind two or three heck have to have one before two or three.

  • http://Website Tim

    If this is where android is headed, then i believe that all devices should be able to be easily rooted. Without the ability to fully customize our device with a different rom, then the platform is not truly open. Yes most devices have been able to be rooted, but i highly doubt Verizon of all carriers will allow this to continue. You can say that there is always a way around it, but many devices need an OTA before they can be rooted. One thing is for sure, i wont leave my current device(Moto Droid) until another device i want has been rooted.

    • http://www.typhon4android.org/ Mike Leahy

      Indeed it would be nice if every device could be easily rooted and custom ROMs installed. I agree with you whole heartedly that if it really went to a strict yearly update especially with some devices not being able to be rooted that could definitely be a bummer.

      I do believe the direction I described above is a good middle ground that allows SDK and perhaps even NDK support to be updated officially between major yearly releases. It also opens the door to device specific official SDK updates to exist along side normalized SDK functionality (eh one step technical level I’ll skip for now via this discussion).

      In many ways I’d like to see the yearly updates cover low level / core architecture updates. Just like splitting off Google apps away from the firmware/ROM it should be possible to split away OS/SDK level API updates from being directly tied to the firmware/ROM.

      With the middleware I’m releasing, TyphonRT, once it hits v1.0 later this year the 1st app that uses TyphonRT is small (just app code; for instance Auriga3D / 3D FPS engine built with TyphonRT is just ~200k) will download ~1-2 megs of TyphonRT support API / middleware to the SD card first time it’s run. The next app that is built with TyphonRT will also be small (~200-500k or smaller) and not have to download the 1-2 megs of TyphonRT middleware since it is already on the SD card. SD card is switched then TyphonRT middleware is downloaded again when the next app powered by the middleware is run. If there is an update to a TyphonRT component or a new component is required by a new app then just that new component is downloaded once the first time the app runs.

      Now if that could just be extended to the entire Android SDK / app level API for app dev that would be fantastic and solve a lot of concerns about innovation being stifled by a yearly major release cycle. It also provides an official mechanism for updates that doesn’t require end users rooting or installing anything custom or not approved.

  • http://Website Franklin

    its a good idea, cos i am using my touch 3g 2.1 and unlike any other android phone it cant be rooted easily , so i cant install custom OS. but if they release the OS on a yearly base, it will give phone makers time to send out the updates to the costumers. i am still stock on android 1.6 whilst the actual android OS has hit 2.2 thats the third update after the 1.6. Android people have to slow down cos most of their customer dont get the updates in time then another update comes out. And the other problem i have with android is that, there are so many android phones. they should do something about it, like each phone make should make one super android phone like the sprint EVO 4G and the Samsung galaxy S .they too should update their phones yearly just in time for an android update.

  • http://facebook.com/xxgearratioxx sal

    Yes, I do agree with this. And believe it would be in googles interest and the well being for the android platform to slow down updates there’s only so much you can do without running out of features at such a quick pace. Although it is nice with all the updates, but I think it will prolong the life of android. If my opinion is off. I welcome all views. And I welcome all android lovers like my self to network with me on facebook. :)

  • chris0101

    It was inevitable. As a platform matures, it gets fewer and fewer updates. To be honest, I would have liked to have updates every 6 months as is, if only because we can stay ahead of the ball.

    I think that custom ROMs like Cyanogen are going to become more and more popular due to slower updates … for phones that can install them anyways.

  • http://www.typhon4android.org/ Mike Leahy

    There is a humongous catch-22 insofar that presently three devices (G1 / Ion / N1) have been the major focus for Google surrounding Android feature development. So far Google has been playing catch up with exposing features that these three devices were capable of (N1 for Android 2.x+). Not all devices are created equal though as we see with custom APIs HTC / Sprint provided for the EVO 4G for the front facing camera and HDMI. What API does the next front facing camera device use for instance; the one Sprint created or their own custom API.
    Google has to play catch up and these will be serviced quite likely with official APIs in Gingerbread. The red herring is that these are just 2 handset devices (N1 and EVO 4G) that are presently guiding Google’s Android development process. Unless there is an official system as I outlined above that can update SDK features in an organized manner in between full firmware / ROM updates fragmentation will just explode as manufacturers of all sorts of devices (not just handsets) will be forced to add custom API code into a custom ROM/firmware of their particular build of Android. It’s one thing to think hey we’ve uncovered all of the useful functionality / main core functionality for smartphones as we know it today. What happens to the feature phone or wacky smartphone that has two displays released right in the middle of a year long release cycle. Well, the manufacturer is going to have to create a custom API and a custom ROM based on a 6 month old official build of Android. Everything I described in this post and the ones above I knew as the solution from day 1 that I got the G1 and started work with Android.

    I’m afraid lengthening the release cycle to 1 year will help with some aspects of fragmentation, but hinder many others without a comprehensive way to deal with custom SDK additions.especially as Android has a wider application outside smartphones.

