ie8 fix
Click Here

Googling Google

Christopher Dawson

Android "fragmentation", and why it doesn't matter

By | June 8, 2010, 8:11pm PDT

Summary: This is a post about my personal opinion — feel free to disagree. Post your opinion in the Talk Back. There is a relatively small subset of Android users who care about “fragmentation”, and an even smaller subset of those users who are knowingly victims of it. The fact is that when you have [...]

This is a post about my personal opinion — feel free to disagree. Post your opinion in the Talk Back.

There is a relatively small subset of Android users who care about “fragmentation”, and an even smaller subset of those users who are knowingly victims of it. The fact is that when you have a phone, you use it for what it does. If you want a phone that does something else, you either wait until your contract expires, or buy one outright.

Even if you only have Android 1.5, the majority of developers are still developing for your platform. It’s silly to target newer versions if your application doesn’t require it. That means that the majority of applications you can find in the entire marketplace will work on your phone. In fact, you probably won’t even know what you’re missing since the marketplace only shows applications that will work on your device.

So does “fragmentation” really matter? Or is it just a buzz word that people are latching onto for some reason? Let’s hear it in the Talk Back.

Kick off your day with ZDNet's daily e-mail newsletter. It's the freshest tech news and opinion, served hot. Get it.

Topics

Garett Rogers has always had a deep interest in computers and the Internet, which led him to a degree in Computer Information Systems. He is currently employed as a programmer for iQmetrix.

Disclosure

Garett Rogers

Garett Rogers is employed as a programmer for iQmetrix, which specializes in retail management software for the wireless industry. He has no other formal associations with any software or hardware companies.

Biography

Garett Rogers

Garett Rogers has always had a deep interest in computers and the Internet, which led him to a degree in Computer Information Systems. He is currently employed as a programmer for iQmetrix, which specializes in retail management software designed specifically for the cellular and electronics industry.

Garett's journey into Google started with his employer asking him to "get a better rank on Google." Diving into search engine optimization sparked his curiosity for how things work and led him to create a blog dedicated to what interests him most--Google.

27
Comments

Join the conversation!

Just In

Consumers don't care
leosg 14th Jun 2010
I would consider myself a technologist. So, I may not be a good example.

Let's take my wife as an example. She wants a phone that does the basic things like make a call and keep the call going, no dropping etc. Simple browsing and as small a footprint as she prefers to keep in her purse.

Whereas, for myself, I'd prefer a phone that has large screen for work, then a larger screen when I do my workout everyday (it doubles up as a video player) and then as small a screen for when I go to the beach or go for games. The frustration I have today is i use the iPhone 3GS for work, squint through the iPhone screen for workout and switch to a HTC Touch Dual when I go for my games ... now, if only i can get my contacts, messages available on each one ....
I agree and disagree.

I agree on your point that it may not matter that much to most consumers. However it certainly does matter for developers. I am planning development of an Android app right now. As a developer, I want to make use of the newest features of 2.2. In fact my app may require those features to function or become a lot more compelling with those features than it would be if I were to code for 1.5. But I also want my app to be available to the widest number of potential consumers. If a significant portion of the Android consumer base is on an earlier version like 1.5, that version becomes the lowest common denominator that I must develop for if I want the best chances for success.

I also disagree on your suggestion that only a small subset of users are victims of fragmentation (as well as the other columnists who have suggested that fragmentation is a myth). Take a look at Google's Official Platform Version Stats at:

http://developer.android.com/resources/dashboard/platform-versions.html

Based on those official numbers from Google more than half of Android users are still on a 1.x version of the OS - 54.5% as of June 1 2010 to be specific. So as lovely as 2.2 may be, I'm not so stupid as to cut out half of my potential consumer audience by coding for it only. It's just a bad business decision all around.

Also, let me propose that fragmentation may be an unavoidable side-effect of having multiple hardware vendors. I'm not sure that it will ever go away. So what I think would be best is if Google (and platform advocates like yourself) were to accept this as the nature of their beast and instead of trying to deny its existence or belittle its importance to various segments of the ecosystem, rather work on strategies to live with it and work through it effectively.

