Back at CES in January Google Executive Chairman Eric Schmidt took part in CNET's Next Big Thing Panel, where he famously said this when discussing Android fragmentation:
"You have to be careful with that word," he cautioned. He actually prefers to use the term "differentiation" to describe the situation, as his definition of fragmentation is when certain apps are able to run on one particular flavor of Android and not other. "What people really care about is if there's an interoperable ecosystem of apps."
I'll admit that I'm less alarmed about Android fragmentation than it seems because it doesn't touch my life very often. The apps that I want to use are, with only a few exceptions available for my phone and my tablets, and they work quite well. Even though I don't endure fragmentation ugliness very often, I've never bought the notion that fragmentation isn't an issue for Android. It's just not as much of an issue as some want you to believe that it is.
There are too many devices from too many manufacturers to not see it as a problem at some level. Just in the USA there are multiple phones from HTC, Motorola, Samsung, ZTE, Huawei, LG and others that are available on every carrier. What's worse, there are multiple variants of some of these devices based on carrier. The HTC One X that might land on AT&T isn't physically the same HTC One X that you'll be able to buy in Canada because the Tegra 3 quad core processor isn't compatible with LTE.
Most of these types of fragmentation issues are easily ignorable from the end user point of view. If you're an average phone buyer on the AT&T network and you buy the HTC One X, the dual core Snapdragon processor isn't going to be front and center on your mind. You'll care about how the phone looks, how responsive it is when you pick it up and how the phone feels in your hand.
Not really much of an issue. Or is it?
Android app developer OpenSignalMaps collects cellular and WiFi signal data to map networks and present information to assist people with finding the best network in their area. Simple concept, simple app.
OSM has been collecting data from 195 countries about the devices that download their app. They logged device, manufacturer and OS versions for 6 months, and the results were stunning. There were 3,997 different devices from 599 distinct manufacturers.
Folks jumped all over this report as a sign that fragmentation is a huge Android problem, but it doesn't seem that many people actually read the entire blog post from OpenSignalMaps. At the very end of their post, OSM says that:
One of the joys of developing for Android is you have no idea who'll end up using your app.
That sounds like a bunch of hippie crap to me, but that's not the point that many seem to be missing. In the description for the models section of their report, OSM says this (emphasis mine):
One complication is that custom ROMs can overwrite the android.build.MODEL variable that we use for the device model. This partly explains why a staggering 1363 device models appear only once in our database.
In the rush to proclaim that there are over 3,900 different Android devices, all three stories that I read about this report failed to mention the fact that 1,363 of those phones were unique models. Due to rooting and custom ROM usage just over 1/3 of the devices in this report identified as unique devices when in fact they were not. The maximum number of unique devices in this report is actually 2,634, and we don't know what other variables are at play that might lower that number even further.
Look, fragmentation or differentiation exists in the Android world, that's just a fact. Want to know another fact? Unless you're buying an off-brand, no-name Chinese device running a hacked up version of Android you'll have very few issues with all of the differentiation.
Sure, you'll have some app issues from time to time, but that would happen if there were only 12 models of Android devices. Fragmentation isn't your enemy, as bad as it may seem on the surface, fragmentation provides the choice that you have in devices and carriers. Differentiation is actually your friend.
Source & Image: OpenSignalMaps