Android’s Fragmentation Problem Not Going Away

Five Rules of Modern Application Development

The problem of Android fragmentation won’t easily be solved. The danger is that the best developers will get tired of dealing with it and wander off to other operating systems.

The challenge is inherently difficult because it involves millions of operating systems that already are out the door. Doing something different today is not going to alleviate the problem quickly. At best, new procedures will gradual affect and mediate the problem in the future. Until that happens, fragmentation will be a part of life.

Theallineed.com does a nice job of explaining why fragmentation is a big deal to developers. The bottom line is that it necessitates that applications run on all the Android platforms. That, apparently, is as big a pain in the neck as it sounds.

The fragmentation issue is a subtext on numbers released by Google on the progress being made by its KitKat operating system. The top line news, as reported by ZDNet, is that the OS has more than doubled its share, from 2.5 percent to 5.3 percent, in a single month.

That’s impressive. The number of different variations of Android also is impressive, in a negative way. A Google-supplied graphic posted with the story alludes to six different Android codenames. Three of these have different versions. All told, there are 12 flavors of Android (which is appropriate phrasing, since the company uses a dessert naming theme).

The story mentions two problems that such a system creates: The introduction of new versions really is only the start of a litany of steps by original equipment manufacturers (OEMs) and carriers. It is a long process, and the end result is staggered use on phones and general confusion. The other issue is that folks staying with the older versions are subject to malware and other vulnerabilities confronted by the newer versions.

Though reporters and commentators can offer opinions on the impact of fragmentation, the opinions of experts must take priority. They wrote the code, after all. I am a bit confused, however, by a post from mobile app developer Nick Bradbury. He begins by referencing a post from a couple of years ago in which he wrote “Android’s supposed ‘fragmentation problem’ was overblown.”

He backs off a bit, saying that his new job, working on Android apps for WordPress, has increased the number of fragmentation problems he sees. However, he “still maintain[s] that fragmentation is less of an issue than is commonly believed.”

Bradbury is writing from the developers’ perspective, not end users’. Still, he at one time says that accommodating multiple versions of the OS is no big deal and highlights a major issue:

I think the biggest problem is that unless your app is relatively new, you probably have to continue supporting Android 2.3. Making sure your app works on that ugly, buggy OS is a massive pain. Every Android developer I know will dance in the streets the day they can drop support for pre-ICS versions of Android.

I am not trying to pick on Bradbury, who gets paid to write code, not posts. It is clear he is saying that there are some issues but that the sky is not falling. Whether this holds true for end users, however, is another question.

Theallineed sums up a situation that is likely to remain the status quo going forward and that KitKat, which some thought would solve the issue, may not in fact solve:

With as many Android OS devices being produced as there are, it’s highly unlikely that manufacturers will want to give up the amount of control they currently have over the operating system. Furthermore, the very nature of the Android OS — a free and modifiable operating system — does not lend itself easily to consolidation. For the foreseeable future, Android OS developers will simply need to understand and work around fragmentation.

Of course, there is another option. Good developers may decide that the effort is not worth their time and instead focus on iOS, Windows Phone or the smaller emerging mobile OSes. That would be bad news for Google and Android users.