    So yeah.. As much as I’m excited to see things perhaps slow down, it ain’t gonna be perfect or pretty as a whole other side of fragmentation will rear it’s ugly head and no one is talking about that yet. Google really does need to look into the future beyond the next release if they hope to tame the beast of fragmentation; they admitted recently they aren’t even doing that, so get ready for a bumpy ride despite a conceivably longer release cycle.

  • http://Website deeb215

    i think twice a year would suffice and get cracking on that market immediately. i mean revamp the whole design. it’s way too unorganized. come on android, at least give more search features, bookmarks/favorites, and video previews would be sweet as the gingerbread it should come in :)

  • Mr.Droid

    Maybe now there won’t be so many left out Android Phones. Some are still running OS 1.5! Now for those you bought the Droid Incredible can rest at ease knowing they will be kept up to date for one whole year.

  • http://Website JGarrido

    A year is fairly reasonable, but I’d prefer open betas, and point release bug fix updates every six months.

  • http://pryvateiD.tumblr.com Dave

    Yess, so everybody, all manufacturer’s & the consumers will be able to have & receive the update at around the same time frame, instead of having to wait countless months to have the updates. So, like Motorola can be able to work around implementing it’s MotoBlur into the new Android, Samsung would be able to implement it around TouchWiz & so on & so forth.

  • http://Website Ron

    My biggest issue is with T-MO they were the 1st carrier to even start the Android train and in my opinion they are the furthest behind. Months ago they were really pushing the Mytouch, they had all those stupid commercials, and it was supposed to be there flagship phone. Well I bought one and now they are giving them away. They are all stuck with 1.6 which is ridicuous, and it seems with these Android phones if you want the newest version of the OS you have to buy a whole new phone altogether, especially if you’re not tech savvy enough to root it. Imagine if Apple made you buy a new I-phone everytime there was a OS update, I-phone users would go nuts. That’s another huge problem with Android too many phones and they are not all up to spec. Google should tell phone makers “Listen if you want to use our software, the phones need to be built up to our specs, so that the customer can receive firmware updates” .

    I spoke with T-mo and they are now pushing the Mytouch Slide, gimme a break, they are the only comapny that has 3 versions of the same phone (G1,Mytouch,MytouchSlide) Why not come out with a game changer. Sprint has the Evo, Att has the Iphone, hell even Verizon who used to have the worst selection of phones now has some of the best ie……Incredible, Droid…..

    I am now trying to get out of my early termination fee with T-mo to go and buy the Evo, anyone have any good ideas that may work?

    • http://Website Franklin

      I feel u Ron, am in the same situation like u are right now. i wanted switch to sprint when i first saw the EVO 4G…..that phone have everything i want in a phone but Ron remember their 4g speeds its no match for T-MO’s HSPA+ speeds. And trust me t-mo is cooking up something for their 4G speeds so please dont switch to sprint….Give them till the end of the summer to see what HSPA+ phone they come up with before u switch..thanx!

  • http://www.typhon4android.org/ Mike Leahy

    I like this comment regarding fragmentation:

    Essentially in the posts above what I brought up is a scheme similar to what the commenter on the Slashdot article brings up below about a dev mentioning a solution at Google I/O. The relevant portion in regard to my posts above:

    pocopoco from Slashdot article
    “Meanwhile Google isn’t even interested in solutions to these problems from what I’ve seen. One developer brought up another potential solution during a session at Google IO. He suggested making the highest level of Android a distributable framework, like .NET. This would allow updating it much easier. Not nearly as many phones would be stranded with old, buggy versions of the Java portion of Android at least. The Google staffers just brushed the idea off without even discussing it. They said fragmentation should really be called progress and to deal with it.”

    The latter meme has been reported here direct from Chris DiBona:

    It is so painfully obvious that Google is not “eating their own dog food” so to speak regarding app development and Android. I’ve been working intensely with Android for 18 months now and I’ve seen little to no effort by Google to avoid OS fragmentation.

    Slowing down the release cycle will not solve it; it will only make it worse for the manufacturers that continue to innovate as Google takes the back seat.

    A distributable manner to handle the top level Java SDK /. API and even NDK (driver level issues) that allows 3rd party manufacturers to easily add SDK additions without custom ROMS / firmware and or Google / manufacturers to back port fixes to the SDK without the need of a new ROM / firmware OTA is a very valid solution. It really isn’t that hard to see that this makes a lot of sense. I also don’t think it’s overly complicated to implement.

    80% engineering and lack of foresight or future planning via Android team in addition to the culture of beta at Google is gonna bite them hard once they slow down without addressing all of this. It sort of works for web apps, but is dreadful for OS development.

    The answer is clear as day; Google just needs to fill their own dog bowl and chow down.

