[cobalt @ Sep 29th 2006 9:10AM]

I followed the Origami announcements and the first UMPC releases with a lot of interest.

The hardware elements were clearly almost there but not quite. Almost every component was just a little off from where you wanted it to be.

The software on the other hand didn't get much thought. Every piece of software for touchscreen has basically been optimized for stylus, not touch. Windows Mobile phones are a good example of this. Dialkeys is not an adequate substitute for even a simple thumbpad. The reason everyone complained about the lack of keyboard on the 1st gen UMPCs is because it is a computer, not a mobile device. Finally, if the device is meant to be mobile in general, the UI has to represent that fact, and not just the main application launcher, but every piece of software that provides useful mobile functions. There is a reason why Streetdeck is the most applicable software set for UMPCs, and that's because it wsa built from the ground up around in-car use. Windows OS on its own does not have what it takes to deal with a small high resolution touchscreen.

IMO, the UMPC needs to be more of a lifestyle device to succeed than just a smaller computer, with or without a keyboard device. In order to be usable in multiple contexts, it needs to be optimized for at least one context. Some poeple want what represents simply a smaller computer, but not most people. That's where the initial UMPC release failed: a lack of vision and hard work to realize the implied concept. Maybe MS was not so keen on the first gen UMPCs but went ahead with it anyway. Or maybe they devoted very few resources to it to start off, while busy working things out in Vista, including support for these small-screen mobile PCs. Maybe Intel wanted to push ahead to break in the consumers to the general idea, just as they are doing with Viiv, even though no one really understands what it might be good for over another PC.

In order for the UMPC to succeed, they need to work on integrated applications, speed, and responsiveness. They need to focus on the non CPU components more than they realize, which means the costs will be higher. No matter how fast the CPU is or how much RAM, a 1.8" HDD will only be so fast, while a 2.5" HDD will be larger and possibly create more heat and increase the power requirements, which are already thinly stretched. And they need to focus on actual usage scenarios to understand what software and hardware is required. I don't think they did enough work on this initially, although what I see here looks promising.