It is perhaps this sort of comment that leads analyst Brian Blau, with Gartner, to conclude that Google "must not care" about the fragmentation issue.
To be fair, Google faces a difficult problem. Both OEMs and mobile carriers customize the open-source OS, so that there are differences even for users running the same version of Android on different devices or with different providers. And neither OEMs nor carriers are in the business of providing OS updates, and have incentives for encouraging users to buy new phones instead.
The PDK won't change any of those factors. Its goal is simply to get phones running a particular version of the OS into consumers' hands faster, Google representatives told developers.
Google does have at least one way to fix the problem, according to analyst Ezra Gottheil, with Technology Business Research. The company could make a version of Android that is more uniform across devices and carriers, he said. Users could then update the OS directly from Google, much as iOS users do from Apple. But the OEMs and carriers want to be able to customize the OS to make it look like their product, according to Gottheil.
Google must therefore navigate between angering its hardware partners and its developers.
"They're trying to knock the edges of the problem," Gottheil said.
But fragmentation contributes to the higher cost of making apps for Android, according to mobile analytics firm Flurry, meaning Google risks chasing its developers away if it does too little to address the problem.
"The danger there would be they get a reputation of having no decent apps, but I think they're kind of far away from that," Gottheil said.
Indeed, developers at the conference seemed to accept that fragmentation can be an occupational hazard of open-source development. They also noted that Google offers tools in its support libraries to help them work with the multiple versions of Android.
"Android is really the first time that we have an operating system that can run on this widely different array of devices, and more or less you can write the same code that's really easily portable, which has been the dream. I really think that the pros of that far outweigh a more integrated system where you only have a few different pieces of hardware," said Zack Juhasz, of a yet-to-launch startup, Tenkiv.
Sign up for CIO Asia eNewsletters.