Saturday, December 14, 2002

Wednesday, December 11, 2002

Monday, December 09, 2002

In reading Dan's comments on the TabletPC I'm reminded of the importance of shipping a product in order to learn what it really is. We shouldn't be very surprised that there have not been great advances in pen software since our attempts in the early 1990's. We've had to wait for more suitable hardware and a more powerful software base. Now that we have a device that is useful as a laptop we can start experimenting with the possibilities of the tablet. It is just the start of a learning process.

In the early 1990's we had a chance to work with similar devices but the emphasis was on using the pen and handwriting. The devices were expensive and cumbersome for just filling out forms and handwriting was not a feature in itself. Typing is a far better way to work with the computer. But Dan observed that the real value of the pen was in drawing and direct interaction.

The NEC-Versa was an interesting precursor of the tablet PC. The screen was detachable and could be mounted face out. But there was no way to interact directly with the screen. The feature was useful for sales people giving presentations but didn't give me the ability to use the machine more like a book.

With the Tablet PCs we can interact with the surface and use it more like a book. One reason we had to wait is that the table PC is not, in itself, a major product and thus is built out of existing technologies designed for other purposes. This is typical -- it doesn't make sense to make major speculative investments when you can make incremental changes and learn what is really important.

The floppy disk drive is a good example. The early (5�") floppy drives were twice as high as today's disk bays because they were built using motors from washing machines. Eventually there were enough PCs to make it worth developing motors just for them and floppy drives could become very thin.

This first generation of tablet PC's is limited by the digitizing surface so we have screens that are only 1024x768 and the resolution of inking is barely acceptable. The fact that these machines exist at all is thanks to Microsoft's interest in such machines, an interest that goes back to the pen computing in the early 90's and their experiments with Windows for the Pen. The growth of this market will depend on Microsoft's commitment through the learning phase. I do expect there will be some enthusiasts but it will take time to understand the native capabilities and many will be put off by the price and bulk of the initial machines.

Microsoft's persistence is evidenced in the PocketPC which started as a weak competitor to the Palm PDAs but has become an interesting platform in its own right. Yet this same persistence has kept the platform innovation within tight bounds. To be fair I have not seen much innovation on the non-Microsoft PDAs either though this may be because they find it most economical to build on the platforms defined for PDAs. I hope that that TabletPC follow Eric von Hippel's advice and observe how people (mis)use the devices and rush to make them more than just PC's with pens.

For me, it's the smart surface coupled with local computer that is most interesting. But I'll have to use one for a while to get a better understanding of the possibilities. I take the lack of availability as a sign of strong interest yet caution on the part of the manufacturers. The one machine that is available is from Acer and it seems like a very good machine but the 256MB limit is a problem. I feel strange saying that given that I grew up with machines that had 16KB but the memory budget for applications continues to grow. Memory is abundant and it makes sense to use more of it in order to gain programmer productivity and add capabilities. There is a tension between the compromises necessary to reduce the current costs and the need to preserve the opportunities in the future and not providing sufficient memory is a common mistake since the there isn't a problem when testing with just a few programs and it isn't obvious when slow performance in the field is due to a lack of memory and not a slow processor.

The memory limitation is relatively minor but the tendency to focus on optimizing the technology for a particular niche is strong, especially in the consumer electronics industry. The excitements about "network computers" that had no local storage and were entirely dependent on the office network and their cousin in the consumer world, the browser machine, are examples of this.

The strength of the hardware/software split is in allowing independent innovation in hardware and software. The TabletPC illustrates the limitations on this particular marketplace architecture. When form is a key part of the function, the same dynamic limits innovation because we have to waiting for the new form. While I can easily innovate in software it is harder for me, on my own, to innovate in hardware. Networks give me a capability to mix and match hardware. What's missing are the building blocks for new kinds of devices -- figuring out what those building blocks should be is at the forefront of taking computing from the fixation on the glass screen and browsing. Sun's JXTA and Microsoft's SPOT are part of this along with many others.

In the meantime I do plan to get a TabletPC because I want to learn. And, given that I worked with Dan at Slate, I also feel obliged. More important is the opportunity to learn about new ways to use computers.

The art of folding laptops reminds me of the challenge of reading newspapers on the subway in New York. In order to read the New York Times you needed to learn to fold the newspaper to one fourth the size. The skill was in turning the page without having to fully unfold the paper. Those who couldn't do that had to read the tabloids. As proud as I am of that skill, I look forward the ability to read on my tablet PC without having to unfold and sit it on my lap.