Kinoma And Marvell Attempt To Augment Android

On Monday, Marvell announced they were acquiring Kinoma. To the man on the street, that sounds more like the plot of a comic book and not the merging of some pretty serious players in the mobile industry, which is what it actually is. And they’ve got some ambitious plans, but I found myself questioning whether many end users will feel their effect.

Marvell is a player on the pre-OEM level, providing tech and chipsets to companies putting together devices like handsets, tablets, and e-readers. Kinoma, which I hadn’t heard of until today, is working on a sort of thin OS layer that goes on top of Android and a few other OSes, adding functionality and connectivity while keeping the footprint small and the compatibility level high. But could they be mistaken on where the platform is going?

Kinoma right now exists as a separate layer from the OS; you launch into it and it has its own interconnected set of call-outs, search protocols, rich displays, and so on. The Facebook app calls in the YouTube app when a friend links a video, for instance, and it plays in-line without leaving the app. Searching actually runs discrete searches within every application that has a search function and returns the results, sorted by app or library. It’s very well put together and the Android team could learn a lot from it.

Or could they? The object of Kinoma seems to be ubiquitous connectivity by using every app as a service and calling those pieces that are required by the user. A marvel of modularity, but while it’s comprehensive, it doesn’t seem… how do I put this — knowledgeable. That is, it’s producing a lot of data for you to peruse (and very quickly and efficiently), but it doesn’t seem to know what you want. Instead of doing or finding just what you want, it seems that Kinoma tends to do everything all at once, and assume that what you want is in that sea of data.

If you search for Pandora, for instance, there’s a very good chance you are trying to launch the app. But Kinoma will still query the Twitter app, Facebook, Wikipedia, even CrunchBase, and present that information as equally relevant. You can reorder the result fields, but it seems like giving the user a lake when they ask for a glass of water.

But to be fair, that’s the “full” Kinoma experience, and no one is under any pressure to use it as such. The framework is complete in each piece, since they’re written in near-universal Javascript and XML. It breaks down very easily into components and even single apps could be distributed — say, an “enhanced” Facebook app on handsets with an extra OS layer like Sense.

Speaking with Kinoma VP Peter Hoddie, I expressed concern that the look and feel of Android (such as it is) might be interrupted by this service butting in and providing extra whatnot in some places, while apps downloaded from the Market would lack any integration. That’s really more of a problem for Samsung and HTC, though — integrating the framework in a way that doesn’t interrupt the normal Android experience.

It also seemed to me that this environment is, in a way, a glorified browser layer with web apps. The interconnectivity and complementary features rather diminish that complaint, though this would be a very bad time to underestimate the versatility of web apps.

So what’s Marvell’s piece of this pie? My guess is that they’re going to do some smart deep integration with Marvell hardware, and shop the OS enhancements to OEMs that want to set their Android devices apart. That’s an increasingly difficult thing to do: here at MWC we’ve seen many, many Android devices, and apart from LG’s rather grotesque gestures towards 3D, they are to the end user almost indistinguishable. Being able to say “Hey, in addition to all the stuff our competitors have, our Facebook and YouTube apps are demonstrably better, and also we have this great dashboard, etc.” They do that already to some extent, but most of the “enhancements” provided by manufacturers are worse than useless.

Kinoma will also be released with an open-source license and the SDK will be made freely available (you can sign up for the beta here). Whether they’ll actually pull any developers in is an open question, but I’m guessing they need to get a little reach first. At any rate, it’s an interesting platform that, either appealingly or fatally, does not fall in line with current mobile platforms.