Fragmentation is real and it will limit innovation on the platform if strategies are not developed to live with it. I want this platform to be as successful as any of you, but unless we deal head on with the issues we come across, it will fail and we will all be worse off for it. Simple as that.
0 Votes
+ -
Otherwise he'd understand how pissed off it is having to maintain several code bases for one product.
0 Votes
+ -
RE: Android
brister@... 9th Jun 2010
Real developers don't have multiple code bases: they have just one that gets built different ways. Where I work all our products run on about 10 different systems. One code base.
0 Votes
+ -
RE: Android
BIGELLOW 9th Jun 2010
@aslam_nathoo So, I assume you first developed your application for Blackberry devices, right? Then, for the iPhone. And, finally, for Android?

Currently, this is how the market share is arranged... so if your only goal is to reach the most hands as possible, this is what you'd be doing.

If, instead, your main concern is creating a cool app that relies on particular functionality to accomplish, then you wouldn't worry about market share and percentages. You'd develop your app for the API version that makes sense... then, you'd reach everyone who has that API level on their phone. Eventually, the others will move over (either by getting upgrades on their phones, or by replacing their phones with newer ones) and you will see your audience grow and grow and grow.

The only ones who have an issue with fragmentation on Android are the ones who are just paying too much attention to numbers.
0 Votes
+ -
Hoping they upgrade is not a business plan
aslam_nathoo 9th Jun 2010
@BIGELLOW: Thanks for the response.

So to give you some more context, the app I am planning will require a touchscreen device with a decent sized screen, strong video and audio playback capabilities as well as mapping/GPS functions. Those are the minimum requirements to have the app be useful.

My goal is to launch the apps simultaneously on multiple platforms. And while I have investors, our pockets are not super deep. So we had to look at the numbers and see where the best place was to spend our limited dollars. Here's what I discovered when I looked at the numbers:

You're right on the mark when it comes to the total numbers of "smartphone" devices when you say that BB is first, then iPhone and then Android. However since most BB devices have a small screen, limited video/audio functionality and no GPS, they are not suitable for our app. Therefore when I took out all the BB devices that don't meet the minimum specs and looked at the numbers again, I discovered that for our potential market iPhone is first Android second and BB a distant third. Since we can't afford to build for all 3, we are planning to build for iPhone and Android and launch them simultaneously.

As well, keep in mind that our app will not be something that 100% of people will want. Our business plan economics are modelled on us being able to get 10% of the potential customers on each platform to buy our app. If we can get that, we will have enough revenue to keep improving the app, provide a really high quality ongoing product/service to our customers, make a comfortable living and a small profit to re-invest into our business so as to make more apps like the first.


So now lets go back to your suggestions and focus on just our Android build strategy (since thats what this post is about). If I understand you correctly, you are proposing that I build for Android 2.2 even though it has a smaller market share right now but because it is the superior system. And then I am to hope and pray that enough handset makers develop the upgrade, that enough of their customers seek out that upgrade, and that enough of those people buy my app to make a living? That's a big gamble for a small app developer to take and probably not a smart one unless one has deep pockets.

Also going with that strategy would severely change the economics of my business plan. Since I can't predict the future, I can only build my business plan economics based on what the reality on the ground is today. And if I were to go with 2.2, as of today they it has less tan half of the total Android installed base. Therefore my 10% number would have to jump to 20% to get the same revenue. That's a big change to the economics and makes the Android build a lot more risky.

Also I'm not sure if you've ever had to deal with investors or financiers, but if I were to pitch that strategy in my business plan, I would be laughed out of their offices.


Having said that, I could genuinely see your plan working if I were developing a landmark/destination app. Something so high profile and juicy that it would drive people to switch or seek out the necessary upgrade just to get it. Alas my app is not that landmark nor that high profile. Indeed the vast majority of apps out there are not destination or high profile apps. And that's my point. The economics for the vast majority of app developers mean that if they target only the latest and greatest, they will severely limit their potential customer base and will be required to get an increasingly higher percentage of that customer base to buy before they will have a viable business.


