Every smartphone worth discussing this year was operated with a thumb on a keypad or a stylus on a resistive panel, and I have come to think that single fact explains why the category is still niche.
Look at what actually shipped. The N73 put a genuinely good camera behind a menu system that requires patience. The Pearl gave RIM its first device aimed at somebody who was not issued it by an employer, and did it by shrinking the keyboard until two letters shared a key. Palm shipped the Treo 680, which is the same Treo it has been shipping for years with the antenna tidied away. Windows Mobile 5 devices arrived from four manufacturers in a dozen carrier variants, each with its own firmware and its own idea of what the home screen is for.
Underneath, the year’s real progress was in the radio. HSDPA turned up on networks that had spent two years describing 3G as a data service without delivering data speeds that made anybody want to use it. That matters more than any handset released in 2006, because it is the thing that makes everything else worth building. A browser is intolerable at 2G speeds and merely bad at 3G ones, which is progress of a sort.
The camera race carried on producing the same result. Three megapixels became normal, autofocus stopped being remarkable, and the pictures did not get much better, because the constraint was never the sensor count. It was the lens diameter, the sensor size and the light available, none of which a marketing department can change. The one meaningful improvement all year was somebody putting a xenon tube in a phone, and almost nobody copied it.
What did not happen in 2006 is the thing I keep expecting. Nobody rethought the interface. Every device on sale assumes you will navigate a tree of nested lists using a directional pad, and the more capable the phone, the deeper the tree gets. Symbian, Windows Mobile and Palm all solved this the same way, which is to say none of them solved it. The stylus is an admission of failure wearing the costume of a feature: it exists because the targets on screen are too small for a finger, and the targets are too small because the interface was designed for a mouse and then shrunk.
I do not think you get mass adoption out of that. The people who own smartphones today are people who tolerate friction because the capability is worth it to them, and there are only so many of those. Everybody else looks at a device that needs a stylus and a manual and decides a phone that makes calls is fine.
The fragmentation is the other tax, and it falls hardest on developers, which means it eventually falls on users. Writing software for this market means picking a platform with maybe a third of the smartphone share, then handling the variants within it. Almost nobody bothers. The result is that the most capable computing devices most people have ever carried run essentially no third-party software, which is an absurd outcome for hardware this good.
My prediction for next year is that none of this gets fixed by the current players, because the current players are all winning at the game as it is scored today. Nokia has volume, RIM has the enterprise, Microsoft has the license revenue. Nobody in that group is desperate enough to throw out the interface and start again. The change, when it comes, comes from somebody who has nothing to defend.