Wednesday, January 26, 2011

Android on the Playbook? Cool but unlikely

I have used the BlackBerry Playbook and I have been honestly impressed by the OS. It is very quick. And the UI is beautiful. It looks like the combination of Palm WebOS (beautiful but slow) and Android (fast ;-)

However, it is missing one little thing: developers.

I know developers (or so I believe, being modest is not my thing ;-) and I see a hard path for RIM to attract them to the Playbook.

The SDK for the Playbook requires Air, which is cool and has a good amount of developers. However, they are not mobile developers. You still need to convince the Air developers to build for the Playbook. It is a totally new platform, because it is a tablet, not a PC. Today the people building for tablets are mobile developers, not PC developers (because they come to the iPad from iPhone and to Android tablets from Android phones). RIM is asking PC developers to move to mobile, which they eventually do. But it is an additional step, a significant one.

You would assume RIM will add a JVM to the device. At least, to allow BlackBerry developers to port their Java apps to the device. Not doing it will piss off the BlackBerry developer community, which is their captive audience. Microsoft has done that with Windows Phone 7, but I would not recommend it to RIM. Bad idea to piss off your community, albeit small. They are your core, those building enterprise apps.

Now, I read about a rumor today: RIM might be thinking about allowing Android apps to run on the Playbook. Technically, I believe it might be doable, because the JVM Android uses (Delvik) is open source and it runs on a Linux derivative (Playbook OS is QNX, a Unix derivative, close enough). However, there are a million internal calls that would have to be rewritten and most apps would not work at all. Android is even moving towards allowing native apps to run (with the NDK), so imagine what it would mean porting... A mess.

However, it makes business sense. Android has a ton of developers and the idea of easy porting to the Playbook must be appealing to RIM. Also, because QNX is the future BlackBerry OS, for all BlackBerries (my personal bet). Therefore, if they can make it really compatible to Android, they would solve the developers issue.

Easier said than done. Knowing OSS and virtualization, the probability of this attempt turning into a major failure is huge. Unless Google wants this, which might make sense to attack Apple iOS. With Google behind this move, maybe there is a chance this will succeed.

If you want my opinion, it is not going to happen. The Playbook will eventually have a JVM, allowing existing BlackBerry apps to run on it. This way, the transition from an old BlackBerry with BBOS to a new BlackBerry with QNX will be smooth. They will also provide Air for easy porting of PC apps and to add a cool factor. Lastly, they will allow web-based technologies (e.g. HTML5) on QNX, because it is the future (and they are already doing it with WebWorks). RIM will want their own developers, not someone else developers. Developers are the king in mobile. You do not want your king managed by someone else...

For all of you thinking fragmentation is a bad thing, I agree. It would be better to have just one platform to write to. However, I am afraid it is not going to happen, at least not soon.