Keep in mind that I'm not saying 2.2 isn't better and that users shouldn't move to it. It clearly is and users should get it (whenever their handset maker chooses to make it available to them). I'm not fighting a philosophical war of which one is better. This is a business. I'm trying to provide a valuable service to customers who would want it, make a living, and give a return to my investors and your strategy as well as the fragmentation in the Android market severely constrains my ability to do so.
@aslam_nathoo points out that consumers really don't care--as long as it works. However, at least for the moment it is practically impossible to upgrade any of the AN 1.x versions to 2.1 or even 2.2. As such, a developer has to either target a specific version, or write several different apps to take advantage of each version's separate features. This immediately makes the work 4x to 5x harder than if they can concentrate on a single distro. What could upset the consumer even more is when they find out that a competing device using, say, the first version of its OS can and has been upgraded almost seamlessly to the second, third and soon to be fourth version of the OS.

So yes, fragmentation is a real problem for both groups, but the consumer is less likely to notice it compared to the developer.
android to begin with, if they even know it. But android app devs with boxes full of incompatible handsets that they have dozens of hacks all over their code to try to support care.

What a lame red herring post!
0 Votes
+ -
RE: Android
aslam_nathoo 9th Jun 2010
@Johnny Vegas: On the point of "android app devs with boxes full of incompatible handsets that they have dozens of hacks all over their code to try to support"... I agree wholeheartedly!! All the small differences in the various device hardware (such as screen size and density) makes life so difficult for us. But that's not the more important part. I don't mind difficult problems. I can solve them. However the time it takes to solve them really makes it much more expensive for me to develop for Android. Those increased costs change the economics of my business significantly and that's a big problem.
0 Votes
+ -
It matters most to iPhone owners
NonZealot 8th Jun 2010
For some reason, it is essential to them that they try to convince everyone that all Android phones suck. Insecurity perhaps?
0 Votes
+ -
RE: Android
aslam_nathoo Updated - 9th Jun 2010
@NonZealot: First of all this is not about which is better or worse. This is about honestly recognizing and solving the problems that exist in the Android ecosystem.

Secondly, I've been following this thread very carefully and not a single post I've read that was made before yours says that the iPhone is better. All one person said was that that it has a higher market share as of today. That is a plain and simple, indisputable fact.

Your post looks like you are the one who is insecure about how good Android is. I have no doubt about how good it is. I also have no doubt about how good the iPhone is. They are both great in my opinion. This is not a beauty contest. This is about making devices that are functional and building a healthy ecosystem around them.

Edit: @shryko: I'm not sure how much I appreciate you telling me how I should use my time. Nonetheless, you are correct about NonZealot's comment being troll-esque. Since I am not a regular commenter on these boards, I didn't recognize him as a known troll. I will be more wary of engaging that user in the future. Thanks for the warning.
0 Votes
+ -
know your trolls...
shryko 9th Jun 2010
@aslam_nathoo: for all you talk about how much effort everything is to do the multiple versions, and how much time it takes...

...you sure have a lot of time to reply to flamebait trolling by a known troll. And even if he's not a known troll, that post is so clearly anti-Apple, it could be considered trolling in and of itself.
0 Votes
+ -
solution
kikl 9th Jun 2010
The fragmentation occurs because hardware vendors include some proprietary software on top of android. They are too lazy or too slow to port their software to the newest version of android.
Solution:
Google should make the OS available in their appstore and possibly provide information about compatibility with certain devices. So people can choose. Do you prefer the older version of android with htc sense or is the updated version of googles android without sense more compelling?

This will force hardware vendors to update their software more frequently or even make it open source. Most people will upgrade if it is easily done. Android already backs up all your settings and files in the cloud. So this should be fairly easy to accomplish.

Regards

Kikl
0 Votes
+ -
RE: Android
aslam_nathoo 9th Jun 2010
@kikl: interesting idea and one that I had this morning as well. The problem is that all of the different devices have slightly different specs. Google would then be in the job of coding/building installers for each of those specific hardware specs and they do not have access to the necessary hardware design schematics to do so. Plus it's a lot of work and what revenue are they getting from it. Google is afterall a business and they have shareholders to answer to. Also it would mean that they would have so many branches in their Android codebase to support all the various device installers which would make future development incredibly difficult and slow.