  • http://me-a-droid.blogspot.com/ A-Droid

    I don’t think it’ll be a benefit if we will stick with the same version for a year.
    The real benefit would be to automate the update process regardless the manufacturer/UI/ROM version.
    Make the a core layer that can’t be changed by manufacturers in any ways and update that core automatically for everyone, at once. ..and you can make it each month if you like, everyone will be glad. Even developers like me.


    • http://www.typhon4android.org/ Mike Leahy

      Yes.. This is essentially what I’m advocating for regarding how the Java SDK including specific OEM APIs specific to a given device could be distributed. It may though I can’t say if it’s possible even cover native drivers such as being able to update the native OpenGL drivers when a bug is found. This later issue is so key to providing a useful experience for end users and peace of mind for developers. Having to work around OpenGL bugs in the drivers for particular devices is no fun. If an OEM could update it without having to push out a complete new ROM / firmware (again not too sure how this will play out for drivers or if it’s possible) that would be fantastic. I certainly know it should be possible for the Java SDK and Java API additions that OEMs may add. This would go so far in making Android a powerhouse especially between yearly updates. Google please consider this and get off the fragmentation is progress meme!

  • http://Website Al Buckner

    I personally think that they need to try to stick with making the updates at least every six months. I understand the need to keep up with the market, and the android manufacturing issues, but at the same time the more they release the new updated OS systems for all of the new android phones the 1st Generations are getting left behind. Im still running the 1.6 OS, which is ridiculous. I can understand possibly slowing down the productioal updates for the new OS’s if you release the current Android 2.2 OS for any of the first generation android phones. We spent just as much money on the G1/Mytouch3g as anyone with a new android phone but we arent even kept up to date. I dont think that its fair for every new phone to have the new updates but the original android phones can even recieve these updates with out having to root our phones, with the possibility of breaking our phones at the same time if we dont know exactly what were doing. I think before Google decides to go with this plan to possibly slow down to once a year updates or every 6 month updates, they need to send out the Android 2.2 OS version to all of the phone carriers so that anyone with a 1st Generation Android will be up to date with all of the new phones that are coming stock with the newest version of 2.2 already on it. Its like we were the guiney pigs of the android market being the first ones to recieve and purchase the android phones, but we arent reaping any benefits of the progress google and the android operating systems current version/updates have made. It seems like some of my previous android friends are saying, we wont get these updated version of the operating system and all the possible apps for them unless we spend another $400.00-$500.00 on a newer android phone that already has the OS installed on it or trying to root our current phones. Ive thought plenty of times of rooting my phone but dont want to possibly destroy a $500.00 phone and be stuck with nothing but a paperweight just because google and t-mobile carriers cant get their stuff together. Before Google moves towards annual updates, all previous androids running 1.5 or 1.6 need to recieve the newest 2.2 OS so that we are all kept equal and dont mind waiting the 6 months to a year for the next update. Im tired of going looking for apps in the market and i can use them because im still on the original OS. FIX THIS GOOGLE, YOU ALREADY HAVE PLENTY OF UNSATIFIED ANDROID USERS. THIS NEEDS TO BE ADDRESSED ASAP TO MAKE ALL OF YOUR ANDROID USERS HAPPY. I DONT MIND WAITING ON AN UPDATE IF IM RUNNING THE LATEST VERSION OF MY OS. PLEASE FIX THIS PROBLEM. I WANT THE ANDROID OS 2.2 THAT COMES STOCK ON ALL THE NEWEST PHONE. I MEAN REALLY, ITS COMING ON THE NEW MYTOUCH SLIDE BUT I CANT GET IT ONE MY MYTOUCH 3G AT ALL. THATS CRAZY. GOOGLE AND T-MOBILE NEED TO GET TOGETHER AND FIX THIS SO EVERYONE IS A HAPPY ANDROID USER AND CONTINUES TO USE GOOGLE AND NOT LEAVE T-MOBILE FOR ANOTHER PHONE PROVIDER.

  • http://Website Fahd Bhatti

    In my opinion the average update cycle should be no more than 6 months long. Because apple only completely change the iOS after a year but during that year they release a lot of minor updates. So, in my opinion to wait whole year for the update is a very long time but reducing the time to 6 months would be a good idea. The other option is that if they start following what apple is doing is also the right way, complete OS in a year with minor updates during the year in quarterly cycles.

  • http://Website Soloren

    I think Android still has a long way to go. I had an iPhone and I switched to an Android. I find it quite disappointing for now.

    What’s great is:
    - the openess, customization, the potential all over…

    What’s not so great:
    - The battery hardly last a day, UI is all over the place, Operators added crap apps that can’t be removed on mine, for developers the screen sizes varies, too much is left to the handset manufacturer which don’t do as good of a job as Apple. etc…

    So I am actually disappointed that releases are slowing down. Also they are barely catching up with the iPhone and they are slowing down??? I thought Google wanted to be at the forefront of innovation!