Sunday, December 6, 2009

Why Open source phones still fail

The ultimate hacker phone, the Nokia N900 is the truest expression of Linux—the OS and the philosophy—that you'll find on mobile this year. It's a great niche gadget. But the idea that free, open-source solutions will sweep the mobile world is just as doomed in the U.S. as the idea of popular, open-source desktop PCs—in this case, because mobile networks don't welcome the unexpected, and they don't welcome geeks.

The joy of the open-source world is the joy of the unexpected. Nothing really goes according to plan, because there isn't one plan; there are a lot of plans. Creativity bubbles up from the grass roots. Many projects fail, but some become very disruptive.

Wireless network planners hate the unexpected. They hate disruptive. They didn't expect how much iPhone users would surf the Web, and look where that's gotten them.

Phone manufacturers have been toying with Linux for years, for reasons that have nothing to do with open-source philosophy or communities. Non-smartphones typically use simple operating systems, which either aren't capable of running modern applications or don't have large developer communities. For years, the use of Linux was viewed as a way to leverage a large developer base and a modern kernel for midrange phones.

Most of those attempts failed. Motorola tried to replace its ancient, proprietary P2K platform with Linux, but ended up casting it off and returning to P2K. The LiMo Foundation tried to create a common Linux platform for midrange phones for a while, but none of its devices has broken through in the U.S. It seems that U.S. manufacturers and carriers just don't see enough advantage in Linux to bother putting a new system through the painful approval process here.

With one exception, of course: Android. Android is a Linux-based OS, although it doesn't sign on to that namby-pamby free, open-source philosophy. Third-party apps on Android (unless you kick the phone into developer mode) are sandboxed into a little Java virtual machine and locked into a single internal memory partition.

That's because while Linux the OS—the kernel, the memory manager—is attractive to phone manufacturers, Linux the philosophy—users banding together ad hoc to create new things—is anathema to wireless carriers.

Living in Fear
You can hate the wireless carriers all you want. Thing is, they're genuinely scared. They seem to consider their networks to be fragile, gossamer things, able to collapse in a strong breeze (or under the weight of a bunch of iPhone users.) I've become convinced that they don't cap your Internet connection at 5 GB out of greed. They cap it because if too many people use too much data—poof, it all goes down.

This has made carriers panicky about any uses they haven't pre-approved. Thus, we have the rather restricted nature of the iPhone App Store and the Android and BlackBerry platforms. You just aren't going to get U.S. carriers signing on to the philosophical parts of the Linux platform—experimentation, collaboration, and the unexpected. Network planners wake up in cold sweats in the middle of the night when faced with "the unexpected."

No carrier wants geeks. Geeks use up a lot of network resources, try to find ways around rules, and create problems for tech support. Every time a carrier has flirted with geeks, it has backed away. Helio was originally conceived as a "power user's carrier," but it did an unexplained about-face and decided to go for the social-networking youth when it launched. T-Mobile did a similar thing a few years ago, changing its theme from "get more" to "stick together." And Sprint sent out messages attracting geeks from time to time, but in recent years, it has backed away too, doing things like preventing tethering on handsets.

"But what about the Nokia N900?" you ask. "And for that matter, what about iPhone jailbreakers and the people who put Android phones into developer mode?"

It's always been possible to fly under the radar; there are a few dozen OpenMoko owners out there who can vouch for it. The networks can take a few mavericks that are willing to pay high prices, seek out obscure sales channels, or risk turning their phones into doorstops through jailbreaking. But the carriers know that as long as they subsidize phones, they'll control the phones are sold en masse. And those aren't going to be open devices.

So geek phones will never be mainstream, but a fortunate confluence of events might get the N900 into more people's hands soon.

Regular readers of my columns know that right now, I'm celebrating T-Mobile for offering no-contract plans that don't bake in a device subsidy. T-Mobile is also building a lot of new network capacity right now, and it doesn't seem to be as laden with heavy data users as AT&T or Verizon. It's no coincidence that the N900 comes with T-Mobile's 3G bands rather than AT&T's.

T-Mobile could start stocking unsubsidized N900s to show a different way of thinking from the other major carriers. The carrier could flirt with the geeks to boost its profile, attracting developers and creative types. Maybe this would give the number four carrier a leg up. Geeks may finally find their wireless carrier after all.




From http://news.yahoo.com/s/zd/20091204/tc_zd/246552