Posted by terrencebarr
on November 6, 2007 at 2:54 PM PST
- Update 1: Dalobor Topic has just posted this about the Android license.
- Update 2: Check out this MacWorld article in which Rich Miner talks about the Android non-fragmentation pledge. Proprietary extensions and consistency in the platform? Sounds like wishful thinking to me.
Over the last 24 hours very little news and lots of opinion and speculation hit the airwaves about Google's Android platform and the Open Handset Alliance . Our very own community member Sean Sheedy has this very detailed blog entry to offer.
After digesting the first barrage of information and talking to a couple of folks in the industry I thought I'd offer up my own personal thoughts about Android and the OHA.
First, at a high level, I think the Android announcement shows that the era of proprietary and closed mobile platforms and networks is finally drawing to an end (remember that Google has also indicated they will participate in the auction of the 700 Mhz spectrum). And that's goodness because locked platforms and networks controlled tightly by vendors have been a huge barrier to innovation and ignored the needs of consumers for too long ("Ten things I want from my phone" ). I commend Google for taking this step.
But once you dig a little deeper into what Android is and what the business model and motivations are I think you'll be left scratching your head a bit. The Android press release and OHA website offers surprisingly little information (and I was a bit annoyed by this video which offers meaningless babble and a dog sitting in for Andy Rubin).
The consensus seems to be that Android is a Linux- and Java-based stack including the low-level OS, middleware, application layers, and some basic applications - at first glance very similar to Sun's JavaFX Mobile (but without a key component FX Script ). It appears to be highly customizable allowing, as Eric Schmidt puts it, "thousands of different phone models" (in contrast to JavaFX Mobile which places a strong emphasis on consistency and predictability). The stack is licensed under Apache v2 and an early access SDK will be available next week.
OHA founding members are Google, T-Mobile, HTC, Qualcomm, Motorola, and Aplix. Other notables on the current member list are Esmertec, Intel, KDDI, LG, NTT DoCoMo, Samsung, Sprint Nextel, and Telefonica. That's impressive but also noteworthy for who is not on that list: Nokia, SonyEricsson, and most carriers around the world (except for Japan where OHA has good coverage with DoCoMo and KDDI). Also, remember that being part of a press release is easy and says very little about how much skin you will actually have in the game at a later point in time.
So what does it all mean?
Right away, the questions that come to mind are: How is this different from the existing 40+ Linux-based handset platforms? The LiMo Foundation , the LiPS Forum , and other efforts in this space? And who is the customer for this? Here a very interesting blog by Carl Howe "What's the Point of Google's Phone?" .
But every more perplexing is the question of what the business case and motivation is for OEMs and carriers to be part of Google's push. In the mobile space there is a very strong established pattern of maintaining tight control of the platform itself, the content on the platform, and the data streams to and from the platform to support business models that drive up the ARPU (average revenue per user). OEMs and carriers rely on proprietary features, walled gardens (directly, but more frequently indirectly), and business arrangements to implement these models. That's not to say they are not investigating new business models but history suggests that the closed model is something they are very comfortable with.
Google, on the other hand, built its business around the open Internet and freely accessible content and applications running on standardized platforms (such as the web browser). Of course, the ad business accounts for something like 95% of Google's revenue so I assume Google's main motivation for Android is to build a new ad delivery platform with a potential for billions of new users and a reach they just never could have achieved on the desktop.
Fair enough, but Google's model is diametrically opposed to the approach of OEMs and carriers. If Google's goal is to drive mobile users to the Google portal and use Google services anywhere at any time on any platform where does that leave the carrier and OEM? They basically become a commodity platform and a dumb data pipe with the purpose of driving Google eyeballs. Sure, Google can offer them revenue sharing and other incentives but fundamentally they would loose the leverage and monetization points they have built a huge business around.
This is not a trivial problem. In fact, it is one of the key reasons the mobile space today is what it is, with a multitude of vendor-specific platforms, software stacks, proprietary features, and limitations. It's also a situation Sun and the Java ME ecosystem have been having to deal with since the release of KVM in 1997. One of the key reasons Java ME has been extremely successful (admittedly not without its share of pain and frustration) is because Sun's position allowed it to be a neutral 3rd party that provided a platform and ecosystem but didn't need to enforce any particular business model. So it was a natural symbiosis for OEMs and carrier.
And this point precisely is where the Android news leaves a big void. It's like fitting a square peg into a round hole. I don't see how Google's business model and the OEM and carrier model align sufficiently to make this a happy partnership.
And what's ominous for developers is that the Android license choice and FAQ explicitly encourages proprietary extensions on the platform as well as the removal of functionality - which spells trouble for delivering a widespread integrated platform that allows developers to build compelling services easily and quickly. This seems to haveÂ "fragmentation" of the worst degree written all over it. Imaging thousands of unique Android-based platforms ... Java ME fragmentation, with all its quirks, may be a piece of cake compared to that.
Which leaves us with the final question. What about Sun and Java ME?
Obviously, Sun isn't part of the OHA at this point - I am disappointed at this but I assume the OHA and Sun had disagreements over business and strategic issues that made a joining of Sun at that time impossible. That might change but I am not privy to those discussions.
Next, why is Java ME not mentioned in the press release? Well, since Aplix and Esmertec are part of OHA I think that is a strong indication Java ME is a key component of the Android platform. And Google being famously secretive I am not surprised they are not volunteering that information. Fact is, Google recognized Java to be a necessary part of success. It literally guarantees Java remains the dominant platform for years to come.
So what aboutÂ phoneME ? phoneME is fully open and licensed under GPLv2 in the Mobile & Embedded Community . Sun believes the GPL license is the best choice for reasons of compatibility with GNU/Linux as well as to ensure ongoing innovation benefits everyone in the ecosystem while maintaining a consistent platform. Google has chosen Apache v2 for Android which means the licenses are incompatible and therefore phoneME cannot be part of Android at this point in time. If and how this might be resolved in the future is speculation and not a particularly useful exercise right now.
But with Sun's vision of the JavaFX Mobile as the next-generation mobile software platform Java wins in any case
Welcome to exciting times.
PS: I'll take the opportunity to remind you of the upcoming Java Mobile & Embedded Developer Days conference. Registration is now open. Check it out here.
Powered by Qumana