Mobile minitablets still grounded despite new tech

Whether you call it a UMPC, an MID or a minitablet, it's far from certain you'll ever call it a success. Photos: Intel ultramobiles

When it comes to developing the ultimate mobile device, the PC industry is still milling about in the parking lot trying to decide which on-ramp to take.

PDAs (personal digital assistants) are pretty much dead; smart phones are the domain of the cellular industry; and the newest push, the ultramobile PC, hasn't met with much early success. Of course, new technology isn't always embraced by the masses, especially when it's expensive and people don't understand why they need it.

Intel, one of the primary backers of the concept, is undeterred. Two weeks ago in Beijing, Senior Vice President Anand Chandrasekher took the wraps off the company's latest chips for UMPCs and outlined a newer product called the Mobile Internet Device (MID). He also revealed plans to use Intel's next generation manufacturing technology to build a chip that sips electricity, a must in any battery-powered mobile device.

But Intel's vision of the UMPC seems tied up with its push behind WiMax, a broadband wireless Internet technology that is still in development despite a high-profile commitment from Sprint. Mobile devices aren't really mobile if they won't work 100 feet outside of a Wi-Fi hot spot. Mobile devices like Samsung's BlackJack or Palm's Treo use cellular networks for data, and although Intel's partners plan to add that capability in later versions of the devices, it adds cost and complicates distribution.

Yet, mobile smart phones are precisely where Intel's eyes are trained. "When you look at the high end of the smart-phone category, that will likely get replaced by the UMPC," Chandrasekher said in comments after his keynote speech in Beijing.

A lot of dominoes will have to fall in the right places for that to happen: Intel will have to deliver its low-power chips, get its hardware partners to build compelling devices, convince developers to write software with a mobile experience in mind, and figure out a way to deliver an always-on Internet connection. And the final product has got to cost around $400.

The early iterations of the product aren't there yet. Samsung's Q1 was considered one of the better models, but that device failed to sell even 100,000 units, an executive said in March at the CeBit technology exposition. That's not even a blip in a worldwide PC market of almost 240 million units a year.

The Q1 costs around $1,000, and that's simply too much for something that's designed to be used as a supplement to both a PC and a cell phone, said Samir Bhavnani, an analyst with Current Analysis.

"If it's a secondary device, it cannot cost more than the primary device," he said. The average price of a laptop has fallen below $1,000, and even ultraportable laptops that a UMPC could conceivably displace don't cost too much more than that.

As with many things in the technology world, price is something that comes down over time. Intel and its partners think they can get the price down to around $700 this year, and about half that by 2009, said Roger Kay, an analyst with Endpoint Technologies Associates. "In the middle to low hundreds (of dollars), people say, 'OK, I'll give this a try.'"

Successful UMPCs or MIDs (or whatever name eventually sticks) will also need to come with compelling features, performance and usability. In 2008, Intel will release a chip based on its 45-nanometer manufacturing technology, called Silverthorne, that will deliver much more performance at a fraction of the power consumption of McCaslin, the product introduced at the last Intel Developers Forum. So on that score, there should be plenty of processing power for video entertainment, PowerPoint presentations and e-mail--the essentials for a highly mobile professional or consumer.

On the software front, Microsoft has been one of the biggest proponents of UMPCs and minitablets, which are part of its Origami project. But its might and muscle haven't helped sales of the devices, either. And Intel thinks that Windows Vista's code base might be too large to run effectively in an environment where power consumption is everything.

"We're working closely with Microsoft. They know what they need to do," Chandrasekher said in response to a reporter's question, after his keynote speech, about whether Microsoft and Intel are developing a stripped-down operating system that's somewhere between Vista and Windows Mobile. But Intel is also hedging its bets, announcing plans to promote Linux-based MIDs in 2008, based on its Menlow platform.

Microsoft didn't respond directly to questions about future plans for mobile operating systems or Intel's support for Linux. "We believe in the promise of the ultramobile form factor and the unparalleled strength of the Windows ecosystem. We'll continue to innovate with our hardware partners--including Intel--so people can stay connected, entertained and organized at the touch of a finger," Microsoft said in a statement.

Featured Video