Open Moto
When Motorola announced their 2007 device collection this week, some people were disappointed. Three of the phones debuted in February at 3GSM and the fourth is yet another version of the Razr. In the year of the iPhone, the Prada phone, the Nokia N95, and even the Helio Ocean, Motorola looks weak.
But in one very important respect, Motorola is ahead of all these competitors. Every phone in their 2007 collection is based on an open development platform: Symbian, Windows Mobile, or Motorola’s new home-brew version of Linux, which they call Linux/Java. The iPhone, the Prada phone, and the Helio Ocean are all closed devices.
Sorry, we don’t count web applications, Java running in a cage, Flash, or widgets in our definition of open. For the purposes of this article, if we can write an app that has access to all of the features on your phone and give it to you without your carrier’s permission, your phone is open; otherwise it’s not. We’d like the source code too, but we won’t insist.
The Nokia N95 runs Symbian, but Nokia’s mid-range handsets do not. What makes Motorola’s announcement so interesting is that the Razr2 – the successor to Moto’s 100 million unit monster – will run Linux.
(To be precise, Moto only committed to Linux on certain versions of the Razr2. Some may run AJAR or another proprietary Moto OS, and the SKUs for Verizon will have to run BREW.)
We believe that five years from now it will be very difficult to buy a phone that does not have an open application development environment, in the developed world at least. Why? It’s certainly not because of moral pressure from the developer community. And it’s not because consumers or enterprises are crying out for phones that can run third-party software. It’s because open platforms are the only cost-effective way for carriers to create the services that they want to sell to customers, while still being able to differentiate.
Currently, I believe that Nokia manages 10,000 software releases every month. One release means one version of one application – SMS, email, the browser, the JVM – for one model of handset for one operator, somewhere on earth. As handsets get more complicated and operators seek unique designs and features, this clearly does not scale. Nokia can refuse to customize handsets and lose market share to those OEMs who will, or they can enable operators to customize handsets themselves by building on open platforms. In the long run this approach is cheaper and faster for both Nokia and the operators, who can hire their own developers rather than waiting in line and paying for fully-loaded Nokia engineers. All the OEMs and operators realized this years ago, and began working with Microsoft, Symbian, multiple Linux vendors, and others that are now dead (Geoworks), dying (the Palm OS if not Access itself), or in the process of being resurrected (SavaJe).
What about Java and the browser? These became solutions for third-party developers – caged environments on the handset where we could make lots of noise and draw pretty pictures without hurting ourselves or anyone else. We couldn’t change the UI of the phone in Java or manage the camera or the contact list, which was fine as far as carriers were concerned, but they couldn’t either. As Mike anticipated, Sun wants to break out of the cage with JavaFX, but it may be too late.
Qualcomm managed both tricks with BREW, and for an encore built a very sophisticated billing solution for mobile content. BREW operators like Verizon and KDDI have complete control over the look and feel of their handsets, a bigger but stronger cage for third-party developers, and more ways to charge their customers more money than anyone else outside SK Telecom. But most of the world’s operators are reluctant to commit to Qualcomm for the same reason that they are reluctant to commit to Microsoft. And while they in turn may have some customers who demand Windows Mobile, they have none who demand BREW.
So the battle is down to Windows Mobile, Symbian, and a dozen variants of mobile Linux. There’s no reason to call a winner among those three. The PC market may be a monoculture, but that’s a bad thing, and the conditions that gave rise to it won’t be repeated. A better analogy is the console market, where there have always been at least three successful platforms at any given time, or the wider CE market, which is split between Windows CE, VxWorks, and another dozen variants of Linux.
The high cost of the hardware necessary to run these platforms has kept them confined to the high-end of the market so far, but Moore’s Law is taking care of that. At 3GSM one vendor demonstrated a single-chip implementation of mobile Linux; Symbian 9.5 will run on “feature phone hardware“; the Razr2 has an ARM11 processor; and so on. And then there are economies of scale. Some people don’t need Linux in their phones, but most people don’t need a color screen or a camera – try buying a new phone without them.
For all their recent difficulties, when it comes to open platforms Motorola may be better positioned than any other OEM. Nokia’s mid-price handsets still use the proprietary S40 platform; Sony Ericsson uses EMP; Samsung and LG have multiple independent design teams that use whatever comes to hand. Motorola has invested heavily in Linux, has the best-selling Windows Mobile phone in America, and through its acquisition of Sendo, some real expertise in Symbian. Watch closely.
UPDATE August 7: As raddedas pointed out in the comments, Motorola had never formally announced that they would enable developers to write native applications for their Linux-based phones. That announcement came today.