The other way to approach this would be to have Google dictate a very strict hardware spec. Then they can build one installer that runs across all devices. Hardware vendors or carriers who want to add some Value Added software (e.g. Sense UI) could then do so on top of the universal OS build. Of course that scenario is not perfect either because we would be reliant on users first installing the universal OS upgrade and then going out and finding the corresponding Value Added software upgrade and installing that. It's a lot of work to ask a non-technical customer to do and I suspect a very small percentage would do it. I know my Dad probably wouldn't. Also another problem with the strict hardware spec idea is that it limits innovation on the platform. The sacrifice comes with the benefit of ease of use/upgrade, but it is a sacrifice nonetheless.

So neither solution is perfect but dialog like this is healthy.
0 Votes
+ -
RE: Android
Trufagus 9th Jun 2010
It matters for support.

For the consumer who asks their friends or family questions, or tries to find answers on the web, it is easier if everyone is using a similar version.

Now imagine the problem for the carrier, manufacturer, or corporation who must provide support for these phones.
0 Votes
+ -
RE: Android
aslam_nathoo 9th Jun 2010
@Trufagus: agreed!
0 Votes
+ -
But it won't matter in the long run.

It matters in the short run because it is a brand new OS that is changing rapidly, and 1st and 2nd gen hardware cannot keep up with the new changes as well. So mistakes were made in the original planning, we now know that we need plenty of extra ROM memory for programs and operating system images. As long as developers (both hardware and software) learn from these early mistakes, fragmentation will not be a long term issue.
0 Votes
+ -
@Michael Kelly: ... in getting Android to a stable iteration that can better work across multiple platforms. Google is working on this, and An 2.2 is much closer to that goal; however, some analysts believe that once they reach that goal, they will release it as An 3.0 and probably have managed to get HTC and the other vendors to build to their minimum specs--just as Microsoft has had to do for the last 30 years.
0 Votes
+ -
Eventuality, not ignorable...
shryko 9th Jun 2010
@Michael Kelly: fragmentation is the long-run eventuality.

As iOS 4 comes out, there's now devices (still supported) that are running OS v2. (I know they're still supported, as my friend is using one)

Versions come out over time, and fragmentation occurs. It is inevitable, just... Android is doing it faster than anyone is really used to.
0 Votes
+ -
RE: Android
Zc456 10th Jun 2010
@shryko Of course, Apple does everything right.

Anyway, Garett makes a good point though. I've noticed that many applications for Android still support 1.5 or 1.6. Its only stuff like Flash 10.1 or the official Twitter app, for example, that require 2.1 and above due to the use of newer APIs.

As for device support, the manufacturer just needs to write the drivers. Otherwise Android doesn't have that much hardware limits. I've even played with an Intel fork. Which is actually what got me interested.

As for iOS, even if the iPod Touch and iPHone devices did still support 2.x, not saying they don't, many apps I've seen on the App Store require 3.x and above from what I'm aware of.

Sigh, sorry for the long comment.
0 Votes
+ -
Is it a problem? Sort of
mr1972 9th Jun 2010
Basically if you look at end users, they will buy the phone that does what they want. So they don't really need to worry about the fragmentation of the operating system as long as their favorite programs run on their phone.

for developers, yeah they might need to work on separate code basis but it all comes down to who you want to sell your app to.

It is kind of to bad that there wasn't a hardware abstraction layer that would allow for an application to run on different hardware. Most smartphones probably have enough resources to support a hardware abstraction layer. A hardware manufacturer could put out a sdk for their particular hardware that is compatible to the software standard.
0 Votes
+ -
RE: Android
CptMatt 9th Jun 2010
It matters for us consumers, too. I just found out the app for my local radio station won't work on my phone. My phone is version 1.5, the app requires 1.6.
0 Votes
+ -
Artificial fragmentation
rarsa 9th Jun 2010
It matters when the fragmentation is artificial.

