How many mobile device manufacturers does it take to keep the most successful handset operating system alive? If you guessed "one," you're right. If you guessed "five," you're right. If you're confused, you're in good company.
Nokia recently acquired Symbian Limited, the developer of the Symbian mobile OS licensed by Nokia, LG, Motorola, Samsung, and Sony Ericsson and used as a system software component in their mobile platforms. Perhaps it's just me, but I find this story remarkable on many levels. Knowing no more than the single fact I've presented, it might appear that Nokia, which already held a controlling interest in Symbian, is moving to pull the rug out from under its competitors. It turns out that Nokia's not the latest antitrust bad boy. Put a cape on Nokia, because it is a champion of corporate trust, or whatever anti-antitrust works out to be. Symbian Limited's engineers will wear Nokia badges, but every line of code they crank out will be turned over to Nokia’s competitors, and later, to the world.
Nokia will bring Symbian's development operations in-house, but Nokia won't own Symbian's intellectual property. With the purchase, Nokia simultaneously established the nonprofit Symbian Foundation, whose members include five Symbian licensees as well as a trio of major wireless carriers (NTT DoCoMo, Vodafone, and AT&T) and a pair of embedded semiconductor manufacturers (TI and STMicro). The Foundation effectively owns Symbian OS.
If the Symbian Foundation membership seems an odd assortment, consider this: If Foundation members could agree on a set of objectives, it might be able to drive a new device from concept to wireless network deployment in a fraction of the time it takes today. Just eliminating the duplicated effort that each handset maker wastes in homebrew workarounds for issues that are fixed in later revisions of the Symbian OS would be a huge shared win. Having all of the vendors tap the same code spigot at the same time would make patches easier to distribute and, dare I imagine it, make it possible to write applications to Symbian OS that run on multiple brands of phones. That's pretty much the message of the PR accompanying Nokia's announcement.
Nokia bought Symbian with the stated long-term intention of giving the OS away as proper open source, a detail that has drawn the focus of observers who see the Symbian Foundation as a bulwark against Google's Open Handset Alliance. My own knee-jerk response was to view the Foundation as a circling of wagons against iPhone. The timing of the announcement, about two weeks before iPhone 3G's availability, strains coincidence. Whatever the perceived external challenge, the Symbian licensees felt compelled to come together to do something that they couldn't achieve separately. I wonder if Symbian OS was about to hit a wall that could only be avoided if the strongest engineering force among the Symbian Five took over development.
Handset makers in the Symbian Foundation have to prep themselves for the next wave of high-end handsets. The brick on your belt can't merely make phone calls and swap text messages. It will have to be a camcorder, a still camera, a television, an iPod, an Office desktop, a Wi-Fi/Bluetooth/cellular wireless data gateway, a navigator and presence beacon, a 3-D game console, and an Internet tablet. As outrageous as it seems, especially considering that most high-end devices are targeted at commercial buyers, device manufacturers will have to treat this "prosumer" feature set as baseline and innovate within and above it.
The engineering catch is that these advancing features must be accommodated while simultaneously lengthening battery life from model to model. That job falls largely to the OS, and I think the Symbian Foundation came together because the licensees decided that none of them could get to next-wave devices with Symbian as it is.
That wouldn't be surprising. Google and Apple decided that paving the way for uberphones (I'm running out of neologisms) justified a trip back to the drawing board for system software. Each Symbian licensee has checked out Symbian alternatives; Nokia and Motorola, at least, have non-Symbian devices in their lineups. But each has decided that for its next push, it has to use what it's got, and that's Symbian.
What TI and ST bring to the Foundation is new generations of application processors. Texas Instruments, a hometown favorite of mine, just introduced a nifty $50 chip sandwich that combines two RISC CPUs with a 3-D graphics accelerator, memory, and a plethora of peripheral controllers. $50 for a single component telegraphs the cost of a device that might integrate and use it to full advantage -- at least $500 before wireless carrier subsidies. Since the ARM CPU architecture dominates mobile devices, TI's new component could be used by any modern mobile software platform. But when reviewers start expressing mobile device metrics in terms that include frames per second and megabits per second, an OS' dexterity at juggling hundreds of threads while remaining responsive to the user becomes paramount.
The devices that blow the doors off iPhone 3G already exist in simulators. Market ranking and device capabilities won't be decided by hardware. It will all come down to software. As unglamorous as Symbian or any other operating system may be, no mobile device vendor can lead the market without leading in OS engineering. Whether the Symbian Foundation can put Symbian OS on the space age trajectory is an open question. If it can't, there will be five handset manufacturers standing at the sidelines when the next major mobile leap is taken.