Showing results for 
Search instead for 
Did you mean: 

Android - to be or not to be?


That Android has got a lot of publicity and public endorsement in the US is clear to all of us but the potential impact that Android can have is probably far beyond what you have thought. I am therefore posting some of my 'insights' in the subject.

When Andy Rubin left Sidekick and started Android (and acquired by Google), he and his team did a few things amazingly correct.  The first thing they want to do was to embedded Google services as much as possible; secondly, providing a standardized easy to develop developer experience that is standardized across devices; and thirdly, provide a nice and easy UI. 

What they did not do was to provide a native SDK, nor any tools to easily modify and change the UI, nor any backwards compatibility between 1.0, 1.5, 1.6 and 2.0 - as for developers, it did not matter - they had the Dalvik SDK (which is stable). But the Android team released most of their code under an open source license (MIT) and that opened up for an unrivaled deployment from chipset, ODMs, OEMs and even operators.

And to be fair, Android is now the one and only open source package that provides everything from UI, kernel, sandbox and a bunch of applications that constitutes the basics of a smartphone.  So what has this lead to?! Well, I can't tell you everything as some is information given in customer confidentiality, but some I can share:

- Mobile phones (Google-certified): HTC, Motorola, LGE, Samsung (though Samsung maybe not be totally backing it) and SonyEricsson (HTC);

- Operator endorsement from Verizon, T-Mobile, 3 and more on the way. What the OEMs and ODMs keep hearing is the demand for Android-based devices;

- Mobile phones (non-Google): China Mobile's oPhone (w Dopod, LGE, Dell and other producing phone); a number of shanzhai phones; virtually all ODMs (to mention a few: Yuhuatel, Pegatron, Quanta, etc.);

- Smartbooks:  not supported by Google, but the ODMs and some OEMs are in full development of Android-based Netbooks, driven by operator demand. I have seen demos based on Qualcomm Snapdragon and Freescale iMX.51 done by ODMs in Taiwan;

- STB: I have been shown a few various reference designs by customer who has been using Android;

- In-Car Infotainment systems: One of the biggest Chinese car manufacturers is developing an Android-based solution (want phone capabilities); and

- Chipsets: nearly all chipset manufacturers have a demo with Android (Freescale, Qualcomm, TI, Marvell, etc. - even MIPS now).

At the ARM symposium that I attended in Taipei and Hsinchu last week, ARM itself had a number of speeches to show how to innovate and develop with ARM and Android. Apart from our Qt-demo and Movial's, a few were Windows Mobile and the rest was just Android.

In Japan, the in-official Android community is huge - more than 2,000 open source developers (and growing), compared with some 50-100 in KDE.

I was speaking at Web Wednesday in Beijing the same week in order to share my experience with open source business models.  And most of the developers there were totally ignoring Nokia but a few them was really in favor of Android.  The perception of Android's strengthen a position, is in my opinion, greatly over-rated, as there are clear signs already of fragmentation (and thus developers will not have such an easy life) but perception rules!

What drives this? For any ODM or OEM who wants to e.g. use an ARM-based chipset to drive down cost and to build a device quickly (partly being able to show a demo to a customer to be able to sell in the design), there are a number of things to look out for:

- free open-source solution means no budget spending and lengthy contract negotiations with supplier.  I.e. easy to start the project and low/minimum risk

- avoid plumming that do not add any value.  In the plumming, we have kernel porting, driver development, media-framework (some value in optimizing and performance), back-end features (SQL, UI toolkits, etc.), basic applications (such email, browser, dialer) and a sandbox model for installing 3rd party applications

- 3rd party application development model for developing and deploying applications on devices but also making it easy for the OEM/ODM themselves to add/develop pre-installed applications.  The Java language makes it fairly easy for anyone to develop apps given that large existing Java developer community.

So what could go wrong? What could stop the Android and make him rust.  There are hurdles and deploying on non-ARM chipsets are one; memory consumption can make the devices too expensive (beyond mobile); difficult to develop and customize the UI (important once the screen is not standard screen) and of course, consumer uptake.  Still there are no evidence of real consumer uptake and the next year stand to prove the case.

There are also some large developer houses that questions the viability of Android and ability for 3rd parties to make some serious money.  Gameloft recently announced cutbacks of Android apps development.  A quote: "We have significantly cut our investment in Android platform, just like ... many others," Gameloft finance director Alexandre de Rochefort said at an investor conference.

Android could also fragment itself to non-importance with its open-source model and non-backward compatibility in the framework and Native controllers.  It could become the preferred choice for building anything but would be limited to UI and some on-device preloaded applications and a browser and not much more beyond mobile phone.  Also Chromium OS may divert attention and given that ChromeOS is built upon Canonical's Linux distro, there are also chances that we can easier get Qt and CWRT running on those devices.  But I fear that Google in true non-open source spirit and its aim to rule the world will limit OEMs/ODMs if they want Google services preloaded on their devices

Alternatives to Android?  Moblin maybe, but that is x86-based and not open to non-Intel chipsets (GPL only otherwise) which really limits uptake

And where is Maemo?  Should that have been one of the options for any ODM?  Well, probably but it seems too complicated to build a device based on ODM.