When Vonage or Skype only provide clients for certain carriers when other Android phones are capable of running those same clients.
0 Votes
+ -
RE: Android
skudera@... 9th Jun 2010
Until there are a couple killer, must have 2.0-2.2 apps, that users demand or they purchase another device, the fragmentation will not decrease. Dell is bringing it's new line of Android devices to market with 1.6. That is like bringing out a new line of computers with Windows XP. I thought of waiting for Dell, but not with Android 1.6. I purchased a Motorola Droid, it fits my needs and has 2.1.
Killer apps will drive fragmentation down some when they appear, unfortunately that is one of the weaknesses of Android at the present, there are no real killer apps that have the market pull to influence hardware manufacturers.
0 Votes
+ -
RE: Android
jf79 Updated - 9th Jun 2010
The thing is, developers have to hold off on putting in the latest and greatest features in their apps, if they want as many people as possible to be able to use them. The fragmentation, in my opinion, holds back the market.
0 Votes
+ -
Tower of Babel OS
roger_tee 10th Jun 2010
The issue is in development and you can't just have "one code base with different compile targets" and solve the problem as one commenter suggested. It's not as simple as that because the differences in handsets exist both at OS level and at t handset specific sw layers above the Android OS.

I ghost wrote a blog post for a friend of mine in this topic here.

http://telecomdisruption.blogspot.com/2009/11/is-android-tower-of-babel-os-guaranteed.html

RT
0 Votes
+ -
Mixed reaction
hiraghm@... 12th Jun 2010
I, too, want to develop for the Android platform. Not only are there phones, but tablet devices already using Android 1.5.

However, I just purchased a Samsung Moment. As soon as it arrived, I opened the box to make sure it was there and okay, and rushed down to the Sprint store to have it upgraded to 2.1 and activated (everything I read said it came with android 1.5). There wasn't time for the tech to upgrade it, but when the clerk fired it up to activate it, it already had Android 2.1 installed.

So, I gather that some, if not most, Android 1.x phones will be migrated to 2.1, sometimes automatically.

However... I've discovered that my Moment cannot be upgraded to 2.2. No multi-touch, no live desktop, no built-in tethering.

I'm still trying to figure out how to query a device for its device capabilities, since asking what version of the OS isn't enough to guarantee that a given device will have a certain capability.

I don't see it as fragmentation, just the same issue Microsoft faces as it evolves Windows from 3.1, to 95/98/NT, to 200/XP, to Vista/7.
Heck, I had a heckuva time on my Linux machine because my video card was too old for the ATI drivers.

So while it's an issue, I don't see it as a make-or-break issue.
0 Votes
+ -
Consumers don't care
leosg 14th Jun 2010
I would consider myself a technologist. So, I may not be a good example.

Let's take my wife as an example. She wants a phone that does the basic things like make a call and keep the call going, no dropping etc. Simple browsing and as small a footprint as she prefers to keep in her purse.

Whereas, for myself, I'd prefer a phone that has large screen for work, then a larger screen when I do my workout everyday (it doubles up as a video player) and then as small a screen for when I go to the beach or go for games. The frustration I have today is i use the iPhone 3GS for work, squint through the iPhone screen for workout and switch to a HTC Touch Dual when I go for my games ... now, if only i can get my contacts, messages available on each one ....

Join the conversation!

Formatting +
BB Codes - Note: HTML is not supported in forums
  • [b] Bold [/b]
  • [i] Italic [/i]
  • [u] Underline [/u]
  • [s] Strikethrough [/s]
  • [q] "Quote" [/q]
  • [ol][*] 1. Ordered List [/ol]
  • [ul][*] · Unordered List [/ul]
  • [pre] Preformat [/pre]
  • [quote] "Blockquote" [/quote]
ie8 fix

The best of ZDNet, delivered

ZDNet Newsletters

Get the best of ZDNet delivered straight to your inbox

Facebook Activity

White Papers, Webcasts, & Resources
ie8 fix
ie8 fix