Month: April 2010

I started working on the system architecture and protocols for Wi-Fi in late 1990, when I consulted with Photonics, a little start-up in Los Gatos that had already built the first commercial wireless LAN. The initial Photonics product was a short distance, infrared-based, wire replacement for Apple Talk, and the second generation system was Wi-Fi over infrared. Most people don’t remember that IEEE 802.11 was a single Medium Access Control protocol and two physical layers, one infrared and the other RF. The RF PHY was obviously more successful than the IR version.

Photonics had two large customers, IBM and Toshiba, both of whom wanted to do the same thing with the wireless LAN: integrate it into touchscreen, portable computers. IBM’s portable computer was called Think Pad, and as the name suggests, it was tablet computer with which the user would interact through a stylus; Toshiba had a similar idea. The user interface was based on gestures and handwriting recognition, a very rough science in those days, and the underlying system was Windows 3.15 or so, a keyboard-and-mouse system. These systems were challenged by limited battery life – like 2 hours between charges – and the slow processors of the day. So we figured out how to build a wireless LAN, but didn’t quite end up with a system that could take advantage of it. It wasn’t until nearly 10 years later that CPUs, PC technology in general, and batteries developed to the point that a fully portable computer was really practical, and by then Windows had established dominance in corporate America so the clamshell design that laptops use today won out over the more personal concept of the tablet. Clamshells weren’t all that revolutionary as a concept by the 90s, given clamshell prototypes were built at Texas Instruments, IBM, and other places in the 1970s. Just ’cause you can built a prototype doesn’t mean you can build a product, however: For portables, the technology has to be there to get decent performance, weight, and operating life from a charge. This is very clear in the case of handhelds, and all geeks are familiar with promising devices that never went anywhere because they failed on one of these three dimensions. The Google G1 phone drained its battery too fast, for example, so it was a total flop despite being an otherwise very nice device.

One interesting development that’s been taking place over the past couple of years is the convertible laptop, a clamshell computer with a touch-sensitive screen that swivels to turn the whole thing into a tablet. The Dell Latitude XT2 is the best example of this sort of system today.

This is nice machine, with a dual-core Intel CPU, several hours of battery life, and both pen and multitouch input. It weighs about 4 pounds with the serious battery, which is great for a laptop, and has all the nice connectivity options such as 802.11n and 3G from various sources. Nicely-equipped, it will set you back about $3,000 and it’s probably worth every penny of it if you need the features and functions it offers, especially handwriting recognition and a full-blown Windows platform. I’ve been tempted to buy one of these, to the extent of researching the refurbs you can get on eBay for half the retail price, but haven’t pulled the trigger. The price, weight, and the general concept as a Windows extension ultimately turned me off.

So I was pretty excited to learn that Apple was building a tablet computer. Apple is the one company with all the capabilities in hardware, software, user interfaces, and vision to develop a personal computing device that breaks new ground and isn’t chained to the past. Of all the personal technology companies in the world, Apple is the one with the least reverence for the traditional ways of doing things and the greatest autonomy in system design. Dell, HP, and the others are ultimately constrained by the dependence on either Microsoft or Linux for software, and it’s hard to push either of these highly successful enterprises into a totally new space very quickly.

As the specs came out, it became clear that Apple gets the limitations of the tablets of the past: the weight was only 1.5 pounds, the battery lasts all day (doing serious work, like video rendering), the connectivity options are all in place in terms of dual-band 802.11n and 3G + GPS for the model shipping later this month. The storage is all solid-state, and the OS is a slightly scaled-up version of the scaled-down Mach system known as OS X by Apple. The price is pretty appealing too, at $729.00 for the 3G model with a 32GB SSD.

The announcement was well in advance of the ship date, so I did what any sensible person would do, bought some Apple stock. The stock has now gone up enough to pay for an iPad 3G with a nice set of accessories and to cover taxes on the gain. So I put in the order at Apple’s on-line store for delivery sometime later this month. So we’ll see how it goes. If I don’t like it, I’ll send it back, probably without a vanity video about how disappointed I am in Steve Jobs. And yes, I expect that the second generation model will be faster, cheaper, 4G-enabled, multi-tasking, and camera-equipped. By then, I will have had 6 months or year’s worth of use from the one I’ve ordered, so that’s not too bad. I bought one of the first Macs in the first 100 days as well, since I liked all that bitmappy, mousey, windowey GUI stuff, and used it for 3 years or so before trading up to a Mac SE and then to a machine that would run Windows 95 at a decent speed.

I expect that we’re about three to five years away from general-purpose tablets that don’t need hands-on curated app stores to ensure consistency and quality in their user interfaces, so at some point the prophets of gloom will be able to buy a lightweight, fully functional, open system that does all the things they need.

Contrary to popular opinion, the cycle of innovation doesn’t always move from open devices to closed ones, of course. What we really see in the long arc of platform innovation is that closed devices like the Xerox Star and the iPhone lead the way, only to be cannibalized over the long term by open systems built on generic technology like Wintel machines.

When a new paradigm is emerging, however, the trade-offs between power, usability and cost are too fragile to accommodate the looseness in interface design and the over-engineering you need to have in order to accommodate unknown apps. We’ll get to open tablets eventually, but only as the hardware and software of the underlying platforms develop to the point that we can afford the overhead to provide openness. And by then, Apple will be pioneering a whole new concept in personal communication, computing, and entertainment. This is as it should be, of course.