SonyEricsson provides an alternative view on how to provide a developer experience:


And this is also well in-line with Nokia's push for the WebRuntimes that is on S60 devices and also in some of the S40 and also Maemo. Similar for PalmPre and this is also aligned a bit with Chromium OS. Webapps is the future?


Thanks for the insight David. A few things come to mind that lead me to believe that there will not be universal Android adoption for quite awhile, if ever ...

First if we look at current Mobile OS market share for SDK and trained professionals building apps and systems, we find Google Android comes in last place:

* Nokia Symbian

* RIM Blackberry

* Apple OS X for MACs, iPhones, tablets etc.

* Microsoft

* Linux

* Google Android

What would drive these communities to abandon what they've invested in to jump onto Android? In fact don't some of these communities view Android as "evil" (developers can offer pretty strong opinions ;-)



you highlight an important aspect.  Developers will normally go for the most deployed devices in the market and develop applications.  However, it seems that Nokia (whom I am working for) has managed to break that paradigm - particularly in the US.

If you ask what developers are developing for, you would probably see something like this:

1. iPhone/iPods - everyone

2. Android as the second platform for most

3. RIM or MS

A lot of the US-based developers will not even consider Nokia/Symbian as its developer offering currently is flawed.  There are a lot of works in progress from both Nokia and Symbian Foundation to fix that of course and I believe it will improve.

But it is also regional. In China, S60v3 is the first platform that developers will develop for and then maybe even S60v2 before considering MTK and/or Java.

But from a device manufacturer, the promises of Android makes statement where Android now is in the last place obsolete.  However, there is a very obvious concern that due to the openness of Android and that Google is not able to enforce compatibility between platforms that the Android developer market will be highly fragmented.  Wired posted an interesting article sometime ago.

And that can give Nokia an edge in consolidating the developer market around 100M+ devices yearly.


Android ... to be or not to be? That is the question ... market share perspective is much smaller than iPhone after first year, but is trending up ... but I see Nokia as so powerful in this space it's hard for me to get head around idea that Android could be dominant. Do you see it becoming "The" dominant Mobile device OS? Would Nokia let that happen? ;-)

The blurring of Mobile Internet application use cases between smartphones, MIDs, laptops etc. makes this all the more fascinating to me.

Google offering Android on the one hand, then building their own Google Phone (will all Android device manufacturers truly get the same just-in-time access to everything Android has to offer? Will Google Phone have leading edge to updates?), while supplying Chrome for "bigger?" devices ... doesn't sound like something that exudes great confidence for developers to race after. But then again, as you state, "perception rules"

And what of App Stores? Are they destined to remain client based? Or will there be "emulators" to run your previously purchased apps on a new device?

Finally, what of trending to "The Cloud" ... makes me wonder if the drive to save even more battery life and further reduce device costs will influence the "smart" phones toward a destiny of dumb TTY terminals linked into intelligent clouds ;-)

Whatever the final destination, its great to have insight from people like you in our community, driving and deliberating where to take this ... I bet part of our "discusion" should go into Discusion ... and something tells me that regardless of which OS "wins", having a reliable and ubiquitous network connection in between these Mobile Internet devices will remain pretty important


personally, I think there will always be a number of platforms, a number of choices - as people are different and have different needs.

If not, we would all be on Microsoft PCs by now and we are not. Nokia will of course not sit quiet and let Google take over and the alliance lately with Microsoft can prove to be powerful.

The Cloud fits some people - but the unreliability of network access and tied with the cost of bandwidth (there is a cost for that) will push for more than cloud-based services/computing.  There will always be low-end vs high-end, simple vs full-features, faster vs more complexed, etc.

Who wants the dumb terminal back?


Very interesting thread.  I just finished writing a report on the "Android Invasion" that has occured here in the U.S., starting just before CTIA Wireless in October and continuing non-stop ever since. A lot of very good technical points are raised in this discussion, but I think my perspective is somewhat different since my point-of-view is mostly focused on the business implications.

Here are just a few of my opinions/observations:

