The opinions on Android fragmentation are, well, fragmented.
The basic idea is that the open source platform will be specialized to such an extent that versions will be incompatible. Instead of a single OS that has unique features and abilities based on the vendor - but can be used on any device - it's a fragmented OS that, in essence, is many different operating systems.
Google has been fighting this battle through all the versions of Android and hopes to take a major step with the release of Ice Cream Sandwich, which, among other things, will be able to accommodate both tablets and smartphones. While that doesn't eliminate the danger of fragmentation, it's a big step in the right direction.
The conclusion by the firm is that the problem isn't all it's cracked up to be:
For both screen size and resolution, Android developers have more to deal with than iOS developers, thanks to Apple's single handset form factor. However, with five options accounting for more than 90% of all Android app usage, the fragmentation is not particularly daunting.
Ed Hansberry, in an InformationWeek post that was published before the study was released, clearly doesn't think the fragmentation is innocuous. This is how the post starts:
If there's one complaint you hear about Google's Android platform, it is about fragmentation. It happens at the device level, the OS level, with the UI, and even with specific apps and services some carriers or manufacturers use. There are plusses and minuses to all of these, and it looks like Google has lost control, ceding the problems to the licensees.
He proceeds to run through the issue, and links to a piece by Charlie Kindel. That piece goes into considerable depth about the issue. The bottom line from his perspective is that fragmentation is real, multi-layered (user interface, device operating system, marketplace and service) and here to stay. It will not, however, doom the operating system.
At eWeek, Clint Boulton reports on Google CEO Eric Schmidt's treatment of the issue. It's a complex semantic exercise based on the nuances of the differences between "fragmentation" and "differentiation." It's a bit difficult to follow, but the two terms seem to be more or less synonymous. What's more important is whether what seems to be Schmidt's verbal gymnastics signifies some level of denial on Google's part that there is an issue with which to deal. Whether it is a major challenge - as Hansberry and Kindel suggest - or a minor one as portrayed in the Localytics report, the key is that it shouldn't be ignored.