There’s an interesting interview with James Gosling at eWeek.com.

Regarding Java on Android, James says…

“But everybody I’ve talked to who is building an Android phone or whatever, they’re all going in and they’re just hacking on it. And so all these Android phones are going to be incompatible”

The big question is in what ways will the phones be incompatible. If the OEMs are changing the OS so as to change the public APIs then this will indeed be unfortunate. However, the OEMs have an incentive not to do this. Doing so would make the OS potentially incompatible with 3rd party applications. They would have to ship their own SDK and only applications created with that SDK would be guaranteed to work on their phones. I am not sure an OEM would voluntarily go down this path with its implied limited app compatibility and difficultities with maintaining the code with the Android main branch going forward.

Instead, what the various OEMs are almost certainly doing is modifying the non-public APIs and native code. As a contrived example, an OEM might create a better camera API to allow greater control of the camera settings. This is OK provided developers don’t later need to rely on these new or modified facilities. If the OEM decides to publish these APIs then it becomes more difficult for 3rd party developers because there may need to be different application versions for different phones. It’s better if the OEM contributes their code to the Android source so that it becomes shared across all devices. This obviously requires some sort of compatibility testing and negotiation.

James goes on to say of Sun’s Java ME …

“We charge license fees is because we’ve got organizations of people that do compatibility testing and actual negotiating amongst the different handset makers so that things like GPS APIs look the same.”

However, we all know now that this didn’t really work. Despite compatibility testing (TCKs) there’s significant fragmentation across Java ME phones. While working on Java ME JVMs, I have personally used the TCKs and I can see that it’s impossible for tests such as these to account for every deviation from the expected functionality. Also, Sun’s cross party negotiations resulted in Java ME’s onerous permissions based security that, in my opinion, has limited Java ME’s success.

I also previously wondered about the future of Java (and Java ME) now that Oracle intends to acquire Sun. In second part of the article, James says…

“It’s pretty much impossible to say.”

James says he has ‘no data’ on this. As one of the commenters on the eWeek article points out, this is revealing. It tells us that the future of Java and Java ME hasn’t yet been thought through. I can only guess there are more important factors and reasons weighing in on the Oracle’s acquisition of Sun.