  1. Very simply - Google's goal with Android in smartphones is to do on the mobile internet what they have done already to the desktop. Individually, the basic components (browser, search engine, email, etc.) built on top of the OS are secondary to the end result that is achieved by their integration: a synergistic platform that facilitates commerce. (See the AdMob purchase). On the desktop it was a matter of integrating (mostly) into Windows. Not so on the mobile handset.. hence the need for Android.
  2. It's a numbers game. The U.S. wireless operators that are currently supporting Android (Verizon, Sprint and T-Mobile) account for more than 170M cell phone users.  That is more than 2X the total available market addressed by iPhone-dependent AT&T. I'm not sure what David means by fragmentation, but apps developers can't help but be drawn to redirecting their efforts to that larger market where one app hits more phones and more target users for the same amount of effort.  Verizon's $100M DROID campaign is backing them up, along with all the other carrier's efforts.  Apple is very profitable, but it is repeating the same closed-platform philosophy that it executed on the desktop... and we know how that worked out.
  3. When I see open source I read ecosystem. Nokia may have a strong Symbian ecosystem in Europe, but it has near-zero presence in the U.S., and make no mistake.. the U.S. drives the smartphone business.  The ecosystem of the Open Handset Alliance spans semiconductors, to design tools, to application software, to CE manufacturers, to service providers. If that is the result of being "open" so be it. The bottom line is that this represents a very formidable ecosystem that Apple's closed system can't match, and neither can Nokia.  Nokia should integrate Android into their Qt porting framework, or risk losing any chance of gaining market share in the U.S.
  4. Android will be in set-top boxes and other CE devices, as the convergence of the internet and digital media penetrates further into the digital living room.
  5. Don't forget Google's investment in Clearwire. Look for a WiMax Android smartphone, if not 2010, definitely in 2011.

-Mike Demler


I saw Apple's "closed" offer more as a gateway for the OS X developer community to easily port apps into the rapidly expanding Mobile Internet community. App store volume suggests they got that part right .. iPhone user interface + application combo offered the "Wow" factor to draw a crowd

We just busted through Black Friday in the States ... it'll be interesting to see how the holidays (and Verizon $100M ad campaign!) helps drive Android sales ...

Apple has always had a crowd willing to pay a premium for a small part of the overall market ... Google draws a crowd that says Internet is free ... it'll be interesting to see what "revenue sharing" models work out best between subscribers, operators, and the droid eco-system


@Mike, with fragmentation I mean that the single app no longer can run on any Android device as OEMs have started to fork Android of various reason and it is difficult for Google's Android to enforce full compatibility. E.g. GPS and Camera integration are handware dependent and in order to make it work, the APIs may change.

It is quite interesting how, due to Apple, US regain the edge in the mobile market after losing in the early 90s to Europe. In the early 00s, US mobile market was a joke but it has changed. But 170M cell phone users is not that much. China adds 80-100M cell phone users yearly. India as well (though 80% is low-end phones).

The chinese operators are all forking Android to O-phone (CMCC), U-Phone (Unicom) and C-phone (Telecom) - all on Android, all on their own SDKs and modifications. Here one Android app will for sure not run on the other platforms. And Nokia will probably the uniting factor here.

But the idea of Qt supporting Android is not new and I really think that we need to take Qt to Android (remember, I work for Nokia) but as good as it sounds, it is not as straightforward. Ideas more than happily received.


Hi David,

I wonder how severe the issue of fragmentation of the Android OS is.  I can see backward compatibility being an issue, but is the basic foundation of the OS revs that Google releases even changeable by developers?  The developer is responsible for testing hardware dependencies.  There are numerous emulators available, which is a topic I cover in my recently published report on "The Android Invasion".

A bad (i.e. incompatible) app is a bad app, and they will exist regardless of the platform, iPhone or Android.  As smartphones are just portable computers, it follows that some of the same problems we see with hardware dependencies on the desktop will occur. However, I don't see a problem with developing apps that are targeted at specific hardware as long as that is specified, nor do I see a problem with creating regional variations. This sort of "fragmentation" just creates more opportunities.

I realize that 170M subscribers is a small fraction of the 2B worldwide, but of course I was directing that point at U.S. developers - just to provide some perspective relating to the "gold rush" to iPhone app development. Android opportunities will dwarf iPhone because of the "fragmentation", or perhaps more accurately - "diffusion of innovation", just as has occured with the numerous variations of Windows OS configurations that dominate the PC market.

I am curious about your comment that "Nokia will probably the uniting factor".  Can you say more about that?

Great discussion..



it is difficult now to say how severe the fragmentation will be. It is more likely to happen in the open-source device ODM/OEMs (STB to MID to Netbooks to whatever) rather than the mobile phones. But even in the mobile phones, the OEMs are enhancing Android and when they do, they add features that requires specific SDKs, etc. etc. and immediately, the developers that targets the 170M potential Android users ends up to have to target 5-10 different phones with 100K-2/3M units.

At Nokia we have learnt that lesson the hard way.  We managed to break the paradigm that developers target the most deployed platform.  And by all numbers, Nokia has deployed more smartphones combined than all others but with a number of platforms where binary compatibility have been broken repeatedly.

wrt my comment about Nokia being the uniting factor, I was referring more to China as the market, where if the three operators post their own non-compatible Android SDKs, Nokia with our dominating market share can get the developers to develop for Qt/Webruntime/Java on Nokia devices and be sure that it will be able to run on all operator's handsets (if the user has a Nokia device).

Once Qt is one of the preferred developer platforms, the developers can also port (or rather deploy) their apps on non-Nokia devices. E.g. Windows Mobile phones, Samsung Symbian phones, Asus Netbooks (Linux/Windows), TV/STB running Qt. and so on.

Content for Community-Ad

This widget could not be displayed.