There seems to be some movement, but no solution, to the problem of Android fragmentation. However, one blogger-who also happens to be the CEO of mobile open source company Funambol-doesn't think fragmentation is that big a deal. Indeed, he suggests that it is healthy and necessary.
Android fragmentation is caused by the release of new versions in such quick succession that any given application doesn't work on all versions. At this point, there are versions 1.5, 1.6, 2.0 and 2.1 of Android, with at least two more, Froyo and Gingerbread, to come. Some observers fear that Android will become an electronic Tower of Babel.
Research released this week by IMS says that fragmentation hurts developers, service providers and handset vendors. Developers have to pick and choose between versions of Android and end up aiming at at only a small piece of the Android market. Android, IMS points out, is smaller than many proprietary OS universes to start with, so subdividing it further is not a good thing. Costs to network operators and vendors rise as more versions of the OS are supported.
Google is confronting the issue, at least according to Engadget. The site reported at the end of March that Android will fight fragmentation by enabling a user's platform to stay current through downloads that are independent of the carrier:
[Android] will start by decoupling many of Android's standard applications and components from the platform's core and making them downloadable and updatable through the Market, much as they've already done with maps.
There may be a problem with this strategy, however. IMS research analyst Chris Schreck suggests that the Android Marketplace might not be in a position to help the user, at least in some cases, because it won't have the update:
The Apache software license, which Android uses, does not require licensees to contribute modifications of the Android platform back to Google. While this license characteristic serves to encourage licensees to innovate on the platform (without fear of having to share IP), it also allows licensees to alter the platform in isolation, a recipe for fragmentation.
Engadget and IMS assume that fragmentation is a problem that threatens Android's long-term prospects. Fabrizio Capobianco, the Funambol CEO, has a different take, and a lot of people in the open source community no doubt hope he's right. He suggests that the danger is overblown. To him, there are two types of "forks" in which innovators moved significantly from the core OS: Those from outside of the main sponsor of the software -- Google in this case -- and those from within. The first half of the post deals with the outside moves, which Capobianco thinks can be vehicles of innovation. The one example -- from China Mobile -- will fade away, he predicts.
The second is the fragmentation emerging from Google. This is what IMS and Engadget are discussing. Capobianco suggests that Google has not moved too fast. Early versions of a new operating system, he writes, must evolve quickly to reach equality with established OSes. Google realized this, and therefore minted four versions in a short time. He believes that much of that early spade work is done. The next versions of Android -- Froyo and Gingerbread -- will have far fewer core changes and will be highly backward compatible with existing versions.